A unified vision-language model for precision oncology and biomarker prediction in neuroblastoma

2026-07-09 · Source: Machine learning : nature.com subject feeds · Field: Health & Wellbeing — Medical Devices & Health Technology, Health & Medical Research, Artificial Intelligence & Machine Learning · Depth: Expert, medium

Summary

NEVA, a novel multimodal foundation model, addresses challenges in neuroblastoma management by integrating vision and language for precision oncology and biomarker prediction. Developed and evaluated across 1,238 patients from multiple institutions, NEVA employs a pathologist-inspired hierarchical workflow with end-to-end optimization, distinguishing it from conventional methods using frozen encoders. The model surpasses ten existing foundation models, including TITAN, UNI, and Virchow, across most of 11 clinical tasks. NEVA demonstrates strong diagnostic capabilities, achieving AUROC scores of 0.916 for subtype classification, 0.823 for Shimada classification, and 0.806 for risk group stratification. It also accurately predicts molecular alterations like *NMYC* amplification (AUROC 0.924) and 1p36 deletion (AUROC 0.830), while enabling prognostic stratification for survival.

Key takeaway

For oncologists and computational pathologists developing precision medicine strategies for neuroblastoma, NEVA offers a robust AI framework to enhance diagnostic accuracy and prognostic stratification. Your team can leverage its ability to predict key molecular alterations and risk groups directly from routine pathology data, potentially streamlining patient management. Consider integrating such multimodal foundation models to improve clinical decision support and overcome limitations of traditional molecular profiling.

Key insights

NEVA is a multimodal vision-language AI for neuroblastoma diagnosis and biomarker prediction using a pathologist-inspired workflow.

Principles

• Integrate vision and language for comprehensive oncology insights.
• Mimic expert workflows for improved model architecture.
• End-to-end optimization surpasses frozen encoder methods.

Method

NEVA utilizes a pathologist-inspired hierarchical workflow with end-to-end optimization, moving beyond conventional frozen encoders and multiple instance learning to process multimodal data for neuroblastoma analysis.

In practice

• Predict *NMYC* amplification and 1p36 deletion from pathology.
• Stratify neuroblastoma risk and support clinical decisions.
• Localize histologically relevant regions via attention maps.

Topics

• Neuroblastoma
• Vision-Language Models
• Precision Oncology
• Biomarker Prediction
• Pathology AI
• Foundation Models

Best for: AI Scientist, Research Scientist, Domain Expert

Related on AIssential

• I Gave a Model Eyes and a Genome. Here’s What It Learned About Brain Cancer. — Cross-attention fusion of pathology and genomics significantly improves glioblastoma survival prediction over unimodal approaches.
• Intelligent Healthcare Imaging Platform: A VLM-Based Framework for Automated Medical Image Analysis and Clinical Report Generation — A VLM-based framework automates medical image analysis and clinical report generation across multiple imaging modalities.
• GigaTIME: Scaling tumor microenvironment modeling using virtual population generated by multimodal AI — GigaTIME uses multimodal AI to convert common H&E slides into virtual mIF images, enabling population-scale precision immuno-oncology research.
• A deep learning–based digital biopsy for predicting early recurrence in gastric cancer — The RSA model combines deep learning with clinical data to predict gastric cancer recurrence.
• Learning neuroimaging models from health system-scale data — Health system-scale data enables AI foundation models to achieve superior, fair, and explainable neuroimaging diagnostics.
• AI Improving Dose Decisions and Patient Outcomes in Oncology- with Shefali Kakar of Novartis — AI-driven data integration refines oncology dose decisions, safety, and patient variability, accelerating drug development.

See Counsel's argued verdicts on the open AI decisions leaders are weighing →

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Machine learning : nature.com subject feeds.