Vision Transformers for Preoperative CT-Based Prediction of Histopathologic Chemotherapy Response Score in High-Grade Serous Ovarian Carcinoma

2026-04-10

AI Summary

This study addresses the clinical challenge of preoperatively assessing histopathological chemotherapy response scores (CRS) in patients with high-grade serous ovarian cancer, which is critical yet currently inaccessible for guiding multidisciplinary treatment decisions. To this end, the authors propose a novel 2.5D multimodal deep learning framework that, for the first time, leverages Vision Transformers (ViTs) to analyze preoperative CT images of omental deposits. The approach employs a pretrained ViT encoder and incorporates an intermediate fusion module to integrate imaging features with clinical variables, enabling non-invasive early prediction of CRS. Evaluated on an internal test set, the model achieves a ROC-AUC of 0.95, accuracy of 95%, and precision of 80%; it demonstrates preliminary generalizability on an external cohort with a ROC-AUC of 0.68, establishing a new paradigm for effective fusion of radiological and clinical data in oncology.

Technology Category

• Computer Vision: Multi-modal Vision
• Machine Learning: Multimodal Learning
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Responsible Web: Algorithmic accountability and transparency on the web

Abstract

Purpose. High-grade serous ovarian carcinoma (HGSOC) is characterized by pronounced biological and spatial heterogeneity and is frequently diagnosed at an advanced stage. Neoadjuvant chemotherapy (NACT) followed by delayed primary surgery is commonly employed in patients unsuitable for primary cytoreduction. The Chemotherapy Response Score (CRS) is a validated histopathological biomarker of response to NACT, but it is only available postoperatively. In this study, we investigate whether pre-treatment computed tomography (CT) imaging and clinical data can be used to predict CRS as an investigational decision-support adjunct to inform multidisciplinary team (MDT) discussions regarding expected treatment response. Methods. We proposed a 2.5D multimodal deep learning framework that processes lesion-dense omental slices using a pre-trained Vision Transformer encoder and integrates the resulting visual representations with clinical variables through an intermediate fusion module to predict CRS. Results. Our multimodal model, integrating imaging and clinical data, achieved a ROC-AUC of 0.95 alongside 95% accuracy and 80% precision on the internal test cohort (IEO, n=41 patients). On the external test set (OV04, n=70 patients), it achieved a ROC-AUC of 0.68, alongside 67% accuracy and 75% precision. Conclusion. These preliminary results demonstrate the feasibility of transformer-based deep learning for preoperative prediction of CRS in HGSOC using routine clinical data and CT imaging. As an investigational, pre-treatment decision-support tool, this approach may assist MDT discussions by providing early, non-invasive estimates of treatment response.

Problem

• Chemotherapy Response Score
• High-Grade Serous Ovarian Carcinoma
• Preoperative Prediction
• CT Imaging
• Neoadjuvant Chemotherapy

Innovation

• Vision Transformer
• multimodal deep learning
• chemotherapy response prediction
• preoperative CT imaging
• 2.5D fusion