SAN DIEGO — A 15-year study of pancreatic cancer trends in the United States found more patients are receiving potentially curative surgery, but only those with node-positive disease benefit from adjuvant radiation, Dr. Nancy N. Baxter reported at a symposium sponsored by the Society of Surgical Oncology.
Although cancer-directed surgery improved survival for patients with localized disease, median and long-term outcomes remained dismal, she said.
“Pancreatic cancer has the worst survival rate of any malignancy,” said Dr. Baxter of the department of surgery at the University of Minnesota in Minneapolis.
She and her colleagues used the Surveillance, Epidemiology, and End Results (SEER) database to track 48,658 patients over age 18 who were diagnosed with pancreatic cancer between 1988 and 2002. The age-adjusted incidence of pancreatic cancer did not change over this period, she said. While 17.5% of patients were not staged, she reported that staging improved significantly over time. About a third of all patients had nonmetastatic disease.
The data showed that median survival was 4 months overall, but 9 months for patients with nonmetastatic disease, Dr. Baxter said. Disease-specific survival for all patients was 4.8% at 5 years. Patients with nonmetastatic disease did better—at 8.8%—than did patients with metastases, at 1.5%.
A total of 12% of all patients underwent cancer-directed surgery. Among 15,031 patients with nonmetastatic disease, 4,462 (30%) had surgery. The proportion of patients with nonmetastatic cancer who received surgery increased from 19% in 1988 to 34% in 2002.
Two subgroups were much less likely to be operated on, however: the elderly and African Americans. A person under age 50 was more than 11 times as likely to have surgery as someone over age 80. Compared with white patients, the odds ratio predicting surgery for African Americans was 0.85.
“There is no rationale for disparities in rate,” Dr. Baxter said, in answer to an audience question about the age differential. “We need to think about treating patients who are older more aggressively.”
For patients with nonmetastatic disease, median survival was longer for those who had surgery than for those who did not: 16 months vs. 7 months, respectively. Those who had surgery also had better 5-year disease-free survival (19.7% vs. 3.8%) and better 5-year overall survival (14.9% vs. 2%).
About 42% of patients who had surgery also were irradiated. The SEER database does not include chemotherapy data, Dr. Baxter said. But because radiochemotherapy is standard for pancreatic cancer, most patients receiving radiation probably had radiochemotherapy, she acknowledged.
Patients with localized disease were less likely to have radiation therapy after surgery (odds ratio 0.78). Among those operated on, radiation was more likely in patients under age 50 than among those over age 80 (odds ratio 5.7) and was less likely in African Americans than in whites (odds ratio 0.74).
To evaluate the effects of radiation, the investigators considered only those patients who survived long enough for radiation to be an option. They assessed 3,756 patients who were irradiated and lived at least 3 months after diagnosis.
If patients had radiation therapy, 5-year disease-specific survival was slightly worse with node-negative localized disease and about the same with node-negative extensive disease, but was better with node-positive disease (hazard ratio 0.73), she said.
“The effect of adjuvant radiation is dependent on the extent of disease,” she said. “For patients with limited disease, there is no apparent benefit. For patients with node-positive disease, there is benefit.”
Dr. Baxter added that the analysis supports the results of the adjuvant European Study Group for Pancreatic Cancer-1 (ESPAC-1) study, which found benefits for surgery and chemotherapy but not for chemoradiation (Ann Surg. 2001;234:758–68). Radiation's apparent lack of efficacy in patients with limited disease may explain the difference between the ESPAC-1 results and results of single-institution series, she said, in that the latter typically enroll sicker patients.