Baseline MRI habitat imaging for predicting treatment response to neoadjuvant chemoradiotherapy in locally advanced rectal cancer

Motivation: Habitat imaging analysis is an innovative radiomics technology that identifies tumor subregions with similar characteristics through quantitative imaging markers. Currently, there are no relevant research predicting the efficacy of neoadjuvant chemoradiotherapy (nCRT) for locally advanced rectal cancer (LARC). Goal(s): To investigate the value of MRI-based habitat imaging in predicting the efficacy of nCRT in cases of LARC. Approach: Employed K-means to segment the entire tumor ROI based on the combined data from the T2WI and ADC maps, and developed a model using habitat features. Results: The habitat model we developed accurately predicts the efficacy of nCRT. Impact: Habitat models built exclusively with first-order histogram features provide a more accurate representation of biological characteristics. Additionally, a hybrid model was developed by integrating clinical features with habitat features, enhancing model interpretability and acceptance by radiologists.