Haili Chui PhD, Abstract Co-Author: Nothing to Disclose

Feng Ma, Presenter: Nothing to Disclose

Alex Schneider, Abstract Co-Author: Nothing to Disclose

Susan Alyson Wood PhD, Abstract Co-Author: Nothing to Disclose

Philip F. Judy PhD, Abstract Co-Author: Nothing to Disclose

To evaluate the performance of a CAD system that can provide automatic monitoring/management (detection, tracking and measurement) of lung nodules for temporal MDCT scans.

A CAD system (R2 Technology, Sunnyvale, CA) is designed to help monitor/manage lung nodules present in temporal MDCT scans. It includes a number of streamlined automatic processing modules: lung nodule detection, nodule correspondence tracking and nodule volume measurement. The system was evaluated using a clinical data set comprised of 40 (including 8 metastatic cases) current-prior pairs of MDCT chest examinations (acquired 1-18 months apart) using 1.0-2.5 mm collimations. Double reading by radiologists identified 88 nodules that were present in both current and prior scans. These nodules were determined to require intervention or surveillance. The CAD system was deployed to study its performance in detecting this set of lung nodules and its success rate in tracking their correspondences (and subsequently measuring their volumes and growth rates). A study on phantom nodule data (using phantom spheres and cubes in the size range of [3mm, 10mm] acquired under different exposure settings) was also done to validate the accuracy and reproducibility of the automatic volume measurements.

In the clinical study, 76 nodules (76/88=86%) were successfully detected by the CAD system in both current and prior scans; and, 82 nodules (82/88=93%) were detected in at least one of the two scans. The CAD system successfully tracked all 76 (76/76=100%) nodules that were detected in both scans.In the phantom study, the mean relative bias and the related standard deviation of CAD volume measurement for each phantom nodule were both less than 2%. No statistically significant change in accuracy or reproducibility of volume measurement was observed due to exposure variation.

The studies demonstrated that an automated lung nodule monitoring/management system has the potential to accurately detect, track and measure lung nodules in chest MDCT examinations.

H.C.,F.M.,S.A.W.,A.S.: H. Chui, F. Ma, A. Schneider and S. Wood are employees at R2 Technology, Inc.

Chui, H, Ma, F, Schneider, A, Wood, S, Judy , P, Performance Evaluation of an Automatic Lung Nodules Monitoring/Management System.  Radiological Society of North America 2004 Scientific Assembly and Annual Meeting, November 28 - December 3, 2004 ,Chicago IL. http://archive.rsna.org/2004/4410047.html