# Development and validation of a CT-based habitat radiomics model for predicting pathological grading in non-small cell lung cancer

## TL;DR

This study develops a CT-based radiomics model to predict the pathological grading of non-small cell lung cancer, showing improved accuracy and specificity using habitat radiomics features.

## Contribution

The novel contribution is the development of a habitat radiomics model that outperforms traditional radiomics and clinical features in predicting NSCLC grading.

## Key findings

- Habitat radiomics achieved an AUC of 0.89 and 0.87 for predicting Grades 3 vs. 1–2 and 1 vs. 2, respectively.
- The combined model (Clf Total) achieved the highest AUC of 0.91 and 0.88 for the two grading models.
- The multimodal model showed high specificity (0.84 and 0.77) and balanced accuracy (0.82 and 0.81).

## Abstract

To develop and validate models for predicting pathological grading of non-small cell lung cancer (NSCLC) using habitat radiomics and clinical semantic features.

In this retrospective study of 800 NSCLC patients, a whole tumor volume (WTV) was delineated by applying a 3 mm expansion to the gross tumor volume (GTV) on non-contrast CT scans. Habitat subregions within the WTV were identified using K-means clustering. A two-step binary classification model was constructed to predict pathological grades: Model-1 distinguished Grade 3 from combined Grades 1–2, and Model-2 further differentiated Grade 1 from Grade 2. Predictive models were built with logistic regression based on four distinct feature sets: WTV radiomics (Clf WVOI), habitat radiomics (Clf Habitats), clinical features (Clf Clinical), and a combined feature set (Clf Total).

In both Model-1 and Model-2, the classification performance of Clf Habitats was generally superior to that of Clf WVOI and Clf Clinical, achieving an AUC of 0.89 and 0.87, specificity of 0.73 for both models, and BACC of 0.78 and 0.79, respectively, on the test set. The combined model, Clf Total, achieved the best predictive performance on the test set, with AUC values of 0.91 and 0.88, specificity of 0.84 and 0.77, and BACC of 0.82 and 0.81.

Habitat radiomics significantly improves NSCLC pathological grading. The multimodal model offers robust performance and high specificity, aiding personalized treatment planning.

Left: Schematic of the multicenter retrospective study and methodological pipeline, including K means–based habitat segmentation, radiomics feature extraction, and a two step classification framework. Right: Comparison of four classifiers shows that integrating habitat radiomics improves prediction of NSCLC pathological grading.Infographic illustrating a retrospective study on a CT-based habitat radiomics model for predicting pathological grading in non-small cell lung cancer, detailing participant sources, classification models, feature extraction methods, performance ROC curves for four classifiers, and the conclusion that habitat radiomics enhances grading accuracy and specificity.

Left: Schematic of the multicenter retrospective study and methodological pipeline, including K means–based habitat segmentation, radiomics feature extraction, and a two step classification framework. Right: Comparison of four classifiers shows that integrating habitat radiomics improves prediction of NSCLC pathological grading.