Adjuvant Therapy in Rectal Cancer: Improving Local Control Leads to Improved Survival

ABSTRACT & COMMENTARY

Numerous retrospective and randomized clinical trials have shown that the use of radiotherapy as an adjunct to surgical resection of rectal carcinoma decreases local failure rates following curative resection. The optimal timing of the radiotherapy, preoperative or postoperative, has been controversial and, until now, no study has demonstrated an improvement in survival with either approach. However, older trials have used suboptimal radiation techniques, inadequate doses, and, because of small patient numbers, had insufficient power to detect a survival benefit. Therefore, the Swedish rectal cancer trial was designed to test whether a short course of preoperative radiotherapy using optimal techniques could substantiate previously observed reductions in local control and evaluate its effect on survival.

Eleven hundred patients were randomized to either immediate surgery alone or preoperative radiotherapy delivering 25 Gy in five fractions, using a three- or four-beam technique, which corresponds to a dose of approximately 45 Gy given in conventional fractionation. An anterior resection or abdominal perineal resection was performed within one week of the completion of radiotherapy. Three percent of patients in both groups did not undergo resection because the tumor was unresectable or metastatic disease was found at the time of surgery. In-hospital mortality rate was 4% in the radiotherapy group and 3% in the surgery alone group. The distribution of Dukes’ stages between the two groups was different, with more Dukes’ stage A patients and fewer Dukes’ stage C patients in the radiotherapy plus surgery group reflecting the downstaging that occurs following preoperative therapy. The patients were followed biannually for five years with laboratory tests or imaging performed only if a recurrence was suspected. Any recurrence within the pelvis was considered a local recurrence.

The results, after a minimum follow up of five years, show that only 11% of patients in the irradiated group had a local recurrence, compared to 27% in the group undergoing surgery alone. There was a significant reduction in local recurrence rates for patients with all three of the Dukes’ stages of disease. The recurrence rates for Dukes’ stage A, B, and C patients treated with preoperative radiotherapy were 4%, 10%, and 20%, respectively, vs. 12%, 23%, and 40% following surgery alone. The overall reduction in the rate of local recurrence was 58%. The overall rate of recurrence was 28% in the radiotherapy plus surgery group vs. 38% for surgery alone (P < 0.001). The overall rate of distant metastatic disease was 23% for the radiotherapy plus surgery group and 24% in the surgery alone group. The five-year survival rate was 58% in the radiotherapy group and 48% in the surgery alone group (P = 0.004) with a relative hazard of death from all causes in the radiotherapy plus surgery group of 0.79. A Cox regression analysis was performed that included the effects of downstaging seen with the radiotherapy and the relative hazard of death from all causes changed marginally to 0.81. Long-term toxicities were not discussed; however, the quality of life of patients in the two groups was comparable despite more problems with ano-rectal function in the irradiated group. (Swedish Rectal Cancer Trial. N Engl J Med 1997;336:980-987.)

COMMENTARY

Adenocarcinoma of the rectum is ideally suited to test the hypothesis that an improvement in local control can lead to an improvement in overall survival. In this region, patients with local failure are seldom salvageable and, because of the significant morbidity, local recurrence is best avoided. Thus, the rationale for adjuvant radiation therapy is well grounded. However, if local failure is a marker for ultimate distant failure, then improvement in local control will have no effect on overall survival. The findings of this trial indicate that this is not the case. Distant metastatic rates were similar between the two groups, whereas local control was substantially improved by the addition of the short course of preoperative radiotherapy. Why have previous studies failed to demonstrate a similar improvement in survival? The GITSG study,2 which randomized patients in the postoperative setting to chemotherapy alone, radiotherapy alone, or combined modality treatment, had only 50 patients in each arm, and by today’s standards, the dose of radiotherapy used would be considered insufficient to control microscopic disease. Thus, there was only a minimal improvement in local control with radiotherapy alone compared to the control group. The NSABP R-01 trial3 compared postoperative radiotherapy to systemic chemotherapy. The doses of radiation were only slightly higher than in the GITSG study. Local failure was reduced from 25% to 16% with the addition of postoperative radiotherapy, but no improvement in disease-free survival was seen. There may be some benefit in using a higher dose of radiotherapy in the postoperative setting. A trial from Sweden found that 25 Gy given preoperatively was more effective than 60 Gy administered postoperatively; preoperative radiation reduced local failure rates from 21% to 12%.4 A single institution retrospective analysis from Thomas Jefferson University found that higher doses of preoperative radiotherapy (55 Gy) were more effective than 45 Gy in reducing local failures as well.5 Thus, the positive results of the Swedish trial may be due to the more effective dose of radiation delivered, and the statistical power provided by the larger number of patients treated.

The authors suggest that the improvement in survival seen in this study is comparable to adjuvant chemotherapy trials reported above. However, they overlook the benefits of adjuvant chemotherapy in the reduction of distant metastases seen in the initial GITSG study, the R-01 trial, and a later GITSG trial,6 which randomized patients between radiation therapy alone to higher doses in the postoperative setting and the same radiation therapy combined with chemotherapy. In the Swedish study, the rates of distant metastatic disease were similar in the two groups. Thus, radiotherapy is probably not preventing further dissemination of disease, and it must be improving survival by preventing deaths from uncontrolled local disease. If maximization of local control can lead to improved survival, as this study has shown, then improvement in systemic control may lead to further improvements in survival, as was found in the studies mentioned above. Therefore, it seems quite logical that the optimum strategy for the adjuvant treatment of patients with rectal cancer would combine both chemotherapy and radiation therapy given in the preoperative setting. The benefits of this approach not only include avoiding the increased toxicity from postoperative radiation after an abdominal perineal resection but also has the advantage of improving resectability in patients with locally advanced lesions. In addition, combined modality adjuvant therapy may enable a surgeon to perform an anal sparing procedure if adequate shrinkage of the primary tumor can be achieved. The disadvantages of preoperative treatment are perhaps the overtreatment of patients with Dukes’ stage A lesions or patients with distant metastatic disease. The use of endoscopic ultrasound can readily identify those patients at low risk for local and systemic recurrence. The characteristics of the primary tumor, including size, ulceration, and differentiation, can also identify those patients at risk for nodal involvement and distant metastatic spread. Preoperative CT scans should identify patients with hepatic metastases. However, as reported in this trial, even patients with hepatic metastases are at high risk for local recurrence. And because of the long natural history of metastatic colon cancer, such patients may also benefit from adjuvant local therapy to decrease their risk of local recurrence. Thus, the use of preoperative radiotherapy for rectal carcinoma is encouraged. Whether a short course, high fraction per dose approach as used in this trial should be used or a more conventional fractionation scheme as typically employed should be used is unknown. We believe the optimal strategy would employ both preoperative radiation therapy and chemotherapy with continuation of chemotherapy postoperatively based upon the operative findings.

References

1. Swedish Rectal Cancer Trial. N Engl J Med 1997;336:980-987.

2. GITSG. N Engl J Med 1985;312:1465-1472.

3. Fisher B, et al. J Natl Cancer Inst 1988;80: 21-29.

4. Pahlman L, Glimelius B. Ann Surg 1990;211: 187-195.

5. Ahmad N, et al. Int J Radiat Oncol Biol Phys 1993;27:773-778.

6. Krook JE, et al. N Engl J Med 1991;324:709-715.