Pancreatic Cancer Detection on CT Scans with Deep Learning: A Nationwide Population-based Study

被引:72
作者
Chen, Po -Ting [1 ]
Wu, Tinghui [3 ]
Wang, Pochuan [4 ]
Chang, Dawei [3 ]
Liu, Kao-Lang [1 ,6 ]
Wu, Ming-Shiang [2 ,5 ]
Roth, Holger R. [7 ]
Lee, Po-Chang [8 ]
Liao, Wei-Chih [2 ,5 ]
Wang, Weichung [3 ]
机构
[1] Natl Taiwan Univ, Natl Taiwan Univ Hosp, Coll Med, Dept Med Imaging, Taipei, Taiwan
[2] Natl Taiwan Univ, Natl Taiwan Univ Hosp, Div Gastroenterol & Hepatol, Coll Med,Dept Internal Med, Taipei, Taiwan
[3] Natl Taiwan Univ, Inst Appl Math Sci, 1 Sect 4, Roosevelt Rd, Taipei 10617, Taiwan
[4] Natl Taiwan Univ, Dept Comp Sci & Informat Engn, 1 Sect 4, Roosevelt Rd, Taipei 10617, Taiwan
[5] Natl Taiwan Univ, Dept Internal Med, Coll Med, 1 Sect 4, Roosevelt Rd, Taipei 10617, Taiwan
[6] Natl Taiwan Univ, Canc Ctr, Dept Med Imaging, Taipei, Taiwan
[7] NVIDIA, Bethesda, MD USA
[8] Minist Hlth & Welf, Natl Hlth Insurance Adm, Taipei, Taiwan
关键词
D O I
10.1148/radiol.220152
中图分类号
R8 [特种医学]; R445 [影像诊断学];
学科分类号
1002 ; 100207 ; 1009 ;
摘要
Background: Approximately 40% of pancreatic tumors smaller than 2 cm are missed at abdominal CT. Purpose: To develop and to validate a deep learning (DL)-based tool able to detect pancreatic cancer at CT. Materials and Methods: Retrospectively collected contrast-enhanced CT studies in patients diagnosed with pancreatic cancer between January 2006 and July 2018 were compared with CT studies of individuals with a normal pancreas (control group) obtained between January 2004 and December 2019. An end-to-end tool comprising a segmentation convolutional neural network (CNN) and a classifier ensembling five CNNs was developed and validated in the internal test set and a nationwide real-world validation set. The sensitivities of the computer-aided detection (CAD) tool and radiologist interpretation were compared using the McNemar test. Results: A total of 546 patients with pancreatic cancer (mean age, 65 years +/- 12 [SD], 297 men) and 733 control subjects were ran-domly divided into training, validation, and test sets. In the internal test set, the DL tool achieved 89.9% (98 of 109; 95% CI: 82.7, 94.9) sensitivity and 95.9% (141 of 147; 95% CI: 91.3, 98.5) specificity (area under the receiver operating characteristic curve [AUC], 0.96; 95% CI: 0.94, 0.99), without a significant difference (P = .11) in sensitivity compared with the original radiologist report (96.1% [98 of 102]; 95% CI: 90.3, 98.9). In a test set of 1473 real-world CT studies (669 malignant, 804 control) from institutions throughout Taiwan, the DL tool distinguished between CT malignant and control studies with 89.7% (600 of 669; 95% CI: 87.1, 91.9) sensitivity and 92.8% specificity (746 of 804; 95% CI: 90.8, 94.5) (AUC, 0.95; 95% CI: 0.94, 0.96), with 74.7% (68 of 91; 95% CI: 64.5, 83.3) sensitivity for malignancies smaller than 2 cm. Conclusion: The deep learning-based tool enabled accurate detection of pancreatic cancer on CT scans, with reasonable sensitivity for tumors smaller than 2 cm.
引用
收藏
页码:172 / 182
页数:11
相关论文
共 30 条
[1]  
Antonelli M, 2021, Arxiv, DOI arXiv:2106.05735
[2]   Automated cardiovascular magnetic resonance image analysis with fully convolutional networks [J].
Bai, Wenjia ;
Sinclair, Matthew ;
Tarroni, Giacomo ;
Oktay, Ozan ;
Rajchl, Martin ;
Vaillant, Ghislain ;
Lee, Aaron M. ;
Aung, Nay ;
Lukaschuk, Elena ;
Sanghvi, Mihir M. ;
Zemrak, Filip ;
Fung, Kenneth ;
Paiva, Jose Miguel ;
Carapella, Valentina ;
Kim, Young Jin ;
Suzuki, Hideaki ;
Kainz, Bernhard ;
Matthews, Paul M. ;
Petersen, Steffen E. ;
Piechnik, Stefan K. ;
Neubauer, Stefan ;
Glocker, Ben ;
Rueckert, Daniel .
JOURNAL OF CARDIOVASCULAR MAGNETIC RESONANCE, 2018, 20
[3]   External validation demonstrates limited clinical utility of the interpretable mortality prediction model for patients with COVID-19 [J].
Barish, Matthew ;
Bolourani, Siavash ;
Lau, Lawrence F. ;
Shah, Sareen ;
Zanos, Theodoros P. .
NATURE MACHINE INTELLIGENCE, 2020, 3 (1) :25-27
[4]   A Human-Centered Evaluation of a Deep Learning System Deployed in Clinics for the Detection of Diabetic Retinopathy [J].
Beede, Emma ;
Baylor, Elizabeth ;
Hersch, Fred ;
Iurchenko, Anna ;
Wilcox, Lauren ;
Ruamviboonsuk, Paisan ;
Vardoulakis, Laura M. .
PROCEEDINGS OF THE 2020 CHI CONFERENCE ON HUMAN FACTORS IN COMPUTING SYSTEMS (CHI'20), 2020,
[5]   Detecting early pancreatic cancer: Problems and prospects [J].
Chari, Suresh T. .
SEMINARS IN ONCOLOGY, 2007, 34 (04) :284-294
[6]   Applications of artificial intelligence in pancreatic and biliary diseases [J].
Chen, Po-Ting ;
Chang, Dawei ;
Wu, Tinghui ;
Wu, Ming-Shiang ;
Wang, Weichung ;
Liao, Wei-Chih .
JOURNAL OF GASTROENTEROLOGY AND HEPATOLOGY, 2021, 36 (02) :286-294
[7]   Statistics notes - Diagnostic tests 4: likelihood ratios [J].
Deeks, JJ ;
Altman, DG .
BRITISH MEDICAL JOURNAL, 2004, 329 (7458) :168-169
[8]   Comparison of endoscopic ultrasound and computed tomography for the preoperative evaluation of pancreatic cancer: A systematic review [J].
Dewitt, John ;
Devereaux, Benedict M. ;
Lehman, Glen A. ;
Sherivian, Stuart ;
Imperiale, Thomas F. .
CLINICAL GASTROENTEROLOGY AND HEPATOLOGY, 2006, 4 (06) :717-725
[9]   A guide to deep learning in healthcare [J].
Esteva, Andre ;
Robicquet, Alexandre ;
Ramsundar, Bharath ;
Kuleshov, Volodymyr ;
DePristo, Mark ;
Chou, Katherine ;
Cui, Claire ;
Corrado, Greg ;
Thrun, Sebastian ;
Dean, Jeff .
NATURE MEDICINE, 2019, 25 (01) :24-29
[10]  
Gibson E., 2018, Multi-organ abdominal CT reference standard segmentations, DOI 10.5281