Spatial patterns and MRI-based radiomic prediction of high peritumoral tertiary lymphoid structure density in hepatocellular carcinoma: a multicenter study

• DOI: 10.1136/jitc-2024-009879
• Publication Date: 2024-12-16T01:50:11Z
• Authors: Shichao Long; Mengsi Li; Juan Chen; Linhui Zhong; Aerzuguli Abudulimu; Lan Zhou; Wenguang Liu; Deng Pan; Ganmian Dai; Kai Fu; Xiong Chen; Yigang Pei; Wenzheng Li
• Keywords: 0303 health sciences; 03 medical and health sciences; Immunotherapy Biomarkers; Neoplasms. Tumors. Oncology. Including cancer and carcinogens; RC254-282

Tertiary lymphoid structures (TLS) within the tumor microenvironment have been associated with cancer prognosis and therapeutic response. However, immunological pattern of a high peritumoral TLS (pTLS) density its clinical potential in hepatocellular carcinoma (HCC) remain poor. This study aimed to elucidate biological differences related pTLS develop radiomic classifier for predicting HCC, offering new insights diagnosis treatment. Spatial transcriptomics (n=4) RNA sequencing data (n=952) were used identify critical regulators evaluate their prognostic significance HCC. Baseline MRI images from 660 patients HCC who had undergone surgery treatment between October 2015 January 2023 retrospectively recruited model development validation. included training (n=307) temporal validation (n=76) cohorts Xiangya Hospital, external three independent hospitals (n=277). Radiomic features extracted intratumoral regions interest analyzed using machine learning algorithms predictive classifier. The classifier's performance was evaluated area under curve (AUC), value assessed across four dual-center outcome cohort 41 received immunotherapy. Patients experienced prolonged median overall survival (p<0.05) favorable immunotherapy response (p=0.03). Moreover, immune infiltration by mature B cells observed region. pseudotime analysis immunohistochemistry staining revealed that expansion elevated CXCL9 CXCL10 co-expression. We developed an optimal radiomic-based excellent discrimination density, achieving AUC 0.91 (95% CI 0.87, 0.94) cohort. also exhibited promising stratification ability terms (p<0.01), relapse-free (p<0.05), (p<0.05). identified key proposed non-invasive capable assisting