Development of a model to predict pathologic response to chemotherapy in patients with colorectal liver metastases.

Abstract

Background\nPreoperative chemotherapy has widely been used in colorectal cancer liver metastasis (CRLM). Pathological response to chemotherapy is very important in evaluating tumor biology. However, there is still a lack of a non-invasive and accurate method to evaluate pathological response before surgery.\n\n\nMethods\nWe retrospectively analyzed the clinicopathologic data of patients with CRLM who underwent liver resection after preoperative chemotherapy between January 2006 and December 2018. Pathological responses were defined as minor when there are ≥50% remnant viable cells and as major when 0-49% remnant viable cells exist.\n\n\nResults\nA total of 482 patients were included and randomly divided into training (n=241) and validation (n=241) cohorts. The proportion of major pathologic response was similar between the two groups (51.5% and 48.5%). Multivariate analysis determined the disease-free interval (DFI), tumor size, tumor number, and RAS status as independent predictors of major pathologic response to preoperative chemotherapy. The nomogram incorporating these variables showed good concordance statistics in the training cohort (0.746, 95% CI: 0.685-0.807) and validation cohort (0.764, 95% CI: 0.704-0.823). In addition, the nomogram showed good applicability in patients with different characteristics.\n\n\nConclusions\nThe established nomogram model performed well in predicting pathological response in patients with CRLM.