Tri-Reader: An Open-Access, Multi-Stage AI Pipeline for First-Pass Lung Nodule Annotation in Screening CT

TL;DR

Tri-Reader is a multi-stage AI pipeline that combines lung segmentation, nodule detection, and malignancy classification, aiming to improve lung nodule annotation in screening CT scans with high sensitivity and broad applicability.

Contribution

It introduces a publicly accessible, integrated pipeline leveraging open-access models for comprehensive lung nodule annotation in screening CTs.

Findings

High sensitivity in nodule detection across datasets

Effective reduction of candidate annotations for radiologists

Comparable performance to expert annotations

Abstract

Using multiple open-access models trained on public datasets, we developed Tri-Reader, a comprehensive, freely available pipeline that integrates lung segmentation, nodule detection, and malignancy classification into a unified tri-stage workflow. The pipeline is designed to prioritize sensitivity while reducing the candidate burden for annotators. To ensure accuracy and generalizability across diverse practices, we evaluated Tri-Reader on multiple internal and external datasets as compared with expert annotations and dataset-provided reference standards.