A. M. Cohen

PURPOSE: More than 50,000 patients in the United States will present each year with liver metastases from colorectal cancers. The current study was performed to determine if liver resection for colorectal metastases is safe and effective and to evaluate predictors of outcome. MATERIALS AND METHODS: Data for 456 consecutive resections performed between July 1985 and December 1991 in a tertiary referral center were analyzed. RESULTS: The perioperative mortality rate was 2.8%, with a mortality rate of 4.6% for resections that involved a lobectomy or more. The median hospital stay was 12 days and only 9% of patients were admitted to the intensive care unit. The 5-year survival rate is 38%, with a median survival duration of 46 months. By univariate analysis, nodal status of the primary lesion, short disease-free interval before detection of liver metastases, carcinoembryonic antigen (CEA) level greater than 200 ng/mL, multiple liver tumors, extrahepatic disease, large tumors, or positive resection margin was predictive of poorer outcome. Sex, age greater than 70 years, site of primary tumor, or perioperative transfusion was not predictive of outcome. By multivariate analysis, positive margin, size greater than 10 cm, disease-free interval less than 12 months, multiple tumors, and extrahepatic disease were independent predictors of poorer outcome. Short disease-free interval or multiple tumors were nevertheless associated with a 5-year survival rate greater than 24%. CONCLUSION: Liver resection for colorectal metastases is safe and effective therapy and currently represents the only potentially curative therapy for metastatic colorectal cancer. The only absolute contraindication to resection is extrahepatic disease. A randomized trial to examine efficacy of surgical resection cannot ethically be performed. Liver resection should be considered standard therapy for all fit patients with colorectal metastases isolated to the liver.

PURPOSE: This study was designed to review experience with neuroendocrine carcinomas of the colon and rectum at a single institution, with emphasis on the pathology and clinical characteristics of this uncommon malignancy. METHODS: A study group of patients was identified from a prospective colorectal service database. Pathology was reviewed and neuroendocrine tumors were classified by a single pathologist. Medical records were retrospectively reviewed. RESULTS: From March 1975 to September 1998, 38 patients with neuroendocrine carcinomas were identified from the colorectal service database comprising 6495 patients (0.6 percent). These neuroendocrine carcinomas did not include carcinoid tumors. Average patient age was 57 years (range, 29-86 years). There were 17 males (44.7 percent) and 21 females (55.3 percent). Tumors were located as follows: 17 colon, 14 rectum, 6 anal canal, and 1 appendix. The diagnosis of neuroendocrine carcinoma was suggested preoperatively from tissue biopsy in 59.3 percent (16/27) of patients evaluable. Pathology was reviewed and tumors were categorized as small cell carcinoma (n = 22) or large cell neuroendocrine carcinoma (n = 16). Most tumors (20/25 evaluable, 80 percent) stained positive by means of immunohistochemistry for neuroendocrine markers, including chromogranin (18/19), synaptophysin (10/15), and/or neuron-specific enolase (14/15). Metastatic disease was detected at the time of diagnosis in 69.4 percent of the patients (25/36). Tumors were advanced at the time of diagnosis, with American Joint Committee on Cancer (AJCC) Stage I (n = 6), Stage III (n = 7), and Stage IV (n = 25) tumors. As a group, these tumors had a poor prognosis, with a median survival of 10.4 months. One-year, two-year, and three-year survival was 46 percent, 26 percent, and 13 percent, respectively. There was no significant difference in survival based on pathologic subtypes. Median follow-up time was 9.4 months (range, 0.6-263.7 months). CONCLUSIONS: Neuroendocrine carcinomas of the colon and rectum are uncommon, comprising less than 1 percent of colon and rectal cancers. Pathologically, these tumors are poorly differentiated carcinomas with distinctive cytoarchitectural features and are often immunoreactive for markers of neuroendocrine differentiation. The prognosis for high-grade neuroendocrine carcinomas is poor, as most patients have metastatic disease at the time of diagnosis.

PURPOSE: To determine if fluorouracil (5-FU) plus high-dose leucovorin (LV) enhances local response in patients receiving preoperative radiation therapy (RT) for adenocarcinoma of the rectum, we compared the degree of downstaging in patients receiving preoperative RT with or without chemotherapy. PATIENTS AND METHODS: For this comparison, three groups of patients who were treated with identical doses and techniques of preoperative pelvic RT (total dose of 5,040 cGy) were examined. Group 1 included 20 patients with unresectable disease who received combined RT and LV/5-FU. Group 2 included 11 patients with unresectable disease who received preoperative RT. Group 3 included 21 patients with invasive, resectable, primary disease who received preoperative RT. RESULTS: Patients with unresectable disease who received LV/5-FU had a higher rate of pathologic complete response (20% v 0%) and a lower incidence of positive nodes (30% v 64%) compared with those who did not receive chemotherapy. Even when the most favorable group of patients was included (group 3), patients who received LV/5-FU still had a higher complete response rate (20% v 6%) and a lower incidence of positive nodes (30% v 53%) compared with those who received RT without LV/5-FU. Of those patients with initially unresectable disease, the resectability rate was higher in those who received LV/5-FU compared with those who did not receive LV/5-FU (90% v 64%). Patients who received LV/5-FU experienced slightly more grade 1 to 2 fatigue, stomatitis, nausea, and grade 3 diarrhea, tenesmus, and dysuria. CONCLUSIONS: Despite the fact that patients who received chemotherapy (group 1) had more advanced disease compared with those with resectable disease (group 3), the addition of LV/5-FU increased the resectability and downstaging rates. The ultimate impact of a complete response as well as a decrease in the incidence of pelvic nodes on local control and survival remains to be determined. However, given the enhancement of down-staging in patients with unresectable rectal cancer, we are encouraged by the combined modality approach.

PURPOSE: We compared the combined radiation therapy (RT) plus chemotherapy segments of two separate parallel phase I trials to determine if combined pelvic RT, fluorouracil (5-FU), and high-dose leucovorin (LV) had less acute toxicity when delivered preoperatively versus postoperatively in patients with rectal cancer. PATIENTS AND METHODS: Patients with unresectable disease received preoperative RT plus LV and 5-FU followed by surgery and postoperative LV and 5-FU. Patients with resectable disease received identical doses, techniques, and schedules of RT and LV and 5-FU except all therapy was delivered postoperatively. On day 1, patients received LV and 5-FU times one cycle. RT began on day 8. A second cycle of LV and 5-FU was given concurrently with the fourth week of RT. RESULTS: Although more patients (75% v 32%; P = .02) received the higher dose level of 5-FU (250 mg/m2), significantly fewer experienced acute grade 3 to 4 toxicity with preoperative versus postoperative therapy (13% v 48%; P = .045). There was no grade 3 to 4 myelosuppression in either group. The two grade 3 toxicities in the preoperative group were gastrointestinal. The grade 3 toxicities in the postoperative group included seven gastrointestinal and two genitourinary; four patients had a grade 4 toxicity. CONCLUSION: Given the high incidence of grade 3 to 4 toxicity also reported in the postoperative combined modality adjuvant randomized trials, future adjuvant trials should explore the preoperative approach.

PURPOSE: In this study, we investigated the association between matrix metalloproteinase (MMP)-9 RNA expression in primary colorectal cancer (CRC) and standard clinicopathologic variables and determined whether levels of MMP-9 RNA predict relapse and survival. PATIENTS AND METHODS: Tumor and paired normal mucosa specimens from 71 primary CRC patients following resections were assessed. RNA levels were determined via Northern blot hybridization and quantitated with laser densitometry. Results were expressed as tumor/normal mucosa (T/N) fold-increase calculated after normalizing for RNA loading using 28S expression. Statistical analysis was performed using the SAS statistical package procedure. Kaplan-Meier survival curves were compared with the two-sided log-rank test. RESULTS: The mean T/N MMP-9 RNA fold-increase was 9.4 +/- 1.0 (mean +/- SE) (P < .001). Overexpression of MMP-9 RNA correlated significantly with status of synchronous distant metastases (M stage) (P = .004) and Dukes' stage (P = .008). A T/N fold-increase of 5.0 was used to discriminate between high (> 5.0) and low (< or = 5.0) T/N MMP-9 expression. High T/N MMP-9 RNA expression was associated with a significantly shorter disease-free (P = .0001) and overall (P = .0002) survival duration. In univariate and multivariate analyses, T/N MMP-9 RNA level was found to be an independent prognostic factor for disease-free survival. CONCLUSION: This report provides the first evidence that increased MMP-9 RNA production in primary human CRC may be a powerful, independent predictor of recurrence and outcome.