One-hundred-ninety-four eligible and evaluable patients with histologically confirmed locally unresectable adenocarcinoma of the pancreas were randomly assigned to therapy with high-dose (6000 rads) radiation therapy alone, to moderate-dose (4000 rads) radiation + 5-fluorouracil (5-FU), and to high-dose radiation plus 5-FU. Median survival with radiation alone was only 51/2 months from date of diagnosis. Both 5-FU-containing treatment regimens produced a highly significant survival improvement when compared with radiation alone. Forty percent of patients treated with the combined regimens were still living at one year compared with 10% of patients treated with radiation only. Survival differences between 4000 rads plus 5-FU and 6000 rads plus 5-FU were not significant with an overall median survival of ten months. Significant prognostic variables, in addition to treatment, were pretreatment performance status and pretreatment CEA level.

This study explores the patterns of recurrence after "curative" operation for colorectal cancer. For an 11-year period, 1960-1971, 281 evaluable patients were resected at the Peter Bent Brigham Hospital. Crude five-year survival in these patients was 49%, but only 10% of those with recurrence lived five years. A total of 69 patients relapsed during their lifetime and 34 additional patients were found to have metastases at death. The initial site of metastases were regional in 23 patients (33%) and distant in 32 (46%). Simultaneous regional and distant metastases were found in 13 (19%) for a total of 65% of patients having initial distant metastases. Approximate recurrence rates by site were: 30% for sigmoid and rectum, 20% for right colon, and 10% for transverse and left colon. Tumor size was a significant determinant of recurrence but did not select for regional or distant sites. Recurrence by Astler-Coller modification of the Dukes-Kirklin classification revealed 10% for A + B1, 33% for B2, 35% for C1, and 50% for C2. More than half of the patients with distant metastases (18/32) had solely hepatic metastases yet the total incidence of liver metastases as the initial site was only 8% of the total. In general, the site of the primary cancer was the most important determinant of the type of recurrence; the stage and site of the primary tumor were most predictive for eventual relapse.

In an effort to determine the patterns of failure and survival of colon cancer, a retrospective review of 294 patients who underwent potentially curative surgery at the New England Deaconess Hospital (NEDH) was performed. For the entire group, the 5-year crude survival rate was 68% and the actuarial rate was 80%. Survival decreased with increasing bowel wall penetration by tumor and the presence of lymph node metastasis. Although survival varied with the tumor site, none of the differences was statistically significant. Other variables, including the grade of adenocarcinoma, size, and the type of surgery had a significant impact on survival. Patterns of failure, expressed as the actuarial incidence of first diagnosed failure at 5 years, were examined by stage and site. There was a trend toward increased failure with increasing bowel wall penetration by tumor and the presence of lymph node metastasis. Abdominal failure, either as the only site or as a component of failure, was the most common type of failure. When compared by site, patients with cecal carcinoma had a significantly lower incidence of local and distant failure than patients with disease in other selected sites. No differences in patterns of failure were seen in patients with carcinomas in the mobile sections of the colon compared with those who had disease arising in the nonmobile sections of the colon. These data may be useful in identifying those patients who might benefit most from adjuvant therapy.

To determine the clinicopathologic significance of colloid carcinoma in carcinoma of the colon and rectosigmoid/rectum, a retrospective review of 462 patients who underwent potentially curative surgery at the New England Deaconess Hospital was performed. Seventy-seven patients (17%) were identified who had tumors with some component of colloid present. Colloid carcinoma occurred in 49 (11%). The remaining 28 (6%) had adenocarcinoma with colloid features. Compared to patients with pure adenocarcinoma, the 5-year actuarial survival of patients with colloid carcinoma was lower in the colon, rectosigmoid/rectum, and colorectum. Patterns of failure, expressed as the actuarial incidence of failure at 5 years, were examined by histologic condition and stage. Patients with Dukes' Stage B colloid carcinoma had a higher incidence of total failure, and patients with Dukes' Stage C colloid carcinoma had a higher incidence of local, abdominal, and total failure. None of the differences reached statistical significance. The presence of colloid carcinoma may have a real but small impact on the patterns of failure and survival in colorectal cancer.

In an effort to determine the patterns of failure and survival of rectosigmoid and rectal cancer, a retrospective review of 168 patients who underwent potentially curative surgery at the New England Deaconess Hospital was performed. The 5-year actuarial survival for the entire group was 67%. Survival rates decreased with increasing penetration of the bowel wall by tumor and the presence of lymph node metastasis, but only the latter reached statistical significance. Those patients who underwent an abdominoperineal resection also experienced a significant decrease in survival compared to a low anterior resection. Patterns of failure, expressed as the actuarial incidence of first failure at 5 years, were examined by stage. With the exception of stages B3 and C3, there was a trend towards increased abdominal, distant, and total failure with increasing bowel wall penetration by tumor. A similar trend was seen in local failure in those patients with positive nodes. Knowledge of these data may help identify those patients who may benefit most from adjuvant therapy.