Independent evaluation of 12 artificial intelligence solutions for the detection of tuberculosis

被引:56
作者
Codlin, Andrew J. [1 ]
Thang Phuoc Dao [2 ]
Luan Nguyen Quang Vo [1 ,2 ]
Forse, Rachel J. [1 ,3 ]
Vinh Van Truong [4 ]
Ha Minh Dang [4 ]
Lan Huu Nguyen [4 ]
Hoa Binh Nguyen [5 ]
Nhung Viet Nguyen [5 ]
Sidney-Annerstedt, Kristi [3 ]
Squire, Bertie [6 ]
Lonnroth, Knut [3 ]
Caws, Maxine [6 ,7 ]
机构
[1] Friends Int TB Relief FIT, Ho Chi Minh City, Vietnam
[2] IRD VN, Ho Chi Minh City, Vietnam
[3] Karolinska Inst, WHO Collaborat Ctr TB & Social Med, Dept Global Publ Hlth, Solna, Sweden
[4] Pham Ngoc Thach Hosp, Ho Chi Minh City, Vietnam
[5] Natl Lung Hosp, Hanoi, Vietnam
[6] Liverpool Sch Trop Med LSTM, Dept Clin Sci, Liverpool, Merseyside, England
[7] Birat Nepal Med Trust, Kathmandu, Nepal
基金
欧盟地平线“2020”;
关键词
CHEST RADIOGRAPHY; PULMONARY TUBERCULOSIS;
D O I
10.1038/s41598-021-03265-0
中图分类号
O [数理科学和化学]; P [天文学、地球科学]; Q [生物科学]; N [自然科学总论];
学科分类号
07 ; 0710 ; 09 ;
摘要
There have been few independent evaluations of computer-aided detection (CAD) software for tuberculosis (TB) screening, despite the rapidly expanding array of available CAD solutions. We developed a test library of chest X-ray (CXR) images which was blindly re-read by two TB clinicians with different levels of experience and then processed by 12 CAD software solutions. Using Xpert MTB/RIF results as the reference standard, we compared the performance characteristics of each CAD software against both an Expert and Intermediate Reader, using cut-off thresholds which were selected to match the sensitivity of each human reader. Six CAD systems performed on par with the Expert Reader (Qure.ai, DeepTek, Delft Imaging, JF Healthcare, OXIPIT, and Lunit) and one additional software (Infervision) performed on par with the Intermediate Reader only. Qure.ai, Delft Imaging and Lunit were the only software to perform significantly better than the Intermediate Reader. The majority of these CAD software showed significantly lower performance among participants with a past history of TB. The radiography equipment used to capture the CXR image was also shown to affect performance for some CAD software. TB program implementers now have a wide selection of quality CAD software solutions to utilize in their CXR screening initiatives.
引用
收藏
页数:11
相关论文
共 49 条
[1]  
Abdat N., 2010, DIGITAL PRODUCT MANA, P154
[2]  
[Anonymous], 2011, Tuberculosis prevalence surveys: a handbook
[3]   Diagnostic Accuracy of Computer-Aided Detection of Pulmonary Tuberculosis in Chest Radiographs: A Validation Study from Sub-Saharan Africa [J].
Breuninger, Marianne ;
van Ginneken, Bram ;
Philipsen, Rick H. H. M. ;
Mhimbira, Francis ;
Hella, Jerry J. ;
Lwilla, Fred ;
van den Hombergh, Jan ;
Ross, Amanda ;
Jugheli, Levan ;
Wagner, Dirk ;
Reither, Klaus .
PLOS ONE, 2014, 9 (09)
[4]   Mobile community-based active case-finding for tuberculosis among older populations in rural Cambodia [J].
Camelique, O. ;
Scholtissen, S. ;
Dousset, J. -P. ;
Bonnet, M. ;
Bastard, M. ;
Hewison, C. .
INTERNATIONAL JOURNAL OF TUBERCULOSIS AND LUNG DISEASE, 2019, 23 (10) :1107-+
[5]   From the help desk: Comparing areas under receiver operating characteristic curves from two or more probit or logit models [J].
Cleves, Mario A. .
STATA JOURNAL, 2002, 2 (03) :301-313
[6]  
Codlin Andrew James, 2018, J Clin Tuberc Other Mycobact Dis, V13, P22, DOI 10.1016/j.jctube.2018.11.001
[7]   When to consult precision-recall curves [J].
Cook, Jonathan ;
Ramadas, Vikram .
STATA JOURNAL, 2020, 20 (01) :131-148
[8]   A Multi-Site Evaluation of Innovative Approaches to Increase Tuberculosis Case Notification: Summary Results [J].
Creswell, Jacob ;
Sahu, Suvanand ;
Blok, Lucie ;
Bakker, Mirjam I. ;
Stevens, Robert ;
Ditiu, Lucica .
PLOS ONE, 2014, 9 (04)
[9]  
Dembrower K, 2020, LANCET DIGIT HEALTH, V2, pE468, DOI 10.1016/S2589-7500(20)30185-0
[10]   World Health Organization cardiovascular disease risk charts: revised models to estimate risk in 21 global regions [J].
Di Angelantonio, Emanuele ;
Kaptoge, Stephen ;
Pennells, Lisa ;
De Bacquer, Dirk ;
Cooney, Marie Therese ;
Kavousi, Maryam ;
Stevens, Gretchen ;
Riley, Leanne ;
Savin, Stefan ;
Altay, Servet ;
Amouyel, Philippe ;
Assmann, Gerd ;
Bell, Steven ;
Ben-Shlomo, Yoav ;
Berkman, Lisa ;
Beulens, Joline W. ;
Bjorkelund, Cecilia ;
Blaha, Michael J. ;
Blazer, Dan G. ;
Bolton, Thomas ;
Bonita, Ruth ;
Brenner, Beaglehole Hermann ;
Brunner, Eric J. ;
Casiglia, Edoardo ;
Chamnan, Parinya ;
Choi, Yeun-Hyang ;
Chowdhury, Rajiv ;
Coady, Sean ;
Crespo, Carlos J. ;
Cushman, Mary ;
Dagenais, Gilles R. ;
D'Agostino, Ralph B. ;
Daimon, Makoto ;
Davidson, Karina W. ;
Engstrom, Gunnar ;
Fang, Xianghua ;
Ford, Ian ;
Gallacher, John ;
Gansevoort, Ron T. ;
Gaziano, Thomas Andrew ;
Giampaoli, Simona ;
Grandits, Greg ;
Grimsgaard, Sameline ;
Grobbee, Diederick E. ;
Gudnason, Vilmundur ;
Guo, Qi ;
Humphries, Steve ;
Iso, Hiroyasu ;
Jukema, J. Wouter ;
Kauhanen, Jussi .
LANCET GLOBAL HEALTH, 2019, 7 (10) :E1332-E1345