Preoperative radiotherapy in rectal cancer : aspects on fractionation and timing of surgery

Abstract

Neo-adjuvant radiotherapy (RT) in rectal cancer (RC) reduces the risk for local recurrence (LR). The optimal fractionation or time to surgery is not determined. The focus areas of this thesis are different RT-courses and timing of surgery in patients with RC. The Stockholm III trial forms the basis of the studies included in the thesis. Between 1998 – 2013, patients with primarily resectable adenocarcinoma of the rectum were randomly allocated to three different RT-courses. SRT - 5 Gy x 5 and surgery within one week, SRT-delay - 5 Gy x 5 and surgery after 4-8 weeks or LRT-delay - 2 Gy x 25 and surgery after 4-8-weeks. Including centres could choose to randomise patients between three courses or between the two courses with 5 Gy x 5. Primary endpoint was time to LR, secondary end points included distant metastases (DM), survival, tumour regression and adverse events. All patients have been registered in the Swedish ColoRectal Cancer Registry.

Paper I. All 840 patients randomised in the Stockholm III trial were analysed after a minimum follow up of 2 years. 357, 355 and 128 patients were allocated to SRT, SRT-delay and LRT-delay respectively. The three armed randomisation was analysed separately and the patients randomised to any of the courses with 5 Gy x 5 were pooled and analysed in a short course RT comparison. About 6 -7 % of the patients with a delay to surgery required hospitalisation between start of RT and surgery due to RT-induced toxicity. In total, 25 patients had a LR within the follow up time, without statistical significant differences between the groups. The cumulative incidence of DM, overall survival (OS) and recurrence free survival (RFS) did not differ between the groups. We found a statistical significant reduction of post-operative complications in SRT-delay compared to SRT (OR 0·61 [95% CI 0·45–0·83] p=0·001). Paper II. The aim this study was to evaluate the post-operative complications in relation to the exact overall treatment time (OTT). Patients were categorized according to OTT and fractionation. Patients that received 5 Gy x 5 were divided into four groups; Group A: 7 days, B: 8-13 days, C: 5-7 weeks, D: 8-13 weeks. Patients that received 2 Gy x 25 were divided in two groups; Group E: 9-11 weeks and F: 12-14 weeks. Main outcome was post-operative complications defined as any-, surgicalor infectious complication. Adjusted odds ratios (any complication) were; A vs. B OR (95 % CI); 0.72 (0.40-1.32) p=0.289, C vs. B 0.50 (0.30-0.84) p=0.009, and D vs. B 0.39 (0.23-0.65) p<0.001. There were no statistical significant differences between group E and F.

Paper III. In this study, all available histopathology slides from the resected tumours have been reassessed by one pathologist. Tumour regression was the main outcome and secondary outcomes were histopathological characteristics and the correlation between tumour response and survival. Patients randomised to SRT-delay showed more tumour regression compared to the other arms. A complete pathology graded tumour regression (pCR) was seen in about 10 % of the patients after SRT-delay. Patients with pCR had improved OS and time to recurrence, compared to patients with lower regression grades. Hazard Ratio pCR vs no-pCR: OS: 0.51 (0.26–0.99) p = 0.046, TTR: 0.27 (0.09–0.86) p = 0.027. Paper IV. Long-term follow up of the Stockholm III trial after a minimum follow-up of 5 years. The endpoints from paper I were analysed. The incidence of LR was 11 of 357 (3,1 %), 13 of 355 (3,7) %, 7 of 128 (5,5%) in SRT, SRT-delay and LRT-delay. Incidence of DM was 88 of 257 (24,7%), 82 of 355 (23,1%), 38 of 128 (29,7%). The median OS was 8.14 (7.23-9.98), 10.18 (8.45-11.68) 10,53 (6.95-11.34) years in SRT, SRT-delay and LRT-delay without statistical differences between the groups, log-rank SRT vs. SRT-delay p=0.162 (short course RT comparison), SRT vs. LRT-delay p=0.738 (three armed randomisation).

In conclusion, we found no statistical differences between the arms regarding oncological outcomes (LR, DM, OS, RFS). SRT-delay is an alternative with less post-operative complications and higher possibility of pCR compared to SRT. LRT-delay demands more RT-resources without any other obvious gain.