MRI‐based qualitative, quantitative, and radiomics/deep learning methods for assessing treatment response after neoadjuvant chemoradiotherapy in patients with locally advanced rectal cancer

Abstract Neoadjuvant chemoradiotherapy (nCRT) followed by total mesorectal excision is standard treatment for locally advanced rectal cancer (LARC). This approach allows a subset of patients to achieve a pathological complete response (pCR), thereby improving surgical outcomes, anal preservation rates, and disease‐free survival. An accurate preoperative assessment of pCR is crucial for guiding treatment decisions. Magnetic resonance imaging (MRI), owing to its superior soft‐tissue contrast and spatial resolution, has become the preferred noninvasive modality for assessing nCRT efficacy. Advances in functional MRI (fMRI) techniques include diffusion‐weighted imaging, derived sequences, perfusion‐weighted imaging, and neuro‐fMRI. fMRI sequences provide not only a qualitative assessment but also quantitative parameters derived from various imaging principles, thereby significantly enhancing the clinical utility of MRI. Beyond conventional and functional MRI, this field is rapidly evolving with the integration of radiomics and deep learning approaches. Radiomics involves the high‐throughput extraction of minimal quantitative features from medical images, which can reveal tumor heterogeneity and phenotypic characteristics that are invisible to the human eye. This review summarizes current research and future perspectives on MRI‐based qualitative, quantitative, radiomic, and deep‐learning approaches for assessing nCRT efficacy in patients with LARC.