Automated detection of lung nodules in low-dose computed tomography

TL;DR

This paper presents a CAD system for detecting lung nodules in low-dose CT scans, utilizing a 3D filter and neural classifier, achieving high sensitivity with manageable false positives.

Contribution

It introduces a novel distributed database and a CAD system combining 3D filtering and neural classification for improved lung nodule detection.

Findings

Achieved 85% sensitivity at 1-9 false positives per scan.

System performs well on both internal and sub-pleural nodules.

Maintains 75% sensitivity at 1-6 false positives per scan.

Abstract

A computer-aided detection (CAD) system for the identification of pulmonary nodules in low-dose multi-detector computed-tomography (CT) images has been developed in the framework of the MAGIC-5 Italian project. One of the main goals of this project is to build a distributed database of lung CT scans in order to enable automated image analysis through a data and cpu GRID infrastructure. The basic modules of our lung-CAD system, consisting in a 3D dot-enhancement filter for nodule detection and a neural classifier for false-positive finding reduction, are described. The system was designed and tested for both internal and sub-pleural nodules. The database used in this study consists of 17 low-dose CT scans reconstructed with thin slice thickness (~300 slices/scan). The preliminary results are shown in terms of the FROC analysis reporting a good sensitivity (85% range) for both internal and sub-pleural nodules at an acceptable level of false positive findings (1-9 FP/scan); the sensitivity value remains very high (75% range) even at 1-6 FP/scan