LungEvaty: A Scalable, Open-Source Transformer-based Deep Learning Model for Lung Cancer Risk Prediction in LDCT Screening

2025-11-25

AI Summary

To address the scalability–performance trade-off in whole-lung modeling for low-dose CT (LDCT) lung cancer screening, this paper proposes the first end-to-end pure Transformer framework capable of processing full-volume lung CT inputs and trained on large-scale pixel-label-free data. The method introduces an anatomy-informed attention guidance (AIAG) mechanism that captures malignancy-related anatomical–pathological correlations without region-level supervision. By integrating self-supervised pretraining with optional AIAG loss, it enables risk prediction using imaging data alone. Trained and validated on over 90,000 CT scans, the model is fine-tuned on a 28,000-case dataset and achieves state-of-the-art performance on a 6,000-case independent test set. It demonstrates significantly improved generalizability and clinical utility in large-scale LDCT screening scenarios.

Technology Category

• Computer Vision: Large Vision Models
• Machine Learning: Large Multimodal Models (LMMs)
• Cognitive Modeling & Cognitive Systems: Agent Architectures

Application Category

• Search and Retrieval-Augmented AI: Agentic search
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Lung cancer risk estimation is gaining increasing importance as more countries introduce population-wide screening programs using low-dose CT (LDCT). As imaging volumes grow, scalable methods that can process entire lung volumes efficiently are essential to tap into the full potential of these large screening datasets. Existing approaches either over-rely on pixel-level annotations, limiting scalability, or analyze the lung in fragments, weakening performance. We present LungEvaty, a fully transformer-based framework for predicting 1-6 year lung cancer risk from a single LDCT scan. The model operates on whole-lung inputs, learning directly from large-scale screening data to capture comprehensive anatomical and pathological cues relevant for malignancy risk. Using only imaging data and no region supervision, LungEvaty matches state-of-the-art performance, refinable by an optional Anatomically Informed Attention Guidance (AIAG) loss that encourages anatomically focused attention. In total, LungEvaty was trained on more than 90,000 CT scans, including over 28,000 for fine-tuning and 6,000 for evaluation. The framework offers a simple, data-efficient, and fully open-source solution that provides an extensible foundation for future research in longitudinal and multimodal lung cancer risk prediction.

Problem

• Predicts 1-6 year lung cancer risk from single LDCT scans
• Processes whole-lung volumes without pixel-level annotations
• Provides scalable open-source framework for screening programs

Innovation

• Transformer-based deep learning model for lung cancer risk
• Processes whole-lung inputs without region supervision
• Uses Anatomically Informed Attention Guidance loss function