A Nomogram for Predicting Lymph Node Metastasis in Submucosal Colorectal Cancer

Abstract

In colorectal cancer (CRC), the possibility of lymph node (LN) metastasis is an important consideration when deciding on treatment. We developed a nomogram for predicting lymph node metastasis of submucosal (SM) CRC. The medical records of 509 patients with SM CRC from 1984 to 2012 were retrospectively investigated. All the patients underwent curative surgical resection at the Osaka Medical Center for Cancer and Cardiovascular Diseases. A total 113 patients with inadequate data were excluded. Using a group of 293 patients who underwent surgery from 1984 to 2008, a logistic regression model was used to develop a prediction model for LN metastasis. The prediction model was validated in an additional group of 103 patients who underwent surgery from 2009 to 2012. Univariate analysis of pathologic factors showed the influence of low histologic grade (muc, por, sig; P < 0.001), positive lymphatic invasion (P < 0.001), positive vascular invasion (P = 0.036), and tumor SM invasion depth (P = 0.098) in LN metastasis. Using these variables, a nomogram predicting LN metastasis was constructed using a logistic regression model with an area under the curve (AUC) of 0.717. The prediction model was validated by an external dataset in an independent patient group with an AUC of 0.920. We developed a novel and reliable nomogram predicting LN metastasis through the integration of 4 pathologic factors. This prediction model may help clinicians to decide on personalized treatment following endoscopic resection.