Society of Surgical Oncology (SSO): Annual Cancer Symposium

ORLANDO – Treatment centers and surgeons tend to play to their strengths when choosing therapy for patients with well-compensated cirrhosis of the liver and early hepatocellular carcinoma, investigators reported at a symposium sponsored by the Society of Surgical Oncology.

Therapy for early HCC with well-compensated cirrhosis is controversial; there is little agreement on when resection, transplantation, or radiofrequency ablation becomes the best approach. Choice of therapy for early HCC often depends on the surgeon’s repertoire of techniques and the therapeutic services the hospital offers, based on the findings of a web-based survey of centers that had at least five HCC cases per year.

"This study demonstrates that nonclinical factors have an important effect of therapy for early HCC, and in particular the choice of therapy depends in part on the surgeon’s portfolio of techniques, as well as the availability of transplantation services," said Dr. Hari Nathan of the department of surgery at Johns Hopkins Hospital in Baltimore.

In a previous analysis of the data from their web-based survey, Dr. Nathan and colleagues found that surgeon specialty was more important than certain patient-specific factors when determining treatment choice (J. Clin. Oncol. 2011;29:619-25).

"Differences in choice of therapy for nontransplant and transplant surgeons were not the result of an across-the-board preference for one therapy vs. another. Rather, some clinical factors impacted surgeons differently, depending on their specialty," he said.

In the new analysis, the authors used the survey data to assess the effect of surgeon and hospital factors on the choice of therapy for early, well-compensated HCC, and the effect of regional liver transplantation services on the surgeon’s choice of therapy.

They defined early HCC according to the Milan criteria as a single tumor less than 5 cm in its largest dimension, or two to three tumors less than 3 cm. Cirrhosis was considered to be well compensated if it was Child-Pugh class A, with no varices, ascites, or encephalopathy.

They presented respondents with case scenarios factoring in age, tumor number and size, type of resection required, etiology of cirrhosis (hepatitis B or C, or alcoholic), biological MELD (Model for End-Stage Liver Disease) score, platelet count, and anticipated transplantation waiting time.

Of the 1,032 invitations they extended, 336 surgeons (33%) responded. Of the respondents, 284 (85%) were in academic practices and 52 (15%) were in community practices for a median of 10 years (range, 4-17 years). About two-thirds (65%) were trained in liver transplantation. Procedures performed for HCC included transplantation and radiofrequency ablation (41% of responders), transplantation alone (14%), or liver resection but not transplantation (45%). Asked which procedures were available at their primary hospital (regardless of whether the respondent performed them personally), 100% said that resections were available, and 99% said that ablations were available. In contrast, transplantations were available at 71% of respondents’ hospitals.

The authors found that neither years in practice, surgical oncology training, nor liver transplantation training had a significant effect on treatment choice. Similarly, regional transplantation variables – such as number of procedures, percentage of transplant recipients with HCC, 30th percentile of liver transplantation wait time, and severity of illness by median MELD score – did not significantly predict treatment choice.

There was, however, significant variation in therapeutic choice based on practice type, adjusted for case presentation, with surgeons in academic practices favoring transplantation 57% of the time, compared with 47% for those in community practice. Community-based surgeons were more likely to favor liver resection (45% vs. 38% for academic surgeons), and radiofrequency ablation (9% vs. 4%).

In regression analysis that controlled for clinical factors, they found that surgeons in academic setting were significantly less likely than community-based surgeons to recommend ablation over liver transplantation (relative risk ratio [RRR], 0.41; P = .01). When they looked at the effect of practice types’ controlling for surgeons’ specialties, however, the significance of the practice type on treatment choice disappeared.

Regression analysis also showed that "higher volume surgeons prefer transplantation over resection more strongly than lower-volume surgeons," Dr. Nathan said.

High-volume surgeons (defined as those performing 30 or more cases annually) were overwhelmingly transplantation surgeons; when the authors adjusted for whether the surgeon performed transplantations, the preference for transplantation disappeared.

Additionally, nontransplantation surgeons who worked at hospitals where transplantations were available were more likely to recommend transplantation over ablation, compared with surgeons working at nontransplantation hospitals.

"Interestingly, they also favored resection over radiofrequency ablation more strongly. This appeared to be a separate phenomenon than the one that we observed for the portfolio – that’s personally performed by each surgeon – and in regression analyses these effects were independent," he said.

Coauthor John F.P. Bridges, Ph.D., provided financial and administrative support for the study. Dr. Nathan reported no relevant financial disclosures.

ORLANDO – Treatment centers and surgeons tend to play to their strengths when choosing therapy for patients with well-compensated cirrhosis of the liver and early hepatocellular carcinoma, investigators reported at a symposium sponsored by the Society of Surgical Oncology.

Therapy for early HCC with well-compensated cirrhosis is controversial; there is little agreement on when resection, transplantation, or radiofrequency ablation becomes the best approach. Choice of therapy for early HCC often depends on the surgeon’s repertoire of techniques and the therapeutic services the hospital offers, based on the findings of a web-based survey of centers that had at least five HCC cases per year.

"This study demonstrates that nonclinical factors have an important effect of therapy for early HCC, and in particular the choice of therapy depends in part on the surgeon’s portfolio of techniques, as well as the availability of transplantation services," said Dr. Hari Nathan of the department of surgery at Johns Hopkins Hospital in Baltimore.

In a previous analysis of the data from their web-based survey, Dr. Nathan and colleagues found that surgeon specialty was more important than certain patient-specific factors when determining treatment choice (J. Clin. Oncol. 2011;29:619-25).

"Differences in choice of therapy for nontransplant and transplant surgeons were not the result of an across-the-board preference for one therapy vs. another. Rather, some clinical factors impacted surgeons differently, depending on their specialty," he said.

In the new analysis, the authors used the survey data to assess the effect of surgeon and hospital factors on the choice of therapy for early, well-compensated HCC, and the effect of regional liver transplantation services on the surgeon’s choice of therapy.

They defined early HCC according to the Milan criteria as a single tumor less than 5 cm in its largest dimension, or two to three tumors less than 3 cm. Cirrhosis was considered to be well compensated if it was Child-Pugh class A, with no varices, ascites, or encephalopathy.

They presented respondents with case scenarios factoring in age, tumor number and size, type of resection required, etiology of cirrhosis (hepatitis B or C, or alcoholic), biological MELD (Model for End-Stage Liver Disease) score, platelet count, and anticipated transplantation waiting time.

Of the 1,032 invitations they extended, 336 surgeons (33%) responded. Of the respondents, 284 (85%) were in academic practices and 52 (15%) were in community practices for a median of 10 years (range, 4-17 years). About two-thirds (65%) were trained in liver transplantation. Procedures performed for HCC included transplantation and radiofrequency ablation (41% of responders), transplantation alone (14%), or liver resection but not transplantation (45%). Asked which procedures were available at their primary hospital (regardless of whether the respondent performed them personally), 100% said that resections were available, and 99% said that ablations were available. In contrast, transplantations were available at 71% of respondents’ hospitals.

The authors found that neither years in practice, surgical oncology training, nor liver transplantation training had a significant effect on treatment choice. Similarly, regional transplantation variables – such as number of procedures, percentage of transplant recipients with HCC, 30th percentile of liver transplantation wait time, and severity of illness by median MELD score – did not significantly predict treatment choice.

There was, however, significant variation in therapeutic choice based on practice type, adjusted for case presentation, with surgeons in academic practices favoring transplantation 57% of the time, compared with 47% for those in community practice. Community-based surgeons were more likely to favor liver resection (45% vs. 38% for academic surgeons), and radiofrequency ablation (9% vs. 4%).

In regression analysis that controlled for clinical factors, they found that surgeons in academic setting were significantly less likely than community-based surgeons to recommend ablation over liver transplantation (relative risk ratio [RRR], 0.41; P = .01). When they looked at the effect of practice types’ controlling for surgeons’ specialties, however, the significance of the practice type on treatment choice disappeared.

Regression analysis also showed that "higher volume surgeons prefer transplantation over resection more strongly than lower-volume surgeons," Dr. Nathan said.

High-volume surgeons (defined as those performing 30 or more cases annually) were overwhelmingly transplantation surgeons; when the authors adjusted for whether the surgeon performed transplantations, the preference for transplantation disappeared.

Additionally, nontransplantation surgeons who worked at hospitals where transplantations were available were more likely to recommend transplantation over ablation, compared with surgeons working at nontransplantation hospitals.

"Interestingly, they also favored resection over radiofrequency ablation more strongly. This appeared to be a separate phenomenon than the one that we observed for the portfolio – that’s personally performed by each surgeon – and in regression analyses these effects were independent," he said.

Coauthor John F.P. Bridges, Ph.D., provided financial and administrative support for the study. Dr. Nathan reported no relevant financial disclosures.

ORLANDO – Treatment centers and surgeons tend to play to their strengths when choosing therapy for patients with well-compensated cirrhosis of the liver and early hepatocellular carcinoma, investigators reported at a symposium sponsored by the Society of Surgical Oncology.

Therapy for early HCC with well-compensated cirrhosis is controversial; there is little agreement on when resection, transplantation, or radiofrequency ablation becomes the best approach. Choice of therapy for early HCC often depends on the surgeon’s repertoire of techniques and the therapeutic services the hospital offers, based on the findings of a web-based survey of centers that had at least five HCC cases per year.

"This study demonstrates that nonclinical factors have an important effect of therapy for early HCC, and in particular the choice of therapy depends in part on the surgeon’s portfolio of techniques, as well as the availability of transplantation services," said Dr. Hari Nathan of the department of surgery at Johns Hopkins Hospital in Baltimore.

In a previous analysis of the data from their web-based survey, Dr. Nathan and colleagues found that surgeon specialty was more important than certain patient-specific factors when determining treatment choice (J. Clin. Oncol. 2011;29:619-25).

"Differences in choice of therapy for nontransplant and transplant surgeons were not the result of an across-the-board preference for one therapy vs. another. Rather, some clinical factors impacted surgeons differently, depending on their specialty," he said.

In the new analysis, the authors used the survey data to assess the effect of surgeon and hospital factors on the choice of therapy for early, well-compensated HCC, and the effect of regional liver transplantation services on the surgeon’s choice of therapy.

They defined early HCC according to the Milan criteria as a single tumor less than 5 cm in its largest dimension, or two to three tumors less than 3 cm. Cirrhosis was considered to be well compensated if it was Child-Pugh class A, with no varices, ascites, or encephalopathy.

They presented respondents with case scenarios factoring in age, tumor number and size, type of resection required, etiology of cirrhosis (hepatitis B or C, or alcoholic), biological MELD (Model for End-Stage Liver Disease) score, platelet count, and anticipated transplantation waiting time.

Of the 1,032 invitations they extended, 336 surgeons (33%) responded. Of the respondents, 284 (85%) were in academic practices and 52 (15%) were in community practices for a median of 10 years (range, 4-17 years). About two-thirds (65%) were trained in liver transplantation. Procedures performed for HCC included transplantation and radiofrequency ablation (41% of responders), transplantation alone (14%), or liver resection but not transplantation (45%). Asked which procedures were available at their primary hospital (regardless of whether the respondent performed them personally), 100% said that resections were available, and 99% said that ablations were available. In contrast, transplantations were available at 71% of respondents’ hospitals.

The authors found that neither years in practice, surgical oncology training, nor liver transplantation training had a significant effect on treatment choice. Similarly, regional transplantation variables – such as number of procedures, percentage of transplant recipients with HCC, 30th percentile of liver transplantation wait time, and severity of illness by median MELD score – did not significantly predict treatment choice.

There was, however, significant variation in therapeutic choice based on practice type, adjusted for case presentation, with surgeons in academic practices favoring transplantation 57% of the time, compared with 47% for those in community practice. Community-based surgeons were more likely to favor liver resection (45% vs. 38% for academic surgeons), and radiofrequency ablation (9% vs. 4%).

In regression analysis that controlled for clinical factors, they found that surgeons in academic setting were significantly less likely than community-based surgeons to recommend ablation over liver transplantation (relative risk ratio [RRR], 0.41; P = .01). When they looked at the effect of practice types’ controlling for surgeons’ specialties, however, the significance of the practice type on treatment choice disappeared.

Regression analysis also showed that "higher volume surgeons prefer transplantation over resection more strongly than lower-volume surgeons," Dr. Nathan said.

High-volume surgeons (defined as those performing 30 or more cases annually) were overwhelmingly transplantation surgeons; when the authors adjusted for whether the surgeon performed transplantations, the preference for transplantation disappeared.

Additionally, nontransplantation surgeons who worked at hospitals where transplantations were available were more likely to recommend transplantation over ablation, compared with surgeons working at nontransplantation hospitals.

"Interestingly, they also favored resection over radiofrequency ablation more strongly. This appeared to be a separate phenomenon than the one that we observed for the portfolio – that’s personally performed by each surgeon – and in regression analyses these effects were independent," he said.

Coauthor John F.P. Bridges, Ph.D., provided financial and administrative support for the study. Dr. Nathan reported no relevant financial disclosures.

ORLANDO – Intensive follow-up of patients who have undergone surgery for hepatocellular carcinoma reduces deaths from tumor recurrence and metastases, Dr. Timothy M. Pawlik said at a symposium sponsored by the Society of Surgical Oncology.

Data from surveillance programs for hepatocellular carcinoma (HCC) and empiric data from centers treating colorectal liver metastases (CRLM) suggest that HCC tends to recur locally, and that recurrent HCC and CRLM, if caught early, can be successfully controlled with a variety of therapeutic options, said Dr. Pawlik, associate professor of surgery and oncology, and hepatobiliary surgery program director at Johns Hopkins Medical Center, Baltimore.

"I would favor high-intensity surveillance for patients with hepatocellular carcinoma. This argument is based on level 1 data showing that high-intensity primary surveillance decreases mortality for patients who have cirrhosis," he said.

Although there is a high HCC recurrence rate following surgery, most recurrences will be contained within the liver, and may be successfully treated with salvage transplantation, ablation, or intra-arterial therapy.

"But if we miss that opportunity and patients recur with advanced disease, all those options are off the table, and their prognosis is abysmal," Dr. Pawlik said.

According to National Cancer Institute data, liver cancer holds the dubious distinction of being the fastest growing cancer in terms of death rate in the United States, outpacing lung cancer in women, esophageal cancer, and thyroid cancer.

Risk factors for HCC include cirrhosis from any cause (hepatitis B and C viruses, alcoholism, nonalcoholic fatty liver disease), hepatitis B primary infection (with or without cirrhosis), and inherited metabolic diseases such as hemochromatosis, alpha-1 antitrypsin deficiency, glycogen storage disease, or tyrosinemia.

"Most patients who have HCC don’t simply have a cancer; they also have significant underlying cirrhosis, so when thinking about the approach to HCC, we have to be thinking about all of the tumor-specific factors such as tumor size, location, and number, and we also have to be thinking about all of the liver-specific factors," Dr. Pawlik said.

Surgical treatment options include resection for HCCs of all sizes; transplantation, for single lesions 5 cm or smaller, or up to three lesions of 3 cm or less or advanced cirrhosis; and ablation for small lesions or inoperable or unresectable tumors as a bridge to transplant.

In various series, overall 5-year survival following resection ranges from 42% to 62%, and following transplantation from 57% to 75%.

But as Dr. Pawlik and colleagues noted in a 2008 study, disease-free survival following resection for HCC is only half of the percentage after transplantation (40% vs. 82%, P less than .01). (J. Gastrointest. Surg. 2008;12:1699-1708).

Intrahepatic recurrence is most frequent among patients with hepatitis C infection, but also occurs with hepatitis B and C coinfection, and hepatitis B alone. A subset of patients who have undergone transplantation (about 20%) will also have recurrence, Dr. Pawlik said.

National Comprehensive Cancer Network guidelines for follow-up of patients with HCC after resection or transplantation include imaging every 3-6 months for 2 years, then every 6-12 months thereafter, and testing of alpha-fetoprotein (AFP) level, if initially elevated, on the same schedule, he noted.

For example, a trial that included 18,816 hepatitis B carriers in China randomized to either intensive surveillance with AFP level and twice-yearly ultrasound versus no screening showed that the intensive surveillance was associated with a significantly lower rate ratio for mortality compared with no screening (RR 0.63) (J. Cancer Res. Clin. Oncol. 2004;130:417-22).

Similar evidence has been found to support intensive follow-up for colorectal cancers, for which there is effective therapy for recurrent disease, Dr. Pawlik noted.

Unlike pancreatic cancer or melanoma, where recurrences tend to be distant metastases, hepatocellular carcinoma is more frequently locally recurrent. It was found that 64%-80% of patients with cirrhosis had an intrahepatic recurrence within 5 years of HCC resection.

In one study, investigators found that 60%-70% of tumors recurred within 2 years, and 30%-40% of patients with recurrences had de novo tumors (J. Hepatol. 2003;38:200-7).

"For patients who have hepatocellular carcinoma, there are two reasons to [monitor] these patients closely. One is that they may truly have recurrence, and two is that a subset of these patients will develop new de novo disease in the cirrhotic liver," Dr. Pawlik said.

Treatment options for late recurrences are very limited and not very effective, he added. Sorafenib (Nexavar) is approved for unresectable HCC, but in the phase III SHARP trial was associated with an improvement in median overall survival of only 2.8 months vs. placebo (P = .00058) (N. Engl. J. Med. 2008;359:378-90).

In contrast, early recurrences generally respond to repeat resection, salvage transplantation, radiofrequency ablation, or transarterial chemoembolization, Dr. Pawlik noted.

Dr. Pawlik said he has no relevant disclosures.

ORLANDO – Intensive follow-up of patients who have undergone surgery for hepatocellular carcinoma reduces deaths from tumor recurrence and metastases, Dr. Timothy M. Pawlik said at a symposium sponsored by the Society of Surgical Oncology.

Data from surveillance programs for hepatocellular carcinoma (HCC) and empiric data from centers treating colorectal liver metastases (CRLM) suggest that HCC tends to recur locally, and that recurrent HCC and CRLM, if caught early, can be successfully controlled with a variety of therapeutic options, said Dr. Pawlik, associate professor of surgery and oncology, and hepatobiliary surgery program director at Johns Hopkins Medical Center, Baltimore.

"I would favor high-intensity surveillance for patients with hepatocellular carcinoma. This argument is based on level 1 data showing that high-intensity primary surveillance decreases mortality for patients who have cirrhosis," he said.

Although there is a high HCC recurrence rate following surgery, most recurrences will be contained within the liver, and may be successfully treated with salvage transplantation, ablation, or intra-arterial therapy.

"But if we miss that opportunity and patients recur with advanced disease, all those options are off the table, and their prognosis is abysmal," Dr. Pawlik said.

According to National Cancer Institute data, liver cancer holds the dubious distinction of being the fastest growing cancer in terms of death rate in the United States, outpacing lung cancer in women, esophageal cancer, and thyroid cancer.

Risk factors for HCC include cirrhosis from any cause (hepatitis B and C viruses, alcoholism, nonalcoholic fatty liver disease), hepatitis B primary infection (with or without cirrhosis), and inherited metabolic diseases such as hemochromatosis, alpha-1 antitrypsin deficiency, glycogen storage disease, or tyrosinemia.

"Most patients who have HCC don’t simply have a cancer; they also have significant underlying cirrhosis, so when thinking about the approach to HCC, we have to be thinking about all of the tumor-specific factors such as tumor size, location, and number, and we also have to be thinking about all of the liver-specific factors," Dr. Pawlik said.

Surgical treatment options include resection for HCCs of all sizes; transplantation, for single lesions 5 cm or smaller, or up to three lesions of 3 cm or less or advanced cirrhosis; and ablation for small lesions or inoperable or unresectable tumors as a bridge to transplant.

In various series, overall 5-year survival following resection ranges from 42% to 62%, and following transplantation from 57% to 75%.

But as Dr. Pawlik and colleagues noted in a 2008 study, disease-free survival following resection for HCC is only half of the percentage after transplantation (40% vs. 82%, P less than .01). (J. Gastrointest. Surg. 2008;12:1699-1708).

Intrahepatic recurrence is most frequent among patients with hepatitis C infection, but also occurs with hepatitis B and C coinfection, and hepatitis B alone. A subset of patients who have undergone transplantation (about 20%) will also have recurrence, Dr. Pawlik said.

National Comprehensive Cancer Network guidelines for follow-up of patients with HCC after resection or transplantation include imaging every 3-6 months for 2 years, then every 6-12 months thereafter, and testing of alpha-fetoprotein (AFP) level, if initially elevated, on the same schedule, he noted.

For example, a trial that included 18,816 hepatitis B carriers in China randomized to either intensive surveillance with AFP level and twice-yearly ultrasound versus no screening showed that the intensive surveillance was associated with a significantly lower rate ratio for mortality compared with no screening (RR 0.63) (J. Cancer Res. Clin. Oncol. 2004;130:417-22).

Similar evidence has been found to support intensive follow-up for colorectal cancers, for which there is effective therapy for recurrent disease, Dr. Pawlik noted.

Unlike pancreatic cancer or melanoma, where recurrences tend to be distant metastases, hepatocellular carcinoma is more frequently locally recurrent. It was found that 64%-80% of patients with cirrhosis had an intrahepatic recurrence within 5 years of HCC resection.

In one study, investigators found that 60%-70% of tumors recurred within 2 years, and 30%-40% of patients with recurrences had de novo tumors (J. Hepatol. 2003;38:200-7).

"For patients who have hepatocellular carcinoma, there are two reasons to [monitor] these patients closely. One is that they may truly have recurrence, and two is that a subset of these patients will develop new de novo disease in the cirrhotic liver," Dr. Pawlik said.

Treatment options for late recurrences are very limited and not very effective, he added. Sorafenib (Nexavar) is approved for unresectable HCC, but in the phase III SHARP trial was associated with an improvement in median overall survival of only 2.8 months vs. placebo (P = .00058) (N. Engl. J. Med. 2008;359:378-90).

In contrast, early recurrences generally respond to repeat resection, salvage transplantation, radiofrequency ablation, or transarterial chemoembolization, Dr. Pawlik noted.

Dr. Pawlik said he has no relevant disclosures.

ORLANDO – Intensive follow-up of patients who have undergone surgery for hepatocellular carcinoma reduces deaths from tumor recurrence and metastases, Dr. Timothy M. Pawlik said at a symposium sponsored by the Society of Surgical Oncology.

Data from surveillance programs for hepatocellular carcinoma (HCC) and empiric data from centers treating colorectal liver metastases (CRLM) suggest that HCC tends to recur locally, and that recurrent HCC and CRLM, if caught early, can be successfully controlled with a variety of therapeutic options, said Dr. Pawlik, associate professor of surgery and oncology, and hepatobiliary surgery program director at Johns Hopkins Medical Center, Baltimore.

"I would favor high-intensity surveillance for patients with hepatocellular carcinoma. This argument is based on level 1 data showing that high-intensity primary surveillance decreases mortality for patients who have cirrhosis," he said.

Although there is a high HCC recurrence rate following surgery, most recurrences will be contained within the liver, and may be successfully treated with salvage transplantation, ablation, or intra-arterial therapy.

"But if we miss that opportunity and patients recur with advanced disease, all those options are off the table, and their prognosis is abysmal," Dr. Pawlik said.

According to National Cancer Institute data, liver cancer holds the dubious distinction of being the fastest growing cancer in terms of death rate in the United States, outpacing lung cancer in women, esophageal cancer, and thyroid cancer.

Risk factors for HCC include cirrhosis from any cause (hepatitis B and C viruses, alcoholism, nonalcoholic fatty liver disease), hepatitis B primary infection (with or without cirrhosis), and inherited metabolic diseases such as hemochromatosis, alpha-1 antitrypsin deficiency, glycogen storage disease, or tyrosinemia.

"Most patients who have HCC don’t simply have a cancer; they also have significant underlying cirrhosis, so when thinking about the approach to HCC, we have to be thinking about all of the tumor-specific factors such as tumor size, location, and number, and we also have to be thinking about all of the liver-specific factors," Dr. Pawlik said.

Surgical treatment options include resection for HCCs of all sizes; transplantation, for single lesions 5 cm or smaller, or up to three lesions of 3 cm or less or advanced cirrhosis; and ablation for small lesions or inoperable or unresectable tumors as a bridge to transplant.

In various series, overall 5-year survival following resection ranges from 42% to 62%, and following transplantation from 57% to 75%.

But as Dr. Pawlik and colleagues noted in a 2008 study, disease-free survival following resection for HCC is only half of the percentage after transplantation (40% vs. 82%, P less than .01). (J. Gastrointest. Surg. 2008;12:1699-1708).

Intrahepatic recurrence is most frequent among patients with hepatitis C infection, but also occurs with hepatitis B and C coinfection, and hepatitis B alone. A subset of patients who have undergone transplantation (about 20%) will also have recurrence, Dr. Pawlik said.

National Comprehensive Cancer Network guidelines for follow-up of patients with HCC after resection or transplantation include imaging every 3-6 months for 2 years, then every 6-12 months thereafter, and testing of alpha-fetoprotein (AFP) level, if initially elevated, on the same schedule, he noted.

For example, a trial that included 18,816 hepatitis B carriers in China randomized to either intensive surveillance with AFP level and twice-yearly ultrasound versus no screening showed that the intensive surveillance was associated with a significantly lower rate ratio for mortality compared with no screening (RR 0.63) (J. Cancer Res. Clin. Oncol. 2004;130:417-22).

Similar evidence has been found to support intensive follow-up for colorectal cancers, for which there is effective therapy for recurrent disease, Dr. Pawlik noted.

Unlike pancreatic cancer or melanoma, where recurrences tend to be distant metastases, hepatocellular carcinoma is more frequently locally recurrent. It was found that 64%-80% of patients with cirrhosis had an intrahepatic recurrence within 5 years of HCC resection.

In one study, investigators found that 60%-70% of tumors recurred within 2 years, and 30%-40% of patients with recurrences had de novo tumors (J. Hepatol. 2003;38:200-7).

"For patients who have hepatocellular carcinoma, there are two reasons to [monitor] these patients closely. One is that they may truly have recurrence, and two is that a subset of these patients will develop new de novo disease in the cirrhotic liver," Dr. Pawlik said.

Treatment options for late recurrences are very limited and not very effective, he added. Sorafenib (Nexavar) is approved for unresectable HCC, but in the phase III SHARP trial was associated with an improvement in median overall survival of only 2.8 months vs. placebo (P = .00058) (N. Engl. J. Med. 2008;359:378-90).

In contrast, early recurrences generally respond to repeat resection, salvage transplantation, radiofrequency ablation, or transarterial chemoembolization, Dr. Pawlik noted.

Dr. Pawlik said he has no relevant disclosures.

ORLANDO – Surgery for ductal carcinoma in situ, with or without adjuvant therapy, saves lives, asserted a breast cancer surgeon at a symposium sponsored by the Society of Surgical Oncology.

Following a surgical biopsy alone, about half of all cases of low-grade ductal carcinoma in situ (DCIS) will progress to invasive cancer within an average of 10-15 years, said Dr. Kimberly J. Van Zee, a surgical oncologist at Memorial Sloan-Kettering Cancer Center in New York.

Additionally, without intervention, low-grade DCIS will result in death from ipsilateral invasive recurrence of breast cancer in about 18% of patients, Dr. Van Zee said.

"With treatment of DCIS, whether it’s breast conservation or mastectomy, with or without radiation, breast cancer–specific survival is over 95%," she noted.

The incidence of DCIS has increased steadily since 1975, when the rate was slightly more than 5 in 100,000 women. In 2009, the rate had reached approximately 36 in 100,000, according to Surveillance, Epidemiology, and End Results (SEER) data. The increase is probably a result of the growing adoption of screening mammography over the same period, Dr. Van Zee commented.

Treatment trends for DCIS showed a gradual but steady decline in mastectomy – from 70% in 1983 to about 28% in 1999 – and a corresponding increase in breast-conserving treatment, which increased from about 25% to 68% over the same period.

Beginning around 2005, however, there was evidence that the trend was reversing, with upticks in both mastectomy for unilateral breast cancer (J. Clin. Oncol. 2010;28:3437-41) and contralateral prophylactic mastectomy, both among women with invasive cancers and DCIS (Ann. Surg. Oncol. 2010;17:2554-62). The trends paralleled the rise in screening mammography in the United States and elsewhere in the world.

The gradual but steady decline in breast cancer deaths that began in the early 1990s appears to be attributable to a combination of increased screening mammography and improvements in adjuvant therapy, Dr. Van Zee noted, citing a 2005 study (N. Engl. J. Med. 2005;353:1784-92).

"They dissected all the various effects of treatment, incidence of screening-detected diseases, etc., and all their analyses concluded that about half of the reduction in death rate was due to screening and the other half was due to adjuvant therapy. So I think this is good circumstantial evidence that screening, with its resultant increased incidence in DCIS and the resulting increased treatment of DCIS, does result in a lower death rate from breast cancer," she said.

Further evidence comes from studies in which pathologists reviewed thousands of slides of biopsy-acquired breast tissue originally reported as benign. In each study (Cancer 1980;46[4 Suppl]:919-25; Cancer 2005;103:2481-84), the investigator identified about 30 samples with evidence of low-grade, relatively low-volume DCIS that was not recognized or treated. After 20-30 years of follow-up, half of the women had developed a clinically apparent ipsilateral breast cancer recurrence. The majority of tumors were invasive. In the second study, the authors noted that 5 of the 28 women (18%) with previously undetected DCIS died of breast cancer.

Evidence from a meta-analysis (Cancer 1999;85:616-28) suggests that the risk for invasive recurrence following a mastectomy for DCIS is 1.1%, and that the risk for breast cancer death is less than 1.1%.

The risk for distant recurrence and/or death from breast-conserving surgery with or without adjuvant radiotherapy in prospective randomized trials of radiotherapy for DCIS was less than 5%. Among patients with invasive local failure in those trials, however, 18%-25% developed metastatic disease, indicating the importance of avoiding local recurrence.

Mastectomy and breast-conserving surgery combined with radiotherapy and/or endocrine therapy all provide excellent disease-specific and overall survival results, Dr. Van Zee said.

"The goal should be avoiding local recurrence and, in particular, invasive recurrence, minimizing morbidity, and perhaps individualizing the treatment to the disease. One could consider age, comorbidities, [and] life expectancy, and weigh those against the morbidity of the treatment and the risk of local recurrence," she said.

Dr. Van Zee reported no relevant financial disclosures.

ORLANDO – Surgery for ductal carcinoma in situ, with or without adjuvant therapy, saves lives, asserted a breast cancer surgeon at a symposium sponsored by the Society of Surgical Oncology.

Following a surgical biopsy alone, about half of all cases of low-grade ductal carcinoma in situ (DCIS) will progress to invasive cancer within an average of 10-15 years, said Dr. Kimberly J. Van Zee, a surgical oncologist at Memorial Sloan-Kettering Cancer Center in New York.

Additionally, without intervention, low-grade DCIS will result in death from ipsilateral invasive recurrence of breast cancer in about 18% of patients, Dr. Van Zee said.

"With treatment of DCIS, whether it’s breast conservation or mastectomy, with or without radiation, breast cancer–specific survival is over 95%," she noted.

The incidence of DCIS has increased steadily since 1975, when the rate was slightly more than 5 in 100,000 women. In 2009, the rate had reached approximately 36 in 100,000, according to Surveillance, Epidemiology, and End Results (SEER) data. The increase is probably a result of the growing adoption of screening mammography over the same period, Dr. Van Zee commented.

Treatment trends for DCIS showed a gradual but steady decline in mastectomy – from 70% in 1983 to about 28% in 1999 – and a corresponding increase in breast-conserving treatment, which increased from about 25% to 68% over the same period.

Beginning around 2005, however, there was evidence that the trend was reversing, with upticks in both mastectomy for unilateral breast cancer (J. Clin. Oncol. 2010;28:3437-41) and contralateral prophylactic mastectomy, both among women with invasive cancers and DCIS (Ann. Surg. Oncol. 2010;17:2554-62). The trends paralleled the rise in screening mammography in the United States and elsewhere in the world.

The gradual but steady decline in breast cancer deaths that began in the early 1990s appears to be attributable to a combination of increased screening mammography and improvements in adjuvant therapy, Dr. Van Zee noted, citing a 2005 study (N. Engl. J. Med. 2005;353:1784-92).

"They dissected all the various effects of treatment, incidence of screening-detected diseases, etc., and all their analyses concluded that about half of the reduction in death rate was due to screening and the other half was due to adjuvant therapy. So I think this is good circumstantial evidence that screening, with its resultant increased incidence in DCIS and the resulting increased treatment of DCIS, does result in a lower death rate from breast cancer," she said.

Further evidence comes from studies in which pathologists reviewed thousands of slides of biopsy-acquired breast tissue originally reported as benign. In each study (Cancer 1980;46[4 Suppl]:919-25; Cancer 2005;103:2481-84), the investigator identified about 30 samples with evidence of low-grade, relatively low-volume DCIS that was not recognized or treated. After 20-30 years of follow-up, half of the women had developed a clinically apparent ipsilateral breast cancer recurrence. The majority of tumors were invasive. In the second study, the authors noted that 5 of the 28 women (18%) with previously undetected DCIS died of breast cancer.

Evidence from a meta-analysis (Cancer 1999;85:616-28) suggests that the risk for invasive recurrence following a mastectomy for DCIS is 1.1%, and that the risk for breast cancer death is less than 1.1%.

The risk for distant recurrence and/or death from breast-conserving surgery with or without adjuvant radiotherapy in prospective randomized trials of radiotherapy for DCIS was less than 5%. Among patients with invasive local failure in those trials, however, 18%-25% developed metastatic disease, indicating the importance of avoiding local recurrence.

Mastectomy and breast-conserving surgery combined with radiotherapy and/or endocrine therapy all provide excellent disease-specific and overall survival results, Dr. Van Zee said.

"The goal should be avoiding local recurrence and, in particular, invasive recurrence, minimizing morbidity, and perhaps individualizing the treatment to the disease. One could consider age, comorbidities, [and] life expectancy, and weigh those against the morbidity of the treatment and the risk of local recurrence," she said.

Dr. Van Zee reported no relevant financial disclosures.

ORLANDO – Surgery for ductal carcinoma in situ, with or without adjuvant therapy, saves lives, asserted a breast cancer surgeon at a symposium sponsored by the Society of Surgical Oncology.

Following a surgical biopsy alone, about half of all cases of low-grade ductal carcinoma in situ (DCIS) will progress to invasive cancer within an average of 10-15 years, said Dr. Kimberly J. Van Zee, a surgical oncologist at Memorial Sloan-Kettering Cancer Center in New York.

Additionally, without intervention, low-grade DCIS will result in death from ipsilateral invasive recurrence of breast cancer in about 18% of patients, Dr. Van Zee said.

"With treatment of DCIS, whether it’s breast conservation or mastectomy, with or without radiation, breast cancer–specific survival is over 95%," she noted.

The incidence of DCIS has increased steadily since 1975, when the rate was slightly more than 5 in 100,000 women. In 2009, the rate had reached approximately 36 in 100,000, according to Surveillance, Epidemiology, and End Results (SEER) data. The increase is probably a result of the growing adoption of screening mammography over the same period, Dr. Van Zee commented.

Treatment trends for DCIS showed a gradual but steady decline in mastectomy – from 70% in 1983 to about 28% in 1999 – and a corresponding increase in breast-conserving treatment, which increased from about 25% to 68% over the same period.

Beginning around 2005, however, there was evidence that the trend was reversing, with upticks in both mastectomy for unilateral breast cancer (J. Clin. Oncol. 2010;28:3437-41) and contralateral prophylactic mastectomy, both among women with invasive cancers and DCIS (Ann. Surg. Oncol. 2010;17:2554-62). The trends paralleled the rise in screening mammography in the United States and elsewhere in the world.

The gradual but steady decline in breast cancer deaths that began in the early 1990s appears to be attributable to a combination of increased screening mammography and improvements in adjuvant therapy, Dr. Van Zee noted, citing a 2005 study (N. Engl. J. Med. 2005;353:1784-92).

"They dissected all the various effects of treatment, incidence of screening-detected diseases, etc., and all their analyses concluded that about half of the reduction in death rate was due to screening and the other half was due to adjuvant therapy. So I think this is good circumstantial evidence that screening, with its resultant increased incidence in DCIS and the resulting increased treatment of DCIS, does result in a lower death rate from breast cancer," she said.

Further evidence comes from studies in which pathologists reviewed thousands of slides of biopsy-acquired breast tissue originally reported as benign. In each study (Cancer 1980;46[4 Suppl]:919-25; Cancer 2005;103:2481-84), the investigator identified about 30 samples with evidence of low-grade, relatively low-volume DCIS that was not recognized or treated. After 20-30 years of follow-up, half of the women had developed a clinically apparent ipsilateral breast cancer recurrence. The majority of tumors were invasive. In the second study, the authors noted that 5 of the 28 women (18%) with previously undetected DCIS died of breast cancer.

Evidence from a meta-analysis (Cancer 1999;85:616-28) suggests that the risk for invasive recurrence following a mastectomy for DCIS is 1.1%, and that the risk for breast cancer death is less than 1.1%.

The risk for distant recurrence and/or death from breast-conserving surgery with or without adjuvant radiotherapy in prospective randomized trials of radiotherapy for DCIS was less than 5%. Among patients with invasive local failure in those trials, however, 18%-25% developed metastatic disease, indicating the importance of avoiding local recurrence.

Mastectomy and breast-conserving surgery combined with radiotherapy and/or endocrine therapy all provide excellent disease-specific and overall survival results, Dr. Van Zee said.

"The goal should be avoiding local recurrence and, in particular, invasive recurrence, minimizing morbidity, and perhaps individualizing the treatment to the disease. One could consider age, comorbidities, [and] life expectancy, and weigh those against the morbidity of the treatment and the risk of local recurrence," she said.

Dr. Van Zee reported no relevant financial disclosures.

ORLANDO – Gastrointestinal stromal tumors of the duodenum can be managed safely with surgery, yielding good overall survival, an investigator reported at a symposium sponsored by the Society of Surgical Oncology.

In addition to surgery for duodenal gastrointestinal stromal tumors (GISTs), adjuvant treatment with imatinib (Gleevec) may reduce the risk of recurrence, and neoadjuvant imatinib may improve surgical outcomes, said Dr. Chiara Colombo, from the sarcoma service of the Fondazione Istituto Nazionale dei Tumori in Milan.

"Imatinib in the neoadjuvant setting might facilitate the surgical resection and possibly help preserve the biliary and pancreatic anatomy," she said.

Only about 5% of GISTs originate in the duodenum, but tumors occurring in this section of the small intestine are anatomically challenging and there is considerable debate about the most appropriate surgical and therapeutic approaches, she added.

To gain a better understanding of the problem, Dr. Colombo and colleagues at her center, as well as the University of Mannheim (Germany), Massachusetts General Hospital in Boston, and the Maria Sklodowska-Curie Memorial Cancer Centre and Institute of Oncology in Gliwice (Poland), pooled data on patients with a primary duodenal GIST treated from February 2000 through August 2011 at their institutions.

There were a total of 39 men and 45 women, median age 58 years, treated with either pancreaticoduodenectomy (28 patients, 33%) or limited resection (56 patients, 67%). In all, 11 patients received preoperative imatinib for a median of 8 months, and 23 received it after surgery for a median of 17 months.

Surgical complication rates were higher among patients who underwent pancreaticoduodenectomy, compared with patients who had limited resection. The complications occurred in 5 patients who underwent limited resection (9%) and 10 who underwent pancreaticoduodenectomy (36%) and included evisceration/wound dehiscence (2 and 1, respectively), infections (1 in each group), gastric obstruction/delayed emptying (0 and 2), and pancreatic/biliary leakage (2 and 6 patients).

Among all patients, overall survival was 98% at 3 years and 89% at 5 years; the median follow-up was 42 months. Overall survival of patients at low or intermediate risk for recurrence was similar whether they received imatinib or did not. Among high-risk patients, however, imatinib was associated with significantly better 5-year overall survival, and the survival curves began to converge after imatinib was withdrawn, Dr. Colombo said.

She noted that the results echo those of a recent phase III study of adjuvant imatinib in patients with operable GISTs (ASCO 2011 Abstract LBA1).

In the intention-to-treat population of the randomized trial, the hazard ratio for overall survival for 36 months vs. 12 months of adjuvant imatinib was 0.45 (P = .019). The longer treatment schedule was also associated with better recurrence-free survival (hazard ratio 0.46, P less than .0001).

In the neoadjuvant setting, imatinib was associated with tumor regression, allowing for safer resection than might otherwise be possible, Dr. Colombo commented. This analysis on tumor regression was based on the combined clinical experience at all the centers.

The study was supported by the participating institutions. Dr. Colombo reported no relevant financial disclosures.

ORLANDO – Gastrointestinal stromal tumors of the duodenum can be managed safely with surgery, yielding good overall survival, an investigator reported at a symposium sponsored by the Society of Surgical Oncology.

In addition to surgery for duodenal gastrointestinal stromal tumors (GISTs), adjuvant treatment with imatinib (Gleevec) may reduce the risk of recurrence, and neoadjuvant imatinib may improve surgical outcomes, said Dr. Chiara Colombo, from the sarcoma service of the Fondazione Istituto Nazionale dei Tumori in Milan.

"Imatinib in the neoadjuvant setting might facilitate the surgical resection and possibly help preserve the biliary and pancreatic anatomy," she said.

Only about 5% of GISTs originate in the duodenum, but tumors occurring in this section of the small intestine are anatomically challenging and there is considerable debate about the most appropriate surgical and therapeutic approaches, she added.

To gain a better understanding of the problem, Dr. Colombo and colleagues at her center, as well as the University of Mannheim (Germany), Massachusetts General Hospital in Boston, and the Maria Sklodowska-Curie Memorial Cancer Centre and Institute of Oncology in Gliwice (Poland), pooled data on patients with a primary duodenal GIST treated from February 2000 through August 2011 at their institutions.

There were a total of 39 men and 45 women, median age 58 years, treated with either pancreaticoduodenectomy (28 patients, 33%) or limited resection (56 patients, 67%). In all, 11 patients received preoperative imatinib for a median of 8 months, and 23 received it after surgery for a median of 17 months.

Surgical complication rates were higher among patients who underwent pancreaticoduodenectomy, compared with patients who had limited resection. The complications occurred in 5 patients who underwent limited resection (9%) and 10 who underwent pancreaticoduodenectomy (36%) and included evisceration/wound dehiscence (2 and 1, respectively), infections (1 in each group), gastric obstruction/delayed emptying (0 and 2), and pancreatic/biliary leakage (2 and 6 patients).

Among all patients, overall survival was 98% at 3 years and 89% at 5 years; the median follow-up was 42 months. Overall survival of patients at low or intermediate risk for recurrence was similar whether they received imatinib or did not. Among high-risk patients, however, imatinib was associated with significantly better 5-year overall survival, and the survival curves began to converge after imatinib was withdrawn, Dr. Colombo said.

She noted that the results echo those of a recent phase III study of adjuvant imatinib in patients with operable GISTs (ASCO 2011 Abstract LBA1).

In the intention-to-treat population of the randomized trial, the hazard ratio for overall survival for 36 months vs. 12 months of adjuvant imatinib was 0.45 (P = .019). The longer treatment schedule was also associated with better recurrence-free survival (hazard ratio 0.46, P less than .0001).

In the neoadjuvant setting, imatinib was associated with tumor regression, allowing for safer resection than might otherwise be possible, Dr. Colombo commented. This analysis on tumor regression was based on the combined clinical experience at all the centers.

The study was supported by the participating institutions. Dr. Colombo reported no relevant financial disclosures.

ORLANDO – Gastrointestinal stromal tumors of the duodenum can be managed safely with surgery, yielding good overall survival, an investigator reported at a symposium sponsored by the Society of Surgical Oncology.

In addition to surgery for duodenal gastrointestinal stromal tumors (GISTs), adjuvant treatment with imatinib (Gleevec) may reduce the risk of recurrence, and neoadjuvant imatinib may improve surgical outcomes, said Dr. Chiara Colombo, from the sarcoma service of the Fondazione Istituto Nazionale dei Tumori in Milan.

"Imatinib in the neoadjuvant setting might facilitate the surgical resection and possibly help preserve the biliary and pancreatic anatomy," she said.

Only about 5% of GISTs originate in the duodenum, but tumors occurring in this section of the small intestine are anatomically challenging and there is considerable debate about the most appropriate surgical and therapeutic approaches, she added.

To gain a better understanding of the problem, Dr. Colombo and colleagues at her center, as well as the University of Mannheim (Germany), Massachusetts General Hospital in Boston, and the Maria Sklodowska-Curie Memorial Cancer Centre and Institute of Oncology in Gliwice (Poland), pooled data on patients with a primary duodenal GIST treated from February 2000 through August 2011 at their institutions.

There were a total of 39 men and 45 women, median age 58 years, treated with either pancreaticoduodenectomy (28 patients, 33%) or limited resection (56 patients, 67%). In all, 11 patients received preoperative imatinib for a median of 8 months, and 23 received it after surgery for a median of 17 months.

Surgical complication rates were higher among patients who underwent pancreaticoduodenectomy, compared with patients who had limited resection. The complications occurred in 5 patients who underwent limited resection (9%) and 10 who underwent pancreaticoduodenectomy (36%) and included evisceration/wound dehiscence (2 and 1, respectively), infections (1 in each group), gastric obstruction/delayed emptying (0 and 2), and pancreatic/biliary leakage (2 and 6 patients).

Among all patients, overall survival was 98% at 3 years and 89% at 5 years; the median follow-up was 42 months. Overall survival of patients at low or intermediate risk for recurrence was similar whether they received imatinib or did not. Among high-risk patients, however, imatinib was associated with significantly better 5-year overall survival, and the survival curves began to converge after imatinib was withdrawn, Dr. Colombo said.

She noted that the results echo those of a recent phase III study of adjuvant imatinib in patients with operable GISTs (ASCO 2011 Abstract LBA1).

In the intention-to-treat population of the randomized trial, the hazard ratio for overall survival for 36 months vs. 12 months of adjuvant imatinib was 0.45 (P = .019). The longer treatment schedule was also associated with better recurrence-free survival (hazard ratio 0.46, P less than .0001).

In the neoadjuvant setting, imatinib was associated with tumor regression, allowing for safer resection than might otherwise be possible, Dr. Colombo commented. This analysis on tumor regression was based on the combined clinical experience at all the centers.

The study was supported by the participating institutions. Dr. Colombo reported no relevant financial disclosures.

ORLANDO – Patients who undergo surgical resection with narrow distal margins for ultra low rectal cancers can have local control of disease and overall survival comparable to those in patients with wider resection margins, investigators reported at a symposium of the Society of Surgical Oncology.

Dr. Wim P. Ceelen and coauthors reviewed 109 patients who underwent neoadjuvant chemoradiation and surgery for rectal cancers 5 cm or less from the anal verge from 1998 to 2010. Overall survival and local control were found to be the same in a comparison of patients with distal resection margins 1 cm wide or less and those with margins greater than 1 cm.

"Close but free distal margins after neoadjuvant radiation and sphincter-saving surgery do not compromise local or systemic control," said Dr. Ceelen, a surgical oncologist at University Hospital in Ghent, Belgium.

Neoadjuvant chemoradiation has made sphincter-sparing procedures available to more patients, but the surgical results often leave distal margins that some surgeons find too close for comfort, because they fall short of the so-called "1-cm rule" holding that margins should be a minimum of 1 cm from the tumor.

"There is some controversy about the oncological safety of having very close margins, i.e., less than 1 cm," Dr. Ceelen said.

To see whether margins matter, he and colleagues reviewed records in a prospective database on the 109 patients with ultra-low rectal cancers (median pretreatment distance of 3 cm from the anal verge). All patients underwent neoadjuvant chemoradiation with 5-fluorouracil and 45 Gy radiation delivered in 25 fractions, followed by sphincter-saving surgery with R0 (cancer-free) resections. In all, 59% of the patients were clinically node positive.

Surgeries included intersphincteric resection and coloanal anastomosis in 38 patients, and stapled anastamosis in 71. The large majority of patients (98) had protective loop ileostomies.

Surgical specimens showed tumors were stage 0 (pathologic complete response to chemoradiation) in 16%, stage 1 (T1-T2) in 30%, stage II (T3-T4) in 21%, and stage III (node-positive) in 19%. The median distal resection margin was 10 mm.

At a median follow-up of 33 months, 5 patients (4.6%) had a local recurrence. Two of these patients had had surgical margins of 1 cm or less, and 3 had margin greater than 1 cm. Of the 30 patients (27.5%) who had a systemic recurrence, 12 had resections with margins of 1 cm or less, and 18 had margins greater than 1 cm. Neither difference was statistically significant.

Overall 5-year survival was 70%, and again there was no significant difference in survival by margin size.

Dr. Ceelen noted that the results echo those of a recently published systematic review (Ann. Surg. Oncol. 2012;19: 801-8).

The study was internally funded. Dr. Ceelen had no disclosures.

ORLANDO – Patients who undergo surgical resection with narrow distal margins for ultra low rectal cancers can have local control of disease and overall survival comparable to those in patients with wider resection margins, investigators reported at a symposium of the Society of Surgical Oncology.

Dr. Wim P. Ceelen and coauthors reviewed 109 patients who underwent neoadjuvant chemoradiation and surgery for rectal cancers 5 cm or less from the anal verge from 1998 to 2010. Overall survival and local control were found to be the same in a comparison of patients with distal resection margins 1 cm wide or less and those with margins greater than 1 cm.

"Close but free distal margins after neoadjuvant radiation and sphincter-saving surgery do not compromise local or systemic control," said Dr. Ceelen, a surgical oncologist at University Hospital in Ghent, Belgium.

Neoadjuvant chemoradiation has made sphincter-sparing procedures available to more patients, but the surgical results often leave distal margins that some surgeons find too close for comfort, because they fall short of the so-called "1-cm rule" holding that margins should be a minimum of 1 cm from the tumor.

"There is some controversy about the oncological safety of having very close margins, i.e., less than 1 cm," Dr. Ceelen said.

To see whether margins matter, he and colleagues reviewed records in a prospective database on the 109 patients with ultra-low rectal cancers (median pretreatment distance of 3 cm from the anal verge). All patients underwent neoadjuvant chemoradiation with 5-fluorouracil and 45 Gy radiation delivered in 25 fractions, followed by sphincter-saving surgery with R0 (cancer-free) resections. In all, 59% of the patients were clinically node positive.

Surgeries included intersphincteric resection and coloanal anastomosis in 38 patients, and stapled anastamosis in 71. The large majority of patients (98) had protective loop ileostomies.

Surgical specimens showed tumors were stage 0 (pathologic complete response to chemoradiation) in 16%, stage 1 (T1-T2) in 30%, stage II (T3-T4) in 21%, and stage III (node-positive) in 19%. The median distal resection margin was 10 mm.

At a median follow-up of 33 months, 5 patients (4.6%) had a local recurrence. Two of these patients had had surgical margins of 1 cm or less, and 3 had margin greater than 1 cm. Of the 30 patients (27.5%) who had a systemic recurrence, 12 had resections with margins of 1 cm or less, and 18 had margins greater than 1 cm. Neither difference was statistically significant.

Overall 5-year survival was 70%, and again there was no significant difference in survival by margin size.

Dr. Ceelen noted that the results echo those of a recently published systematic review (Ann. Surg. Oncol. 2012;19: 801-8).

The study was internally funded. Dr. Ceelen had no disclosures.

ORLANDO – Patients who undergo surgical resection with narrow distal margins for ultra low rectal cancers can have local control of disease and overall survival comparable to those in patients with wider resection margins, investigators reported at a symposium of the Society of Surgical Oncology.

Dr. Wim P. Ceelen and coauthors reviewed 109 patients who underwent neoadjuvant chemoradiation and surgery for rectal cancers 5 cm or less from the anal verge from 1998 to 2010. Overall survival and local control were found to be the same in a comparison of patients with distal resection margins 1 cm wide or less and those with margins greater than 1 cm.

"Close but free distal margins after neoadjuvant radiation and sphincter-saving surgery do not compromise local or systemic control," said Dr. Ceelen, a surgical oncologist at University Hospital in Ghent, Belgium.

Neoadjuvant chemoradiation has made sphincter-sparing procedures available to more patients, but the surgical results often leave distal margins that some surgeons find too close for comfort, because they fall short of the so-called "1-cm rule" holding that margins should be a minimum of 1 cm from the tumor.

"There is some controversy about the oncological safety of having very close margins, i.e., less than 1 cm," Dr. Ceelen said.

To see whether margins matter, he and colleagues reviewed records in a prospective database on the 109 patients with ultra-low rectal cancers (median pretreatment distance of 3 cm from the anal verge). All patients underwent neoadjuvant chemoradiation with 5-fluorouracil and 45 Gy radiation delivered in 25 fractions, followed by sphincter-saving surgery with R0 (cancer-free) resections. In all, 59% of the patients were clinically node positive.

Surgeries included intersphincteric resection and coloanal anastomosis in 38 patients, and stapled anastamosis in 71. The large majority of patients (98) had protective loop ileostomies.

Surgical specimens showed tumors were stage 0 (pathologic complete response to chemoradiation) in 16%, stage 1 (T1-T2) in 30%, stage II (T3-T4) in 21%, and stage III (node-positive) in 19%. The median distal resection margin was 10 mm.

At a median follow-up of 33 months, 5 patients (4.6%) had a local recurrence. Two of these patients had had surgical margins of 1 cm or less, and 3 had margin greater than 1 cm. Of the 30 patients (27.5%) who had a systemic recurrence, 12 had resections with margins of 1 cm or less, and 18 had margins greater than 1 cm. Neither difference was statistically significant.

Overall 5-year survival was 70%, and again there was no significant difference in survival by margin size.

Dr. Ceelen noted that the results echo those of a recently published systematic review (Ann. Surg. Oncol. 2012;19: 801-8).

The study was internally funded. Dr. Ceelen had no disclosures.

ORLANDO – Patients with stage IA to IIA gastric adenocarcinoma are frequently managed with surgery alone, but a subset of these patients may benefit from adjuvant chemotherapy and/or radiation, suggested Boston-based investigators at a symposium sponsored by the Society of Surgical Oncology.

A review of Surveillance Epidemiology and End Results (SEER) registry data on 8,515 patients treated for gastric adenocarcinoma found that stages, age, tumor differentiation status, tumor size, and location are significantly associated with worse disease-specific survival (DSS) among patients with earlier diseases, reported Dr. Jason S. Gold, chief of surgical oncology at the VA Boston Healthcare System.

"Patients with at least two of these adverse features had a 5-year disease-specific survival of 76% or less, and perhaps these patients would be benefited by adjuvant treatment," said Dr. Gold.

Although the benefit of adjuvant therapy for stage IIB to IIIC gastric adenocarcinomas has been well documented in randomized trials, there are patients with less advanced-stage disease who have a poor prognosis and might also benefit from adjuvant therapy, Dr. Gold said.

Consensus guideline recommendations on the use of adjuvant therapy vary, with U.S. (National Comprehensive Cancer Network) and Canadian (Cancer Care Ontario) guidelines recommending adjuvant therapy for all patients except those with T1 tumors with no lymph node involvement. In contrast, European guidelines (Norwegian and Dutch) generally recommend adjuvant treatment for cancers with serosal invasion or nodal positivity, but not for patients with T1 tumors and 0-2 involved nodes, Dr. Gold said.

To tease out the natural history of stage IA- to II cancers and identify predictive factors for worse outcomes, he and colleagues combed through SEER data to identify patients with local or local-regional gastric adenocarcinoma who underwent surgery and pathologic evaluation of at least 15 lymph nodes, and who also had disease-specific survival data. They found that 2,431 patients had stages IA - IIA disease.

Not surprisingly, 887 patients with stage IA disease (T1N0) had the best odds, with a 5-year DSS of 91%. The investigators determined that these patients would be unlikely to benefit from adjuvant therapy, and excluded them from further analyses.

Among patients with stages IB through IIA disease, 5-year DSS rates ranged from 81% for patients with T1N2 and T2N0 disease, to 66% for patients with T1N1, T2N1, and T3NO disease.

In univariate analysis, factors associated with worse outcomes were older age (P less than .001), higher grade disease (P = .03), larger tumor size (P less than .001), proximal vs. distal location (P less than .001), T stage (P = .006), and TN grouping (P = .004).

In multivariate analysis, variables associated with worse outcomes were age (relative risk 1.025, P = .001), moderately or poorly differentiated or undifferentiated vs. well differentiated cancers (RR, 2.160, 3.323, and 3.306, respectively, P = .004 for all), size (RR, 1.027, P = .001) and location relative to the antrum/pylorus (gastric body RR, 1.289, and cardia/fundus RR, 2.508, P = .001 for both comparisons).

However, neither T stage, N stage nor TN grouping were independent predictors of outcome, the investigators found.

Based on these variables, they devised a risk score for stages IB-IIA, with each of the following four factors receiving 1 point: age greater than 60, tumor size greater than 5 cm, proximal location (cardia or fundus), and histologic grade other than well differentiated.

Under this risk classification system, they saw that 5-year DSS with no risk factors (two patients) was 100%, compared with 86% for patients with one risk factor (92 patients), 76% for those with two risk factors (325), 72% for those with three (372 patients), and 48% for those with all four (136 patients, P less than .001).

The study was internally funded. Dr. Gold reported having no relevant disclosures.

ORLANDO – Patients with stage IA to IIA gastric adenocarcinoma are frequently managed with surgery alone, but a subset of these patients may benefit from adjuvant chemotherapy and/or radiation, suggested Boston-based investigators at a symposium sponsored by the Society of Surgical Oncology.

A review of Surveillance Epidemiology and End Results (SEER) registry data on 8,515 patients treated for gastric adenocarcinoma found that stages, age, tumor differentiation status, tumor size, and location are significantly associated with worse disease-specific survival (DSS) among patients with earlier diseases, reported Dr. Jason S. Gold, chief of surgical oncology at the VA Boston Healthcare System.

"Patients with at least two of these adverse features had a 5-year disease-specific survival of 76% or less, and perhaps these patients would be benefited by adjuvant treatment," said Dr. Gold.

Although the benefit of adjuvant therapy for stage IIB to IIIC gastric adenocarcinomas has been well documented in randomized trials, there are patients with less advanced-stage disease who have a poor prognosis and might also benefit from adjuvant therapy, Dr. Gold said.

Consensus guideline recommendations on the use of adjuvant therapy vary, with U.S. (National Comprehensive Cancer Network) and Canadian (Cancer Care Ontario) guidelines recommending adjuvant therapy for all patients except those with T1 tumors with no lymph node involvement. In contrast, European guidelines (Norwegian and Dutch) generally recommend adjuvant treatment for cancers with serosal invasion or nodal positivity, but not for patients with T1 tumors and 0-2 involved nodes, Dr. Gold said.

To tease out the natural history of stage IA- to II cancers and identify predictive factors for worse outcomes, he and colleagues combed through SEER data to identify patients with local or local-regional gastric adenocarcinoma who underwent surgery and pathologic evaluation of at least 15 lymph nodes, and who also had disease-specific survival data. They found that 2,431 patients had stages IA - IIA disease.

Not surprisingly, 887 patients with stage IA disease (T1N0) had the best odds, with a 5-year DSS of 91%. The investigators determined that these patients would be unlikely to benefit from adjuvant therapy, and excluded them from further analyses.

Among patients with stages IB through IIA disease, 5-year DSS rates ranged from 81% for patients with T1N2 and T2N0 disease, to 66% for patients with T1N1, T2N1, and T3NO disease.

In univariate analysis, factors associated with worse outcomes were older age (P less than .001), higher grade disease (P = .03), larger tumor size (P less than .001), proximal vs. distal location (P less than .001), T stage (P = .006), and TN grouping (P = .004).

In multivariate analysis, variables associated with worse outcomes were age (relative risk 1.025, P = .001), moderately or poorly differentiated or undifferentiated vs. well differentiated cancers (RR, 2.160, 3.323, and 3.306, respectively, P = .004 for all), size (RR, 1.027, P = .001) and location relative to the antrum/pylorus (gastric body RR, 1.289, and cardia/fundus RR, 2.508, P = .001 for both comparisons).

However, neither T stage, N stage nor TN grouping were independent predictors of outcome, the investigators found.

Based on these variables, they devised a risk score for stages IB-IIA, with each of the following four factors receiving 1 point: age greater than 60, tumor size greater than 5 cm, proximal location (cardia or fundus), and histologic grade other than well differentiated.

Under this risk classification system, they saw that 5-year DSS with no risk factors (two patients) was 100%, compared with 86% for patients with one risk factor (92 patients), 76% for those with two risk factors (325), 72% for those with three (372 patients), and 48% for those with all four (136 patients, P less than .001).

The study was internally funded. Dr. Gold reported having no relevant disclosures.

ORLANDO – Patients with stage IA to IIA gastric adenocarcinoma are frequently managed with surgery alone, but a subset of these patients may benefit from adjuvant chemotherapy and/or radiation, suggested Boston-based investigators at a symposium sponsored by the Society of Surgical Oncology.

A review of Surveillance Epidemiology and End Results (SEER) registry data on 8,515 patients treated for gastric adenocarcinoma found that stages, age, tumor differentiation status, tumor size, and location are significantly associated with worse disease-specific survival (DSS) among patients with earlier diseases, reported Dr. Jason S. Gold, chief of surgical oncology at the VA Boston Healthcare System.

"Patients with at least two of these adverse features had a 5-year disease-specific survival of 76% or less, and perhaps these patients would be benefited by adjuvant treatment," said Dr. Gold.

Although the benefit of adjuvant therapy for stage IIB to IIIC gastric adenocarcinomas has been well documented in randomized trials, there are patients with less advanced-stage disease who have a poor prognosis and might also benefit from adjuvant therapy, Dr. Gold said.

Consensus guideline recommendations on the use of adjuvant therapy vary, with U.S. (National Comprehensive Cancer Network) and Canadian (Cancer Care Ontario) guidelines recommending adjuvant therapy for all patients except those with T1 tumors with no lymph node involvement. In contrast, European guidelines (Norwegian and Dutch) generally recommend adjuvant treatment for cancers with serosal invasion or nodal positivity, but not for patients with T1 tumors and 0-2 involved nodes, Dr. Gold said.

To tease out the natural history of stage IA- to II cancers and identify predictive factors for worse outcomes, he and colleagues combed through SEER data to identify patients with local or local-regional gastric adenocarcinoma who underwent surgery and pathologic evaluation of at least 15 lymph nodes, and who also had disease-specific survival data. They found that 2,431 patients had stages IA - IIA disease.

Not surprisingly, 887 patients with stage IA disease (T1N0) had the best odds, with a 5-year DSS of 91%. The investigators determined that these patients would be unlikely to benefit from adjuvant therapy, and excluded them from further analyses.

Among patients with stages IB through IIA disease, 5-year DSS rates ranged from 81% for patients with T1N2 and T2N0 disease, to 66% for patients with T1N1, T2N1, and T3NO disease.

In univariate analysis, factors associated with worse outcomes were older age (P less than .001), higher grade disease (P = .03), larger tumor size (P less than .001), proximal vs. distal location (P less than .001), T stage (P = .006), and TN grouping (P = .004).

In multivariate analysis, variables associated with worse outcomes were age (relative risk 1.025, P = .001), moderately or poorly differentiated or undifferentiated vs. well differentiated cancers (RR, 2.160, 3.323, and 3.306, respectively, P = .004 for all), size (RR, 1.027, P = .001) and location relative to the antrum/pylorus (gastric body RR, 1.289, and cardia/fundus RR, 2.508, P = .001 for both comparisons).

However, neither T stage, N stage nor TN grouping were independent predictors of outcome, the investigators found.

Based on these variables, they devised a risk score for stages IB-IIA, with each of the following four factors receiving 1 point: age greater than 60, tumor size greater than 5 cm, proximal location (cardia or fundus), and histologic grade other than well differentiated.

Under this risk classification system, they saw that 5-year DSS with no risk factors (two patients) was 100%, compared with 86% for patients with one risk factor (92 patients), 76% for those with two risk factors (325), 72% for those with three (372 patients), and 48% for those with all four (136 patients, P less than .001).

The study was internally funded. Dr. Gold reported having no relevant disclosures.

ORLANDO – The type of regional chemotherapy given to patients with in-transit or intralymphatic melanoma of the extremities appears to make a difference in out-of-field recurrences and time to distant recurrence, reported investigators at a symposium of the Society of Surgical Oncology.

A study of 214 patients who underwent either first-time hyperthermic isolated limb perfusion (HILP) or isolated limb infusion (ILI) for in-transit melanomas showed that HILP was associated with significantly more in-field complete responses (P = .01), and a longer median time to out-of-field recurrences, compared with ILI, said Dr. Ketan Sharma of Duke University Medical Center in Durham, NC.

"We found that perfusion complete response and infusion complete response exhibit similar degrees of in-field disease control," Dr. Sharma said.

However, "recurrent disease after a regional therapy complete response is complex, and requires a multidisciplinary approach to treatment," he added.

The National Cancer Institute defines an in-transit metastasis as a "type of metastasis in which skin cancer spreads through a lymph vessel and begins to grow more than 2 centimeters away from the primary tumor but before it reaches the nearest lymph node."

The investigators used data from a prospective database of patients with in-transit melanomas to take a retrospective look at complete responders to either of the two isolated limb therapies. They compared patterns of recurrence and effects on outcomes between the two modalities. In all, 81 patients had first-time HILP and 133 had first-time ILI.

Among 36 patients with a complete response to HILP, 24 had recurrences. Of these patients, 11 experienced in-field-only recurrences, 12 had out-of-field-only recurrences, and 1 patient had a mixed recurrence pattern.

In comparison, 28 of 37 patients with complete responses to ILI had recurrences (9 had in-field-only, 16 had out-of-field, and 3 had mixed recurrence patterns).

There were no significant differences between the perfusion or infusion therapies in time to in-field recurrence, but time to out-of-field recurrence to regional nodes was significantly longer with HILP (42 months vs. 14 months; P = .02). When the authors looked at distant out-of-field recurrences, however, the difference between the treatment types was not significant.

Overall survival after all procedures (including partial responses, stable disease, and nonresponses) was also similar among the treatment types. The overall survival rate after a complete response was higher with HILP (77% vs. 54%); the investigators described this as clinically significant, although it was not statistically significant (P = .1).

At last follow-up (median, 4.0 years for HILP and 2.5 years for ILI), 12 patients who had a complete response to HILP were alive without recurrence for a median duration of 6.5 years, and 24 had recurrences at a median of 3.3 years after perfusion therapy and received additional therapy. Among the latter group, 2 had no evidence of disease, 9 were alive with disease, and 13 died, at a median of 3.1 years.

Among the complete responders to ILI, 9 had no recurrence at a median of 2.6 years; 28 had recurrences at a median of 2.3 years, and received additional treatment. At last follow-up, 8 of the 28 had no evidence of disease (median, 3.9 years), 7 were alive with disease (median, 1.2 years), and 13 had died (median, 2.1 years).

The investigators concluded that the higher proportion of recurrences among patients with an initial complete response to ILI may be due to more frequent lymph node recurrences (nine vs. one in patients who had a complete response to HILP).

The study was supported by Roche/Schering-Plough. Dr. Sharma had no disclosures.

ORLANDO – The type of regional chemotherapy given to patients with in-transit or intralymphatic melanoma of the extremities appears to make a difference in out-of-field recurrences and time to distant recurrence, reported investigators at a symposium of the Society of Surgical Oncology.

A study of 214 patients who underwent either first-time hyperthermic isolated limb perfusion (HILP) or isolated limb infusion (ILI) for in-transit melanomas showed that HILP was associated with significantly more in-field complete responses (P = .01), and a longer median time to out-of-field recurrences, compared with ILI, said Dr. Ketan Sharma of Duke University Medical Center in Durham, NC.

"We found that perfusion complete response and infusion complete response exhibit similar degrees of in-field disease control," Dr. Sharma said.

However, "recurrent disease after a regional therapy complete response is complex, and requires a multidisciplinary approach to treatment," he added.

The National Cancer Institute defines an in-transit metastasis as a "type of metastasis in which skin cancer spreads through a lymph vessel and begins to grow more than 2 centimeters away from the primary tumor but before it reaches the nearest lymph node."

The investigators used data from a prospective database of patients with in-transit melanomas to take a retrospective look at complete responders to either of the two isolated limb therapies. They compared patterns of recurrence and effects on outcomes between the two modalities. In all, 81 patients had first-time HILP and 133 had first-time ILI.

Among 36 patients with a complete response to HILP, 24 had recurrences. Of these patients, 11 experienced in-field-only recurrences, 12 had out-of-field-only recurrences, and 1 patient had a mixed recurrence pattern.

In comparison, 28 of 37 patients with complete responses to ILI had recurrences (9 had in-field-only, 16 had out-of-field, and 3 had mixed recurrence patterns).

There were no significant differences between the perfusion or infusion therapies in time to in-field recurrence, but time to out-of-field recurrence to regional nodes was significantly longer with HILP (42 months vs. 14 months; P = .02). When the authors looked at distant out-of-field recurrences, however, the difference between the treatment types was not significant.

Overall survival after all procedures (including partial responses, stable disease, and nonresponses) was also similar among the treatment types. The overall survival rate after a complete response was higher with HILP (77% vs. 54%); the investigators described this as clinically significant, although it was not statistically significant (P = .1).

At last follow-up (median, 4.0 years for HILP and 2.5 years for ILI), 12 patients who had a complete response to HILP were alive without recurrence for a median duration of 6.5 years, and 24 had recurrences at a median of 3.3 years after perfusion therapy and received additional therapy. Among the latter group, 2 had no evidence of disease, 9 were alive with disease, and 13 died, at a median of 3.1 years.

Among the complete responders to ILI, 9 had no recurrence at a median of 2.6 years; 28 had recurrences at a median of 2.3 years, and received additional treatment. At last follow-up, 8 of the 28 had no evidence of disease (median, 3.9 years), 7 were alive with disease (median, 1.2 years), and 13 had died (median, 2.1 years).

The investigators concluded that the higher proportion of recurrences among patients with an initial complete response to ILI may be due to more frequent lymph node recurrences (nine vs. one in patients who had a complete response to HILP).

The study was supported by Roche/Schering-Plough. Dr. Sharma had no disclosures.

ORLANDO – The type of regional chemotherapy given to patients with in-transit or intralymphatic melanoma of the extremities appears to make a difference in out-of-field recurrences and time to distant recurrence, reported investigators at a symposium of the Society of Surgical Oncology.

A study of 214 patients who underwent either first-time hyperthermic isolated limb perfusion (HILP) or isolated limb infusion (ILI) for in-transit melanomas showed that HILP was associated with significantly more in-field complete responses (P = .01), and a longer median time to out-of-field recurrences, compared with ILI, said Dr. Ketan Sharma of Duke University Medical Center in Durham, NC.

"We found that perfusion complete response and infusion complete response exhibit similar degrees of in-field disease control," Dr. Sharma said.

However, "recurrent disease after a regional therapy complete response is complex, and requires a multidisciplinary approach to treatment," he added.

The National Cancer Institute defines an in-transit metastasis as a "type of metastasis in which skin cancer spreads through a lymph vessel and begins to grow more than 2 centimeters away from the primary tumor but before it reaches the nearest lymph node."

The investigators used data from a prospective database of patients with in-transit melanomas to take a retrospective look at complete responders to either of the two isolated limb therapies. They compared patterns of recurrence and effects on outcomes between the two modalities. In all, 81 patients had first-time HILP and 133 had first-time ILI.

Among 36 patients with a complete response to HILP, 24 had recurrences. Of these patients, 11 experienced in-field-only recurrences, 12 had out-of-field-only recurrences, and 1 patient had a mixed recurrence pattern.

In comparison, 28 of 37 patients with complete responses to ILI had recurrences (9 had in-field-only, 16 had out-of-field, and 3 had mixed recurrence patterns).

There were no significant differences between the perfusion or infusion therapies in time to in-field recurrence, but time to out-of-field recurrence to regional nodes was significantly longer with HILP (42 months vs. 14 months; P = .02). When the authors looked at distant out-of-field recurrences, however, the difference between the treatment types was not significant.

Overall survival after all procedures (including partial responses, stable disease, and nonresponses) was also similar among the treatment types. The overall survival rate after a complete response was higher with HILP (77% vs. 54%); the investigators described this as clinically significant, although it was not statistically significant (P = .1).

At last follow-up (median, 4.0 years for HILP and 2.5 years for ILI), 12 patients who had a complete response to HILP were alive without recurrence for a median duration of 6.5 years, and 24 had recurrences at a median of 3.3 years after perfusion therapy and received additional therapy. Among the latter group, 2 had no evidence of disease, 9 were alive with disease, and 13 died, at a median of 3.1 years.

Among the complete responders to ILI, 9 had no recurrence at a median of 2.6 years; 28 had recurrences at a median of 2.3 years, and received additional treatment. At last follow-up, 8 of the 28 had no evidence of disease (median, 3.9 years), 7 were alive with disease (median, 1.2 years), and 13 had died (median, 2.1 years).

The investigators concluded that the higher proportion of recurrences among patients with an initial complete response to ILI may be due to more frequent lymph node recurrences (nine vs. one in patients who had a complete response to HILP).

The study was supported by Roche/Schering-Plough. Dr. Sharma had no disclosures.

ORLANDO – Malignant melanomas of the scalp behave differently from melanomas arising at other body sites, and are associated with poor disease-free and overall survival compared with other head and neck melanomas, investigators reported here.

A retrospective study of more than 11,000 patients with malignant melanomas showed that 5-year melanoma-specific survival was 65% for patients with lesions on the scalp, compared with 78% for patients with tumors on the trunk or elsewhere on the head, face, neck, or ear (P = .0003), said Dr. Junko Ozao-Choy, a fellow at the John Wayne Cancer Institute in Santa Monica, Calif.

Five-year overall survival for patients with melanomas of the scalp was 58%, compared with 72% for those with head, face, neck, or ear lesions, 74% for those with trunk lesions, and 77% for those with tumors on an extremity (P less than .0001), Dr. Ozao-Choy reported at a symposium sponsored by the Society of Surgical Oncology.

Melanomas of the scalp may account for the poor prognosis of head and neck melanoma relative to tumors originating at other body sites, Dr. Ozao-Choy and her colleagues suggested.

"Scalp melanomas may warrant further studies to ascertain whether biology or anatomy contributes to their worse clinical course," she said, adding that the results indicate "scalp melanomas may need closer clinical follow-up."

Compared with melanomas originating at other body sites, scalp melanomas tend to occur in older patients, predominantly men, according to the investigators. The lesions tend to have higher Breslow thickness, advanced nodal stage and overall stage, and more ulceration.

Dr. Ozao-Choy and her colleagues based their findings on a database review of 11,396 patients presenting for treatment within 4 months of diagnosis from 1971 through 2010. In univariate analysis controlling for sex, they found that 80% of the 799 patients with melanoma originating on the scalp were men (P = .0001).

The mean age at presentation was 54 years for those with scalp lesions and 55 for those with head, neck, or ear tumors. Taken together, the mean age at diagnosis for patients with scalp and head melanomas was higher than for patients with lesions on the trunk (age 47 years) or extremities (age 51 years, P less than .0001).

Scalp tumors had greater Breslow thickness, at a mean of 2.5 mm compared with 1.7 mm for other head and neck melanomas, 1.8 mm for trunk tumors, and 1.9 mm for lesions on an extremity (P less than .0001).

Looking at 5-year overall survival by stage, the authors found that patients with stage I/II scalp lesions had worse survival than those with stage I/II lesions at other sites (P less than .0001). Similarly, stage III scalp primary tumors were associated with worse survival than other stage III tumors (P = .009).

Multivariate analysis controlling for age, male sex, Breslow thickness, lymph node status, and ulceration revealed that patients with scalp tumors had worse 5-year disease-free survival, at 47%, compared with 61% for other head and neck tumors, 66% for trunk tumors, and 69% for extremity melanomas (hazard ratio, 1.8; P less than .0001).

In the question and answer session, an audience member commented that the worse prognosis for head and neck melanomas may be related to the greater frequency of aggressive NRAS and BRAF mutations in tumors originating at those sites.

The study was internally funded. The authors had no disclosures.

ORLANDO – Malignant melanomas of the scalp behave differently from melanomas arising at other body sites, and are associated with poor disease-free and overall survival compared with other head and neck melanomas, investigators reported here.

A retrospective study of more than 11,000 patients with malignant melanomas showed that 5-year melanoma-specific survival was 65% for patients with lesions on the scalp, compared with 78% for patients with tumors on the trunk or elsewhere on the head, face, neck, or ear (P = .0003), said Dr. Junko Ozao-Choy, a fellow at the John Wayne Cancer Institute in Santa Monica, Calif.

Five-year overall survival for patients with melanomas of the scalp was 58%, compared with 72% for those with head, face, neck, or ear lesions, 74% for those with trunk lesions, and 77% for those with tumors on an extremity (P less than .0001), Dr. Ozao-Choy reported at a symposium sponsored by the Society of Surgical Oncology.

Melanomas of the scalp may account for the poor prognosis of head and neck melanoma relative to tumors originating at other body sites, Dr. Ozao-Choy and her colleagues suggested.

"Scalp melanomas may warrant further studies to ascertain whether biology or anatomy contributes to their worse clinical course," she said, adding that the results indicate "scalp melanomas may need closer clinical follow-up."

Compared with melanomas originating at other body sites, scalp melanomas tend to occur in older patients, predominantly men, according to the investigators. The lesions tend to have higher Breslow thickness, advanced nodal stage and overall stage, and more ulceration.

Dr. Ozao-Choy and her colleagues based their findings on a database review of 11,396 patients presenting for treatment within 4 months of diagnosis from 1971 through 2010. In univariate analysis controlling for sex, they found that 80% of the 799 patients with melanoma originating on the scalp were men (P = .0001).

The mean age at presentation was 54 years for those with scalp lesions and 55 for those with head, neck, or ear tumors. Taken together, the mean age at diagnosis for patients with scalp and head melanomas was higher than for patients with lesions on the trunk (age 47 years) or extremities (age 51 years, P less than .0001).

Scalp tumors had greater Breslow thickness, at a mean of 2.5 mm compared with 1.7 mm for other head and neck melanomas, 1.8 mm for trunk tumors, and 1.9 mm for lesions on an extremity (P less than .0001).

Looking at 5-year overall survival by stage, the authors found that patients with stage I/II scalp lesions had worse survival than those with stage I/II lesions at other sites (P less than .0001). Similarly, stage III scalp primary tumors were associated with worse survival than other stage III tumors (P = .009).

Multivariate analysis controlling for age, male sex, Breslow thickness, lymph node status, and ulceration revealed that patients with scalp tumors had worse 5-year disease-free survival, at 47%, compared with 61% for other head and neck tumors, 66% for trunk tumors, and 69% for extremity melanomas (hazard ratio, 1.8; P less than .0001).

In the question and answer session, an audience member commented that the worse prognosis for head and neck melanomas may be related to the greater frequency of aggressive NRAS and BRAF mutations in tumors originating at those sites.

The study was internally funded. The authors had no disclosures.

ORLANDO – Malignant melanomas of the scalp behave differently from melanomas arising at other body sites, and are associated with poor disease-free and overall survival compared with other head and neck melanomas, investigators reported here.

A retrospective study of more than 11,000 patients with malignant melanomas showed that 5-year melanoma-specific survival was 65% for patients with lesions on the scalp, compared with 78% for patients with tumors on the trunk or elsewhere on the head, face, neck, or ear (P = .0003), said Dr. Junko Ozao-Choy, a fellow at the John Wayne Cancer Institute in Santa Monica, Calif.

Five-year overall survival for patients with melanomas of the scalp was 58%, compared with 72% for those with head, face, neck, or ear lesions, 74% for those with trunk lesions, and 77% for those with tumors on an extremity (P less than .0001), Dr. Ozao-Choy reported at a symposium sponsored by the Society of Surgical Oncology.

Melanomas of the scalp may account for the poor prognosis of head and neck melanoma relative to tumors originating at other body sites, Dr. Ozao-Choy and her colleagues suggested.

"Scalp melanomas may warrant further studies to ascertain whether biology or anatomy contributes to their worse clinical course," she said, adding that the results indicate "scalp melanomas may need closer clinical follow-up."

Compared with melanomas originating at other body sites, scalp melanomas tend to occur in older patients, predominantly men, according to the investigators. The lesions tend to have higher Breslow thickness, advanced nodal stage and overall stage, and more ulceration.

Dr. Ozao-Choy and her colleagues based their findings on a database review of 11,396 patients presenting for treatment within 4 months of diagnosis from 1971 through 2010. In univariate analysis controlling for sex, they found that 80% of the 799 patients with melanoma originating on the scalp were men (P = .0001).

The mean age at presentation was 54 years for those with scalp lesions and 55 for those with head, neck, or ear tumors. Taken together, the mean age at diagnosis for patients with scalp and head melanomas was higher than for patients with lesions on the trunk (age 47 years) or extremities (age 51 years, P less than .0001).

Scalp tumors had greater Breslow thickness, at a mean of 2.5 mm compared with 1.7 mm for other head and neck melanomas, 1.8 mm for trunk tumors, and 1.9 mm for lesions on an extremity (P less than .0001).

Looking at 5-year overall survival by stage, the authors found that patients with stage I/II scalp lesions had worse survival than those with stage I/II lesions at other sites (P less than .0001). Similarly, stage III scalp primary tumors were associated with worse survival than other stage III tumors (P = .009).

Multivariate analysis controlling for age, male sex, Breslow thickness, lymph node status, and ulceration revealed that patients with scalp tumors had worse 5-year disease-free survival, at 47%, compared with 61% for other head and neck tumors, 66% for trunk tumors, and 69% for extremity melanomas (hazard ratio, 1.8; P less than .0001).

In the question and answer session, an audience member commented that the worse prognosis for head and neck melanomas may be related to the greater frequency of aggressive NRAS and BRAF mutations in tumors originating at those sites.

The study was internally funded. The authors had no disclosures.

ORLANDO  – Triple-negative breast tumors are notorious for their ability to evade hormonal agents and targeted therapies. But a subset of these intractable tumors expresses androgen receptors, and may be vulnerable to antiandrogen drugs, investigators suggested at the symposium of the Society of Surgical Oncology.

An evaluation of cancerous and normal breast tissues from 94 patients with triple-negative breast tumors (lacking estrogen-, progesterone-, and HER2-receptor expression), showed that the androgen receptor (AR) was expressed in 23% of tumors, reported Dr. Barbara Pockaj of the Mayo Clinic in Scottsdale, Ariz., and colleagues.

AR–positive cancers occurred more frequently in older patients and were significantly associated with lymph node metastases, compared with AR-negative, triple-negative breast cancers in this small study, she said.

There was a trend toward higher tumor stage at diagnosis among patients with AR-positive tumors, but AR-positive and AR-negative patients did not differ significantly in overall or recurrence-free survival, said Dr. Pockaj.

Previous studies have suggested that 10%-43% of triple-negative breast cancers bear the androgen receptor, but the antibodies and methods used for characterizing the presence of the receptor varied significantly, making it difficult to nail down actual numbers, Dr. Pockaj said.

The investigators examined 177 tissue biopsy cores from 94 patients with triple-negative breast cancer to see whether AR-receptor expression correlated with patient and tumor factors and survival, and where expression of the receptors in different tissues from the same patients correlated with tumor progression.

The receptor was expressed in 88% of normal breast tissues in the samples, a proportion identical to that of the estrogen receptor in normal tissues from the same patients.

AR expression was detected in all 6 of 6 adjacent DCIS (ductal carcinoma in situ) samples from AR-positive patients, and in 9 of 15 DCIS samples from patients with AR-negative cancer.

All lymph node metastases from AR-positive patients were also positive, whereas no lymph node metastases from AR-negative patients were found to express the androgen receptor.

There were no significant differences between AR-positive and -negative patients in tumor grade, angiolymphatic invasion, TNM (tumor, node, metastasis) stage, or tumor size. In contrast, 16 of the 72 (22%) of the AR-negative patients had lymph node metastases, compared with 10 of 22 (46%) AR-positive patients (P = .033).

There were no significant differences in locoregional recurrences, overall survival, or disease-specific survival between positive and negative patients.

Among all 94 patients, however, the presence of lymph node metastases was associated with a significantly worse recurrence-free survival (hazard ratio, 5.502; P = .017). Chemotherapy significantly reduced that risk (HR, 0.099; P = .0004).

In multivariate analysis, the investigators found that AR expression was associated with older age (63 years vs. 57 years; P = .051) and with the presence of lymph nodes metastases (P = .033).

Although larger studies are needed, the data suggest that AR-positive, triple-negative breast cancer "has unique clinical behavior and may need different treatment," Dr. Pockaj said.

A phase II, open-label trial of the antiandrogen bicalutamide (Casodex) in patients with AR-positive, estrogen- and progesterone-receptor–negative, metastatic breast cancer is currently underway in eight U.S. cancer centers.

The current study was supported by the Translational Genomics Research Institute, Phoenix, and the Mayo Clinic, Scottsdale, Ariz. Dr. Pockaj is an employee of the Mayo Clinic.

ORLANDO  – Triple-negative breast tumors are notorious for their ability to evade hormonal agents and targeted therapies. But a subset of these intractable tumors expresses androgen receptors, and may be vulnerable to antiandrogen drugs, investigators suggested at the symposium of the Society of Surgical Oncology.

An evaluation of cancerous and normal breast tissues from 94 patients with triple-negative breast tumors (lacking estrogen-, progesterone-, and HER2-receptor expression), showed that the androgen receptor (AR) was expressed in 23% of tumors, reported Dr. Barbara Pockaj of the Mayo Clinic in Scottsdale, Ariz., and colleagues.

AR–positive cancers occurred more frequently in older patients and were significantly associated with lymph node metastases, compared with AR-negative, triple-negative breast cancers in this small study, she said.

There was a trend toward higher tumor stage at diagnosis among patients with AR-positive tumors, but AR-positive and AR-negative patients did not differ significantly in overall or recurrence-free survival, said Dr. Pockaj.

Previous studies have suggested that 10%-43% of triple-negative breast cancers bear the androgen receptor, but the antibodies and methods used for characterizing the presence of the receptor varied significantly, making it difficult to nail down actual numbers, Dr. Pockaj said.

The investigators examined 177 tissue biopsy cores from 94 patients with triple-negative breast cancer to see whether AR-receptor expression correlated with patient and tumor factors and survival, and where expression of the receptors in different tissues from the same patients correlated with tumor progression.

The receptor was expressed in 88% of normal breast tissues in the samples, a proportion identical to that of the estrogen receptor in normal tissues from the same patients.

AR expression was detected in all 6 of 6 adjacent DCIS (ductal carcinoma in situ) samples from AR-positive patients, and in 9 of 15 DCIS samples from patients with AR-negative cancer.

All lymph node metastases from AR-positive patients were also positive, whereas no lymph node metastases from AR-negative patients were found to express the androgen receptor.

There were no significant differences between AR-positive and -negative patients in tumor grade, angiolymphatic invasion, TNM (tumor, node, metastasis) stage, or tumor size. In contrast, 16 of the 72 (22%) of the AR-negative patients had lymph node metastases, compared with 10 of 22 (46%) AR-positive patients (P = .033).

There were no significant differences in locoregional recurrences, overall survival, or disease-specific survival between positive and negative patients.

Among all 94 patients, however, the presence of lymph node metastases was associated with a significantly worse recurrence-free survival (hazard ratio, 5.502; P = .017). Chemotherapy significantly reduced that risk (HR, 0.099; P = .0004).

In multivariate analysis, the investigators found that AR expression was associated with older age (63 years vs. 57 years; P = .051) and with the presence of lymph nodes metastases (P = .033).

Although larger studies are needed, the data suggest that AR-positive, triple-negative breast cancer "has unique clinical behavior and may need different treatment," Dr. Pockaj said.

A phase II, open-label trial of the antiandrogen bicalutamide (Casodex) in patients with AR-positive, estrogen- and progesterone-receptor–negative, metastatic breast cancer is currently underway in eight U.S. cancer centers.

The current study was supported by the Translational Genomics Research Institute, Phoenix, and the Mayo Clinic, Scottsdale, Ariz. Dr. Pockaj is an employee of the Mayo Clinic.

ORLANDO  – Triple-negative breast tumors are notorious for their ability to evade hormonal agents and targeted therapies. But a subset of these intractable tumors expresses androgen receptors, and may be vulnerable to antiandrogen drugs, investigators suggested at the symposium of the Society of Surgical Oncology.

An evaluation of cancerous and normal breast tissues from 94 patients with triple-negative breast tumors (lacking estrogen-, progesterone-, and HER2-receptor expression), showed that the androgen receptor (AR) was expressed in 23% of tumors, reported Dr. Barbara Pockaj of the Mayo Clinic in Scottsdale, Ariz., and colleagues.

AR–positive cancers occurred more frequently in older patients and were significantly associated with lymph node metastases, compared with AR-negative, triple-negative breast cancers in this small study, she said.

There was a trend toward higher tumor stage at diagnosis among patients with AR-positive tumors, but AR-positive and AR-negative patients did not differ significantly in overall or recurrence-free survival, said Dr. Pockaj.

Previous studies have suggested that 10%-43% of triple-negative breast cancers bear the androgen receptor, but the antibodies and methods used for characterizing the presence of the receptor varied significantly, making it difficult to nail down actual numbers, Dr. Pockaj said.

The investigators examined 177 tissue biopsy cores from 94 patients with triple-negative breast cancer to see whether AR-receptor expression correlated with patient and tumor factors and survival, and where expression of the receptors in different tissues from the same patients correlated with tumor progression.

The receptor was expressed in 88% of normal breast tissues in the samples, a proportion identical to that of the estrogen receptor in normal tissues from the same patients.

AR expression was detected in all 6 of 6 adjacent DCIS (ductal carcinoma in situ) samples from AR-positive patients, and in 9 of 15 DCIS samples from patients with AR-negative cancer.

All lymph node metastases from AR-positive patients were also positive, whereas no lymph node metastases from AR-negative patients were found to express the androgen receptor.

There were no significant differences between AR-positive and -negative patients in tumor grade, angiolymphatic invasion, TNM (tumor, node, metastasis) stage, or tumor size. In contrast, 16 of the 72 (22%) of the AR-negative patients had lymph node metastases, compared with 10 of 22 (46%) AR-positive patients (P = .033).

There were no significant differences in locoregional recurrences, overall survival, or disease-specific survival between positive and negative patients.

Among all 94 patients, however, the presence of lymph node metastases was associated with a significantly worse recurrence-free survival (hazard ratio, 5.502; P = .017). Chemotherapy significantly reduced that risk (HR, 0.099; P = .0004).

In multivariate analysis, the investigators found that AR expression was associated with older age (63 years vs. 57 years; P = .051) and with the presence of lymph nodes metastases (P = .033).

Although larger studies are needed, the data suggest that AR-positive, triple-negative breast cancer "has unique clinical behavior and may need different treatment," Dr. Pockaj said.

A phase II, open-label trial of the antiandrogen bicalutamide (Casodex) in patients with AR-positive, estrogen- and progesterone-receptor–negative, metastatic breast cancer is currently underway in eight U.S. cancer centers.

The current study was supported by the Translational Genomics Research Institute, Phoenix, and the Mayo Clinic, Scottsdale, Ariz. Dr. Pockaj is an employee of the Mayo Clinic.