PMID- 32055949 OWN - NLM STAT- MEDLINE DCOM- 20201123 LR - 20211206 IS - 1432-1084 (Electronic) IS - 0938-7994 (Linking) VI - 30 IP - 6 DP - 2020 Jun TI - CT-based deep learning model to differentiate invasive pulmonary adenocarcinomas appearing as subsolid nodules among surgical candidates: comparison of the diagnostic performance with a size-based logistic model and radiologists. PG - 3295-3305 LID - 10.1007/s00330-019-06628-4 [doi] AB - OBJECTIVES: To evaluate the deep learning models for differentiating invasive pulmonary adenocarcinomas (IACs) among subsolid nodules (SSNs) considered for resection in a retrospective diagnostic cohort in comparison with a size-based logistic model and expert radiologists. METHODS: This study included 525 patients (309 women; median, 62 years) to develop models, and an independent cohort of 101 patients (57 women; median, 66 years) was used for validation. A size-based logistic model and deep learning models using 2.5-dimension (2.5D) and three-dimension (3D) CT images were developed to discriminate IAC from less invasive pathologies. Overall performance, discrimination, and calibration were assessed. Diagnostic performances of the three thoracic radiologists were compared with those of the deep learning model. RESULTS: The overall performances of the deep learning models (Brier score, 0.122 for the 2.5D DenseNet and 0.121 for the 3D DenseNet) were superior to those of the size-based logistic model (Brier score, 0.198). The area under the receiver operating characteristic curve (AUC) of the 2.5D DenseNet (0.921) was significantly higher than that of the 3D DenseNet (0.835; p = 0.037) and the size-based logistic model (0.836; p = 0.009). At equally high sensitivities of 90%, the 2.5D DenseNet showed significantly higher specificity (88.2%; all p < 0.05) and positive predictive value (97.4%; all p < 0.05) than other models. Model calibration was poor for all models (all p < 0.05). The 2.5D DenseNet had a comparable performance with the radiologists (AUC, 0.848-0.910). CONCLUSION: The 2.5D DenseNet model could be used as a highly sensitive and specific diagnostic tool to differentiate IACs among SSNs for surgical candidates. KEY POINTS: * The deep learning model developed using 2.5D DenseNet showed higher overall performance and discrimination than the size-based logistic model for the differentiation of invasive adenocarcinomas among subsolid nodules for surgical candidates. * The 2.5D DenseNet demonstrated a thoracic radiologist-level diagnostic performance and had higher specificity (88.2%) at equal sensitivities (90%) than the size-based logistic model (specificity, 52.9%). * The 2.5D DenseNet could be used to reduce potential overtreatment for the indolent subsolid nodules or to select candidates for sublobar resection instead of the standard lobectomy. FAU - Kim, Hyungjin AU - Kim H AD - Department of Radiology, Seoul National University Hospital, Seoul National University College of Medicine, 101, Daehak-ro, Jongno-gu, Seoul, 03080, South Korea. AD - Institute of Radiation Medicine, Seoul National University Medical Research Center, 101, Daehak-ro, Jongno-gu, Seoul, 03080, South Korea. FAU - Lee, Dongheon AU - Lee D AD - Interdisciplinary Program in Bioengineering, Graduate School, Seoul National University, 101, Daehak-ro, Jongno-gu, Seoul, 03080, South Korea. FAU - Cho, Woo Sang AU - Cho WS AD - Interdisciplinary Program in Bioengineering, Graduate School, Seoul National University, 101, Daehak-ro, Jongno-gu, Seoul, 03080, South Korea. FAU - Lee, Jung Chan AU - Lee JC AD - Department of Biomedical Engineering, Seoul National University College of Medicine, 101, Daehak-ro, Jongno-gu, Seoul, 03080, South Korea. AD - Institute of Medical & Biological Engineering, Medical Research Center, Seoul National University, 101, Daehak-ro, Jongno-gu, Seoul, 03080, South Korea. FAU - Goo, Jin Mo AU - Goo JM AD - Department of Radiology, Seoul National University Hospital, Seoul National University College of Medicine, 101, Daehak-ro, Jongno-gu, Seoul, 03080, South Korea. AD - Institute of Radiation Medicine, Seoul National University Medical Research Center, 101, Daehak-ro, Jongno-gu, Seoul, 03080, South Korea. AD - Cancer Research Institute, Seoul National University, 101, Daehak-ro, Jongno-gu, Seoul, 03080, South Korea. FAU - Kim, Hee Chan AU - Kim HC AD - Interdisciplinary Program in Bioengineering, Graduate School, Seoul National University, 101, Daehak-ro, Jongno-gu, Seoul, 03080, South Korea. hckim@snu.ac.kr. AD - Department of Biomedical Engineering, Seoul National University College of Medicine, 101, Daehak-ro, Jongno-gu, Seoul, 03080, South Korea. hckim@snu.ac.kr. AD - Institute of Medical & Biological Engineering, Medical Research Center, Seoul National University, 101, Daehak-ro, Jongno-gu, Seoul, 03080, South Korea. hckim@snu.ac.kr. FAU - Park, Chang Min AU - Park CM AUID- ORCID: 0000-0003-1884-3738 AD - Department of Radiology, Seoul National University Hospital, Seoul National University College of Medicine, 101, Daehak-ro, Jongno-gu, Seoul, 03080, South Korea. cmpark.morphius@gmail.com. AD - Institute of Radiation Medicine, Seoul National University Medical Research Center, 101, Daehak-ro, Jongno-gu, Seoul, 03080, South Korea. cmpark.morphius@gmail.com. AD - Cancer Research Institute, Seoul National University, 101, Daehak-ro, Jongno-gu, Seoul, 03080, South Korea. cmpark.morphius@gmail.com. LA - eng GR - 2017R1A2B4008517/Ministry of Science, ICT and Future Planning/ PT - Comparative Study PT - Journal Article DEP - 20200213 PL - Germany TA - Eur Radiol JT - European radiology JID - 9114774 SB - IM MH - Adenocarcinoma of Lung/*diagnosis MH - Aged MH - *Deep Learning MH - Female MH - Humans MH - Logistic Models MH - Lung Neoplasms/*diagnosis MH - Male MH - Middle Aged MH - ROC Curve MH - Radiography, Thoracic/*methods MH - *Radiologists MH - Retrospective Studies MH - Tomography, X-Ray Computed/*methods OTO - NOTNLM OT - Adenocarcinoma OT - Artificial intelligence OT - Computer-assisted radiographic image interpretation OT - Logistic model OT - Multidetector computed tomography EDAT- 2020/02/15 06:00 MHDA- 2020/11/24 06:00 CRDT- 2020/02/15 06:00 PHST- 2019/10/08 00:00 [received] PHST- 2019/12/13 00:00 [accepted] PHST- 2019/11/09 00:00 [revised] PHST- 2020/02/15 06:00 [pubmed] PHST- 2020/11/24 06:00 [medline] PHST- 2020/02/15 06:00 [entrez] AID - 10.1007/s00330-019-06628-4 [pii] AID - 10.1007/s00330-019-06628-4 [doi] PST - ppublish SO - Eur Radiol. 2020 Jun;30(6):3295-3305. doi: 10.1007/s00330-019-06628-4. Epub 2020 Feb 13.