1
|
Salman R, Nguyen HN, Sher AC, Hallam KA, Seghers VJ, Sammer MBK. Diagnostic performance of artificial intelligence for pediatric pulmonary nodule detection in computed tomography of the chest. Clin Imaging 2023; 101:50-55. [PMID: 37301051 DOI: 10.1016/j.clinimag.2023.05.019] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 02/10/2023] [Revised: 05/26/2023] [Accepted: 05/31/2023] [Indexed: 06/12/2023]
Abstract
PURPOSE To test the performance of a commercially available adult pulmonary nodule detection artificial intelligence (AI) tool in pediatric CT chests. METHODS 30 consecutive chest CTs with or without contrast of patients ages 12-18 were included. Images were retrospectively reconstructed at 3 mm and 1 mm slice thickness. AI for detection of lung nodules in adults (Syngo CT Lung Computer Aided Detection (CAD)) was evaluated. 3 mm axial images were retrospectively reviewed by two pediatric radiologists (reference read) who determined the location, type, and size of nodules. Lung CAD results at 3 mm and 1 mm slice thickness were compared to reference read by two other pediatric radiologists. Sensitivity (Sn) and positive predictive value (PPV) were analyzed. RESULTS The radiologists identified 109 nodules. At 1 mm, CAD detected 70 nodules; 43 true positive (Sn = 39 %), 26 false positive (PPV = 62 %), and 1 nodule which had not been identified by radiologists. At 3 mm, CAD detected 60 nodules; 28 true positive (Sn = 26 %), 30 false positive (PPV = 48 %) and 2 nodules which had not been identified by radiologists. There were 103 solid nodules (47 measuring < 3 mm) and 6 subsolid nodules (5 measuring < 5 mm). When excluding 52 nodules (solid < 3 mm and subsolid < 5 mm) based on algorithm conditions, the Sn increased to 68 % at 1 mm and 49 % at 3 mm but there was no significant change in the PPV measuring 60 % at 1 mm and 48 % at 3 mm. CONCLUSION The adult Lung CAD showed low sensitivity in pediatric patients, but better performance at thinner slice thickness and when smaller nodules were excluded.
Collapse
Affiliation(s)
- Rida Salman
- Edward B. Singleton Department of Radiology, Division of Body Imaging, Texas Children's Hospital and Baylor College of Medicine, Houston, TX, USA
| | - HaiThuy N Nguyen
- Department of Radiology, Children's Hospital Los Angeles and Keck School of Medicine, University of Southern California, Los Angeles, CA, USA
| | - Andrew C Sher
- Edward B. Singleton Department of Radiology, Division of Body Imaging, Texas Children's Hospital and Baylor College of Medicine, Houston, TX, USA
| | | | - Victor J Seghers
- Edward B. Singleton Department of Radiology, Division of Body Imaging, Texas Children's Hospital and Baylor College of Medicine, Houston, TX, USA
| | - Marla B K Sammer
- Edward B. Singleton Department of Radiology, Division of Body Imaging, Texas Children's Hospital and Baylor College of Medicine, Houston, TX, USA.
| |
Collapse
|
2
|
Soliman M, Petrella T, Tyrrell P, Wright F, Look Hong NJ, Lu H, Zezos P, Jimenez-Juan L, Oikonomou A. The clinical significance of indeterminate pulmonary nodules in melanoma patients at baseline and during follow-up chest CT. Eur J Radiol Open 2019; 6:85-90. [PMID: 30805420 PMCID: PMC6374500 DOI: 10.1016/j.ejro.2019.02.003] [Citation(s) in RCA: 6] [Impact Index Per Article: 1.2] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 01/09/2019] [Revised: 02/04/2019] [Accepted: 02/05/2019] [Indexed: 02/07/2023] Open
Abstract
Objective This study aims to determine an appropriate timeline to monitor indeterminate pulmonary nodules (IPN) in melanoma patients to confirm metastatic origin. Materials and Methods 588 clinically non-metastatic melanoma patients underwent curative intent surgery during 3 years. Patients with baseline chest CT and at least one follow-up (FU) CT were retrospectively analyzed to assess for IPN. Patients with definitely benign nodules, metastases and non-melanoma malignancies were excluded. Change in volume from first to FU CT, initial diameter (D1) and volume (V1), distance from pleura, peripheral and perifissural location, density and clinical stage were evaluated. Nodules were volumetrically measured on CTs and were considered metastases if they increased in size between two CTs or if increase was accompanied by multiple new nodules or extrapulmonary metastases. Results 148 patients were included. Two out of 243 baseline IPN detected in 70 patients, increased significantly in volume in 3 and 5 months and were proven metastases. During FU, 86% of 40 interval IPN detected in 28 patients, were proven metastases. Interval nodule (p < 0.0001, HR:243,CI:[57.32,1033.74]), 3-month volume change (OR:1.023,CI:[1.014,1.033]), V1 (OR:1.006,CI:[1.003,1.009]), D1 (OR:1.424,CI:[1.23,1.648]), distance from pleura (OR:1.03,CI:[1.003,1.059]), and combined stage IIC + III (OR:11.29,CI:[1.514,84.174]), were associated with increased risk for metastasis. 43%, 72% and 94% of patients with IPN were confirmed with metastases in the first FU CT at 3, 6 and 12 months respectively. Conclusion Baseline IPN are most likely benign, while interval IPN are high risk for metastasis. Absence of volume increase of IPN within 6 months excluded metastasis in most patients.
Collapse
Affiliation(s)
- Magdy Soliman
- Department of Medical Imaging, Sunnybrook Health Sciences Centre, University of Toronto, M4N 3M5, Toronto, ON, Canada
| | - Teresa Petrella
- Department of Medicine, Sunnybrook Health Sciences Centre, University of Toronto, M4N 3M5, Toronto, ON, Canada
| | - Pascal Tyrrell
- Department of Medical Imaging, University of Toronto, M5T 1W7, Toronto, ON, Canada
| | - Frances Wright
- Department of Surgery, Sunnybrook Health Sciences Centre, University of Toronto, M4N 3M5, Toronto, ON, Canada
| | - Nicole J Look Hong
- Department of Surgery, Sunnybrook Health Sciences Centre, University of Toronto, M4N 3M5, Toronto, ON, Canada
| | - Hua Lu
- Department of Medical Imaging, University of Toronto, M5T 1W7, Toronto, ON, Canada
| | - Petros Zezos
- Department of Medicine, North Ontario School of Medicine, ON P7B 5E1, Canada
| | - Laura Jimenez-Juan
- Department of Medical Imaging, Sunnybrook Health Sciences Centre, University of Toronto, M4N 3M5, Toronto, ON, Canada
| | - Anastasia Oikonomou
- Department of Medical Imaging, Sunnybrook Health Sciences Centre, University of Toronto, M4N 3M5, Toronto, ON, Canada
| |
Collapse
|