DR-CSI for In Vivo Tumor Heterogeneity and Lymph Node Metastasis Prediction in Oral Tongue Squamous Cell Carcinoma

Motivation: Preoperative assessment of cervical lymph node metastasis (CLNM) is critical for prognosis and treatment in oral tongue squamous cell carcinoma (OTSCC), but current imaging methods lack the precision to capture tumor microstructure accurately. Goal(s): We aimed to assess the feasibility of DR-CSI in characterizing tumor heterogeneity and predicting CLNM in OTSCC. Approach: We enrolled 57 OTSCC patients who underwent DR-CSI, segmenting D-T2 spectra into five subregions. We compared DR-CSI volume fractions to differentiate normal tissue, CLNM-negative, and CLNM-positive groups. Results: DR-CSI metrics, especially VB and VD, distinguished CLNM-positive cases and outperformed ADC in predicting CLNM. Impact: DR-CSI enables detailed in vivo mapping of tumor microstructure in OTSCC, facilitating precise lymph node metastasis prediction and offering a powerful tool for non-invasive risk stratification. Its use could improve staging accuracy and optimize treatment planning for OTSCC.