Rethinking Lung Cancer Screening: AI Nodule Detection and Diagnosis Outperforms Radiologists, Leading Models, and Standards Beyond Size and Growth

2025-11-28

AI Summary

Current lung cancer screening relies on nodule size and growth rate, leading to delayed detection of early-stage malignant nodules and facing bottlenecks in data scale, interpretability, and diagnostic accuracy. To address this, we propose the first nodule-level joint detection and malignancy diagnosis framework that bypasses traditional temporal dependencies and enables end-to-end diagnosis directly from a single low-dose CT scan. Our method integrates shallow deep learning with domain-informed feature engineering, trained and validated on 25,709 CT scans containing 69,449 annotated nodules. The model achieves an internal AUC of 0.98 and an independent-test AUC of 0.945; at 0.5 false positives per scan, sensitivity reaches 99.3%. It significantly outperforms established benchmarks—including Lung-RADS and Sybil—across all nodule sizes, clinical stages, and growth patterns, and can identify challenging or slow-growing malignant nodules approximately one year earlier than current standards.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Computer Vision: Multi-modal Vision
• Search and Optimization: Learning to Search

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work
• Security and Privacy: Large-scale security measurements

Abstract

Early detection of malignant lung nodules is critical, but its dependence on size and growth in screening inherently delays diagnosis. We present an AI system that redefines lung cancer screening by performing both detection and malignancy diagnosis directly at the nodule level on low-dose CT scans. To address limitations in dataset scale and explainability, we designed an ensemble of shallow deep learning and feature-based specialized models. Trained and evaluated on 25,709 scans with 69,449 annotated nodules, the system outperforms radiologists, Lung-RADS, and leading AI models (Sybil, Brock, Google, Kaggle). It achieves an area under the receiver operating characteristic curve (AUC) of 0.98 internally and 0.945 on an independent cohort. With 0.5 false positives per scan at 99.3% sensitivity, it addresses key barriers to AI adoption. Critically, it outperforms radiologists across all nodule sizes and stages, excelling in stage 1 cancers, and all growth-based metrics, including the least accurate: Volume-Doubling Time. It also surpasses radiologists by up to one year in diagnosing indeterminate and slow-growing nodules.

Problem

• AI system detects and diagnoses lung nodules earlier than size-based methods
• Overcomes dataset and explainability limits with ensemble deep learning models
• Outperforms radiologists and existing AI in accuracy across all nodule stages

Innovation

• AI system detects and diagnoses lung nodules directly
• Ensemble of shallow deep learning and feature-based models
• Outperforms radiologists and leading AI models in screening