Multivisceral resection of advanced colon and rectal cancer: a prospective multicenter observational study with propensity score analysis of the morbidity, mortality, and survival.

Objectives: In the surgical treatment of colorectal carcinoma (CRC), 1 in 10 patients has a peritumorous adhesion or tumor infiltration in the adjacent tissue or organs. Accordingly, multivisceral resection (MVR) must be performed in these patients. This prospective multicenter observational study aimed to analyze the possible differences between non-multivisceral resection (nMVR) and MVR in terms of early postoperative and long-term oncological treatment outcomes. We also aimed to determine the factors influencing overall survival. Methods: The data of 25,321 patients from 364 hospitals who had undergone surgery for CRC (the Union for International Cancer Control stages I-III) during a defined period were evaluated. MVR was defined as (partial) resection of the tumor-bearing organ along with resection of the adherent and adjacent organs or tissues. In addition to the patients' personal, diagnosis (tumor findings), and therapy data, demographic data were also recorded and the early postoperative outcome was determined. Furthermore, the long-term survival of each patient was investigated, and a "matched-pair" analysis was performed. Results: From 2008 to 2015, the MVR rates were 9.9 % (n=1,551) for colon cancer (colon CA) and 10.6 % (n=1,027) for rectal cancer (rectal CA). CRC was more common in men (colon CA: 53.4 %; rectal CA: 62.0 %) than in women; all MVR groups had high proportions of women (53.6 % vs. 55.2 %; pairs of values in previously mentioned order). Resection of another organ frequently occurred (75.6 % vs. 63.7 %). The MVR group had a high prevalence of intraoperative (5.8 %; 12.1 %) and postoperative surgical complications (30.8 % vs. 36.4 %; each p<0.001). Wound infections (colon CA: 7.1 %) and anastomotic insufficiencies (rectal CA: 8.3 %) frequently occurred after MVR. The morbidity rates of the MVR groups were also determined (43.7 % vs. 47.2 %). The hospital mortality rates were 4.9 % in the colon CA-related MVR group and 3.8 % in the rectal CA-related MVR group and were significantly increased compared with those of the nMVR group (both p<0.001). Results of the matched-pair analysis showed that the morbidity rates in both MVR groups (colon CA: 42.9 % vs. 34.3 %; rectal CA: 46.3 % vs. 37.2 %; each p<0.001) were significantly increased. The hospital lethality rate tended to increase in the colon CA-related MVR group (4.8 % vs. 3.7 %; p=0.084), while it significantly increased in the rectal CA-related MVR group (3.4 % vs. 3.0 %; p=0.005). Moreover, the 5-year (yr) overall survival rates were 53.9 % (nMVR: 69.5 %; p<0.001) in the colon CA group and 56.8 % (nMVR: 69.4 %; p<0.001) in the rectal CA group. Comparison of individual T stages (MVR vs. nMVR) showed no significant differences in the survival outcomes (p<0.05); however, according to the matched-pair analysis, a significant difference was observed in the survival outcomes of those with pT4 colon CA (40.6 % vs. 50.2 %; p=0.017). By contrast, the local recurrence rates after MVR were not significantly different (7.0 % vs. 5.8 %; both p>0.05). The risk factors common to both tumor types were advanced age (>79 yr), pT stage, sex, and morbidity (each hazard ratio: >1; p<0.05). Conclusions: MVR allows curation by R0 resection with adequate long-term survival. For colon or rectal CA, MVR tended to be associated with reduced 5-year overall survival rates (significant only for pT4 colon CA based on the MPA results), as well as, with a significant increase in morbidity rates in both tumor entities. In the overall data, MVR was associated with significant increases in hospital lethality rates, as indicated by the matched-pair analysis (significant only for rectal CA).