Rectal cancer surgery: Prognostic Factors Related to Treatment
MetadataVis full innførsel
Background Towards the end of the 80-ties, 28 % of the rectal cancer patients treated for cure developed local recurrence, and only 55 % survived five years. Since 1993, intensified focus on the rectal cancer patients has improved the standards for diagnostics and treatment at a national level. Implementation of total mesorectal excision (TME) and selective application of preoperative radiotherapy (RT) have made the local recurrence rate approaching 9 % and the overall survival rate 71 % at a national level. Continuous research on behalf of the Norwegian Gastrointestinal Cancer Group (NGICG) have resulted in national guidelines and provided knowledge for the clinicians responsible for the treatment of rectal cancer patients. Some of the risk factors in the rectal cancer patient have strong impact on the prognosis. In contrast to risk factors such as T and N-status, and differentiation, circumferential resection margin (CRM), distal resection margin (DRM) and the presence of synchronous colon cancers are strongly influenced by the decision making and skills of the surgeon. Thus, gathering further knowledge on these risk factors is valuable in the continuous approach towards optimised treatment for rectal cancer patients. In the international literature on rectal cancer during the last twenty years, local recurrence has been the dominating endpoint. There has been less focus on how the risk for distant metastases has changed over the years. Aim In contrast to most countries offering modern rectal cancer treatment today, few patients in Norway have had adjunctive treatment. This has made it possible to study the “nature of the raw cancer biology”. The aim of the study was to gather further knowledge on CRM and DRM. In a large national cohort it should be possible to explore the importance and the lack of importance of these risk factors in several clinical settings, i.e. for different T-stages or for different tumour levels from anal verge. Furthermore, the aim was also to explore the role of these risk factors in the prediction of prognosis following neoadjuvant treatment. The aim of the study was also to explore whether the efforts on improving the quality of the local treatment in the pelvis had any impact on the risk of development of distant metastases. The last aim was to explore the presence of synchronous colon cancer and the importance of the increased tumour burden on the prognosis of rectal cancer patients. Material and methods All rectal cancer patients treated with major surgery for cure in Norway with or without preoperative radiotherapy between 1993 and 2004 were included in Paper I and Paper II. In Paper I, all 3196 rectal cancer patients with information on CRM were analysed. In Paper II, the 3571 patients treated with low anterior resection (AR) and information on DRM were analysed. Paper III included the 6501 rectal cancer patients treated for cure between 1993 and 2006. In Paper IV, the patients with R2 status were also included to define the incidence of synchronous colon cancer. Altogether, 7605 patients with solitary rectal cancer were compared with the 327 rectal cancer patients with synchronous colon cancer in the main analyses. This was an open cohort study, thus patients have been included continuously since the initiation of the national database in 1993. Data have been collected on project specific forms and date and cause of death have been obtained from the Norwegian Cause of Death Registry. The completeness of the registry is thought to be close to 100 %. In order to characterize the cohort of each project, Chi-square method was used for association between subgroups. The Kaplan-Meier method was used for survival alanyses. Cox proportional hazrds regression was used for the multivariable analyses. Results After analysing the rates of local recurrences for each millimetres of CRM, a short CRM was defined as CRM ≤ 2 mm in Papers I,II and III. CRM ≤ 2 mm was associated with increased risk of local recurrence, distant metastases and death for patients treated without RT. A DRM ≤ 10 mm was defined as short DRM and was associated with increaesed risk of local recurrence, but not for distant metastases or death. Increased risk of local recurrence was observed among T2 and T3 tumours when CRM was ≤ 2 mm. The role of CRM was unclear among the T4 tumours, however the rates of local recurrence was high among patients with both narrow and wide lateral resection margin. Short CRM and DRM were associated with increased risks of local recurrence when tumours were localized between 6 and 15 cm from anal verge. In Paper I, CRM ≤ 2 mm had an unclear impact on local recurrence when tumour was located 0-5 cm from anal verge, but considering patients with distal tumours in Paper II treated with AR only, though few patients, short CRM and DRM might influence on local recurrence. However, nodal disease was associated with high risk of local recurrence independent of distance from anal verge. Short CRM and residual nodal disease after preoperative RT were associated with increased risk of local recurrence, distant metastases and decreased overall survival. None of the patients was evaluated with pelvic MRI and only 4.5 % were treated with preoperative RT during the first years (1993-1997) of the study. Short CRM was measured in 22 % of the patients. Between 2004 and 2006, more than 4/5 of the patients were evaluated with MRI and more than 1/5 had preoperative RT. During this period, 6.7 % of the patients had inadvertent perforation versus 10 % during the previous period. Only 12 % of the patients during the last period had short CRM. Comparing the first and the last period, local recurrence rate decreased from 13 % to 8 %. A strong association between local recurrence and distant metastases was found, and at a national level, the four-year distant metastasis rate was reduced from 25.2 % to 18.5 %. According to the present thesis, the decrease in distant metastases is thought to be multifactorial, but it will probably be influenced by the reduction of short CRM and the increased use of neoadjuvant therapy. Synchronous colon cancers were found among 3.4% of the Norwegian rectal cancer patients. One third of these had the synchronous colon cancer proximal and 2/3 distal to the splenic flexure. The proximal colon cancers were more advanced and more often had poor differentiation compared to the distal synchronous colon cancers. Comparing patients with a solitary rectal cancer and patients with a rectal cancers and synchronous proximal colon cancer, the latter had inferior survival. They also had a shorter lifespan after the diagnosis of distant metastases. Conclusions The role of short CRM has been explored in different clinical settings, and CRM ≤ 2 mm is associated with inferior prognosis both for the non-irradiated as well as for the radiated patients. The impact of short CRM on local recurrence is strong for patients with T2 and T3 tumours between 6 and 15 cm from anal verge. Short DRM, a distal resection margin ≤ 10 mm in the pathological specimen is associated with increased risk of local recurrence when sphincter preserving surgery is performed. Improved quality on diagnostics and treatment for the primary tumour in the pelvis has also reduced the risk of distant metastases. Patients with rectal cancer and proximal synchronous colon cancers have inferior survival compared with patients with a solitary rectal cancer only.
UtgiverNorges teknisk-naturvitenskapelige universitet, Det medisinske fakultet, Institutt for kreftforskning og molekylær medisin
SerieDoktoravhandlinger ved NTNU, 1503-8181; 2012:284
Dissertations at the Faculty of Medicine, 0805-7680; 576