Abstract Archives of the RSNA, 2007
VP21-05
Computer-aided Detection for the Identification of Pulmonary Nodules in Pediatric Oncology Patients
Scientific Papers
Presented on November 26, 2007
Presented as part of VP21: Pediatric Series: Pediatric Chest
Emma Jane Helm MBBS, Presenter: Nothing to Disclose
Heidi Roberts, Abstract Co-Author: Research Consultant, R2 Technology, Inc, Sunnyvale, CA
Research Consultant, Medicsight, Inc, London, UK
Cicero J. Torres A. Silva MD, Abstract Co-Author: Nothing to Disclose
Paul S. Babyn MD, Abstract Co-Author: Nothing to Disclose
Computer Aided Detection (CAD) has been shown to increase the sensitivity for detection of pulmonary nodules in adults. This study reports initial findings utilizing a CAD system for the detection of pediatric pulmonary nodules.
CT scans from 17 pediatric patients with known primary tumors and lung nodules were analyzed. Scans were performed with slice collimation of 5 mm reconstructed at 2.5 mm or less. Initial reading was performed by 3 radiologists and comparison made with the CAD analysis. The CAD system is designed to detect nodules 4mm and above. Maximum nodule diameter and location were recorded. Location was classified as pleural (abutting the pleura), peripheral (less than 1cm from the pleura), central (more than 1cm from the pleura and more than 1cm from the hilum) and hilar (less than 1cm from the hilar vessels). Sensitivities were calculated for detection according to nodule size and location.
105 nodules were identified in 17 patients (age 3-17 years). Primary tumors were Ewing sarcoma (6), osteosarcoma (4), Wilms tumour (2), rhabdomyosarcoma (3), hepatoblastoma (1) and acinic cell tumor (1). 6 patients had solitary nodules and 11 had multiple nodules. Median nodule size was 3.4mm, range 1 to 19mm. 40 (38%) nodules were 4.0mm or greater and 65 (62%) were 3.9mm or less. 26 (24.8%) nodules were in contact with the pleura, 51 nodules (48.6%) were within 1cm of the pleura, 23 nodules (21.9%) were centrally located and 3 nodules (2.9%) were within 1cm of the hilum.
The sensitivity of CAD for detection of nodules 4.0mm or more was 79.5%. The sensitivity for detection of nodules 3.9mm or less was 18.5%. Overall sensitivity was 41.0%. Sensitivity of CAD was 50% for pleurally based nodules, 35.3% for peripheral nodules, 43.5% for central nodules and 33.3% for hilar nodules.
There were 16 false positives giving an average of 0.9 per patient.
CAD in pediatric oncology patients has a sensitivity of 79.5% for detection of lung nodules 4mm and above with a low number of false positives. However, the sensitivity is significantly less for nodules below 4mm.
CAD is a promising technique for detection of pulmonary nodules in pediatric oncology patients.
Helm, E,
Roberts, H,
Silva, C,
Babyn, P,
Computer-aided Detection for the Identification of Pulmonary Nodules in Pediatric Oncology Patients. Radiological Society of North America 2007 Scientific Assembly and Annual Meeting, November 25 - November 30, 2007 ,Chicago IL.
http://archive.rsna.org/2007/5004808.html