Fig. 3: Kaplan-Meier Curves for LNM risk groups (temporal validation set). | Communications Medicine

Fig. 3: Kaplan-Meier Curves for LNM risk groups (temporal validation set).

From: Predicting lymph node metastasis from primary tumor histology and clinicopathologic factors in colorectal cancer using deep learning

Fig. 3: Kaplan-Meier Curves for LNM risk groups (temporal validation set).

Kaplan-Meier Curves for disease specific survival (DSS) amongst stage II and stage III cases, respectively. A logistic regression model using the five machine-learned features identified for LNM prediction was fit on the development cohort for predicting DSS. Risk groups were defined within each stage by binarizing using the median LNM model score for the most recent 5-years of the development cohort (2003 to 2007). The resulting regression models and risk group thresholds were then evaluated on the temporal validation set. The logistic regression model provided significant risk stratification within both node-negative and node-positive disease groups, suggesting the potential of such a model to aid in improving prognostication and therapeutic decision making.

Back to article page