Development and internal validation of a prognostic nomogram for lung metastasis in locally advanced rectal cancer incorporating molecular markers.

Background: Prediction models for lung metastasis in rectal cancer are limited, and gene mutations have not been incorporated. As genetic testing becomes more common in routine clinical practice, including RAS/BRAF mutation status in lung metastasis risk prediction models may enhance their clinical relevance. This study aimed to identify gene mutations and clinicopathological factors associated with lung metastasis in locally advanced rectal cancer (LARC) patients treated with neoadjuvant chemoradiotherapy (nCRT) followed by total mesorectal excision (TME). Methods: A total of 311 consecutive LARC patients who underwent genetic testing and received nCRT followed by tumor curative resection between April 2014 and August 2023 were retrospectively analyzed. The diagnosis of lung metastasis mainly relied on chest computed tomography (CT) findings indicative of metastatic lesions. Predictors were preliminarily selected using Least Absolute Shrinkage and Selection Operator (LASSO) analysis. A prediction model was then developed via multivariate logistic regression and visualized as a nomogram. The model's performance was evaluated using the receiver operating characteristic (ROC) curve, calibration curves, and decision curve analysis (DCA), and internally validated using the bootstrap validation method. Results: The study cohort included 311 patients with a median age of 55.3 years old (range, 26-76 years old) and 67.5% males. During a median 28-month follow-up period (range, 3-112 months), lung metastasis occurred in 47 patients (15.1%) after surgery. Tumor location, lymphovascular invasion, pathological N stage and RAS mutation were identified as significant factors associated with the risk of lung metastasis. The prediction model had good discrimination, with an area under the receiver operating characteristic curve (AUC) of 0.772. The nomogram's sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were 83.0%, 56.8%, 25.5%, and 94.9%, respectively, at a cutoff value of 142. Conclusions: We identified the risk factors for lung metastasis in LARC patients treated with nCRT and TME and developed a nomogram to facilitate individualized imaging and follow-up strategies. Further studies with larger sample size and external validation are warranted.