COVLIAS 2.0-cXAI: Cloud-Based Explainable Deep Learning System for COVID-19 Lesion Localization in Computed Tomography Scans
Background: The previous COVID-19 lung diagnosis system lacks both scientific validation and the role of explainable artificial intelligence (AI) for understanding lesion localization. This study presents a cloud-based explainable AI, the “COVLIAS 2.0-cXAI” system using four kinds of class activatio...
| Main Authors: | , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-06-01
|
| Series: | Diagnostics |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2075-4418/12/6/1482 |
| _version_ | 1827660929107492864 |
|---|---|
| author | Jasjit S. Suri Sushant Agarwal Gian Luca Chabert Alessandro Carriero Alessio Paschè Pietro S. C. Danna Luca Saba Armin Mehmedović Gavino Faa Inder M. Singh Monika Turk Paramjit S. Chadha Amer M. Johri Narendra N. Khanna Sophie Mavrogeni John R. Laird Gyan Pareek Martin Miner David W. Sobel Antonella Balestrieri Petros P. Sfikakis George Tsoulfas Athanasios D. Protogerou Durga Prasanna Misra Vikas Agarwal George D. Kitas Jagjit S. Teji Mustafa Al-Maini Surinder K. Dhanjil Andrew Nicolaides Aditya Sharma Vijay Rathore Mostafa Fatemi Azra Alizad Pudukode R. Krishnan Ferenc Nagy Zoltan Ruzsa Mostafa M. Fouda Subbaram Naidu Klaudija Viskovic Mannudeep K. Kalra |
| author_facet | Jasjit S. Suri Sushant Agarwal Gian Luca Chabert Alessandro Carriero Alessio Paschè Pietro S. C. Danna Luca Saba Armin Mehmedović Gavino Faa Inder M. Singh Monika Turk Paramjit S. Chadha Amer M. Johri Narendra N. Khanna Sophie Mavrogeni John R. Laird Gyan Pareek Martin Miner David W. Sobel Antonella Balestrieri Petros P. Sfikakis George Tsoulfas Athanasios D. Protogerou Durga Prasanna Misra Vikas Agarwal George D. Kitas Jagjit S. Teji Mustafa Al-Maini Surinder K. Dhanjil Andrew Nicolaides Aditya Sharma Vijay Rathore Mostafa Fatemi Azra Alizad Pudukode R. Krishnan Ferenc Nagy Zoltan Ruzsa Mostafa M. Fouda Subbaram Naidu Klaudija Viskovic Mannudeep K. Kalra |
| author_sort | Jasjit S. Suri |
| collection | DOAJ |
| description | Background: The previous COVID-19 lung diagnosis system lacks both scientific validation and the role of explainable artificial intelligence (AI) for understanding lesion localization. This study presents a cloud-based explainable AI, the “COVLIAS 2.0-cXAI” system using four kinds of class activation maps (CAM) models. Methodology: Our cohort consisted of ~6000 CT slices from two sources (Croatia, 80 COVID-19 patients and Italy, 15 control patients). COVLIAS 2.0-cXAI design consisted of three stages: (i) automated lung segmentation using hybrid deep learning ResNet-UNet model by automatic adjustment of Hounsfield units, hyperparameter optimization, and parallel and distributed training, (ii) classification using three kinds of DenseNet (DN) models (DN-121, DN-169, DN-201), and (iii) validation using four kinds of CAM visualization techniques: gradient-weighted class activation mapping (Grad-CAM), Grad-CAM++, score-weighted CAM (Score-CAM), and FasterScore-CAM. The COVLIAS 2.0-cXAI was validated by three trained senior radiologists for its stability and reliability. The Friedman test was also performed on the scores of the three radiologists. Results: The ResNet-UNet segmentation model resulted in dice similarity of 0.96, Jaccard index of 0.93, a correlation coefficient of 0.99, with a figure-of-merit of 95.99%, while the classifier accuracies for the three DN nets (DN-121, DN-169, and DN-201) were 98%, 98%, and 99% with a loss of ~0.003, ~0.0025, and ~0.002 using 50 epochs, respectively. The mean AUC for all three DN models was 0.99 (p < 0.0001). The COVLIAS 2.0-cXAI showed 80% scans for mean alignment index (MAI) between heatmaps and gold standard, a score of four out of five, establishing the system for clinical settings. Conclusions: The COVLIAS 2.0-cXAI successfully showed a cloud-based explainable AI system for lesion localization in lung CT scans. |
| first_indexed | 2024-03-09T23:59:19Z |
| format | Article |
| id | doaj.art-724250641a304ca5bdab585ac56026fe |
| institution | Directory Open Access Journal |
| issn | 2075-4418 |
| language | English |
| last_indexed | 2024-03-09T23:59:19Z |
| publishDate | 2022-06-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Diagnostics |
| spelling | doaj.art-724250641a304ca5bdab585ac56026fe2023-11-23T16:18:54ZengMDPI AGDiagnostics2075-44182022-06-01126148210.3390/diagnostics12061482COVLIAS 2.0-cXAI: Cloud-Based Explainable Deep Learning System for COVID-19 Lesion Localization in Computed Tomography ScansJasjit S. Suri0Sushant Agarwal1Gian Luca Chabert2Alessandro Carriero3Alessio Paschè4Pietro S. C. Danna5Luca Saba6Armin Mehmedović7Gavino Faa8Inder M. Singh9Monika Turk10Paramjit S. Chadha11Amer M. Johri12Narendra N. Khanna13Sophie Mavrogeni14John R. Laird15Gyan Pareek16Martin Miner17David W. Sobel18Antonella Balestrieri19Petros P. Sfikakis20George Tsoulfas21Athanasios D. Protogerou22Durga Prasanna Misra23Vikas Agarwal24George D. Kitas25Jagjit S. Teji26Mustafa Al-Maini27Surinder K. Dhanjil28Andrew Nicolaides29Aditya Sharma30Vijay Rathore31Mostafa Fatemi32Azra Alizad33Pudukode R. Krishnan34Ferenc Nagy35Zoltan Ruzsa36Mostafa M. Fouda37Subbaram Naidu38Klaudija Viskovic39Mannudeep K. Kalra40Stroke Diagnostic and Monitoring Division, AtheroPoint™, Roseville, CA 95661, USAAdvanced Knowledge Engineering Centre, GBTI, Roseville, CA 95661, USADepartment of Radiology, Azienda Ospedaliero Universitaria (A.O.U.), 09123 Cagliari, ItalyDepartment of Radiology, “Maggiore della Carità” Hospital, University of Piemonte Orientale (UPO), Via Solaroli 17, 28100 Novara, ItalyDepartment of Radiology, Azienda Ospedaliero Universitaria (A.O.U.), 09123 Cagliari, ItalyDepartment of Radiology, Azienda Ospedaliero Universitaria (A.O.U.), 09123 Cagliari, ItalyDepartment of Radiology, Azienda Ospedaliero Universitaria (A.O.U.), 09123 Cagliari, ItalyDepartment of Radiology, University Hospital for Infectious Diseases, 10000 Zagreb, CroatiaDepartment of Pathology, Azienda Ospedaliero Universitaria (A.O.U.), 09124 Cagliari, ItalyStroke Diagnostic and Monitoring Division, AtheroPoint™, Roseville, CA 95661, USAThe Hanse-Wissenschaftskolleg Institute for Advanced Study, 27753 Delmenhorst, GermanyStroke Diagnostic and Monitoring Division, AtheroPoint™, Roseville, CA 95661, USADepartment of Medicine, Division of Cardiology, Queen’s University, Kingston, ON K7L 3N6, CanadaDepartment of Cardiology, Indraprastha APOLLO Hospitals, New Delhi 110076, IndiaCardiology Clinic, Onassis Cardiac Surgery Center, 17674 Athens, GreeceHeart and Vascular Institute, Adventist Health St. Helena, St. Helena, CA 94574, USAMinimally Invasive Urology Institute, Brown University, Providence, RI 02912, USAMen’s Health Center, Miriam Hospital, Providence, RI 02912, USAMinimally Invasive Urology Institute, Brown University, Providence, RI 02912, USADepartment of Radiology, Azienda Ospedaliero Universitaria (A.O.U.), 09123 Cagliari, ItalyRheumatology Unit, National Kapodistrian University of Athens, 17674 Athens, GreeceDepartment of Surgery, Aristoteleion University of Thessaloniki, 54124 Thessaloniki, GreeceCardiovascular Prevention and Research Unit, Department of Pathophysiology, National & Kapodistrian University of Athens, 15772 Athens, GreeceDepartment of Immunology, SGPIMS, Lucknow 226014, IndiaDepartment of Immunology, SGPIMS, Lucknow 226014, IndiaAcademic Affairs, Dudley Group NHS Foundation Trust, Dudley DY1 2HQ, UKAnn and Robert H. Lurie Children’s Hospital of Chicago, Chicago, IL 60611, USAAllergy, Clinical Immunology and Rheumatology Institute, Toronto, ON M5G 1N8, CanadaAtheroPoint LLC., Roseville, CA 95661, USAVascular Screening and Diagnostic Centre, University of Nicosia Medical School, Engomi 2408, CyprusDivision of Cardiovascular Medicine, University of Virginia, Charlottesville, VA 22902, USAAtheroPoint LLC., Roseville, CA 95661, USADepartment of Physiology & Biomedical Engineering, Mayo Clinic College of Medicine and Science, Rochester, MN 55905, USADepartment of Radiology, Mayo Clinic College of Medicine and Science, Rochester, MN 55905, USANeurology Department, Fortis Hospital, Bengaluru 560076, IndiaInternal Medicine Department, University of Szeged, 6725 Szeged, HungaryInvasive Cardiology Division, University of Szeged, 1122 Budapest, HungaryDepartment of ECE, Idaho State University, Pocatello, ID 83209, USAElectrical Engineering Department, University of Minnesota, Duluth, MN 55812, USADepartment of Radiology, University Hospital for Infectious Diseases, 10000 Zagreb, CroatiaDepartment of Radiology, Massachusetts General Hospital, 55 Fruit Street, Boston, MA 02114, USABackground: The previous COVID-19 lung diagnosis system lacks both scientific validation and the role of explainable artificial intelligence (AI) for understanding lesion localization. This study presents a cloud-based explainable AI, the “COVLIAS 2.0-cXAI” system using four kinds of class activation maps (CAM) models. Methodology: Our cohort consisted of ~6000 CT slices from two sources (Croatia, 80 COVID-19 patients and Italy, 15 control patients). COVLIAS 2.0-cXAI design consisted of three stages: (i) automated lung segmentation using hybrid deep learning ResNet-UNet model by automatic adjustment of Hounsfield units, hyperparameter optimization, and parallel and distributed training, (ii) classification using three kinds of DenseNet (DN) models (DN-121, DN-169, DN-201), and (iii) validation using four kinds of CAM visualization techniques: gradient-weighted class activation mapping (Grad-CAM), Grad-CAM++, score-weighted CAM (Score-CAM), and FasterScore-CAM. The COVLIAS 2.0-cXAI was validated by three trained senior radiologists for its stability and reliability. The Friedman test was also performed on the scores of the three radiologists. Results: The ResNet-UNet segmentation model resulted in dice similarity of 0.96, Jaccard index of 0.93, a correlation coefficient of 0.99, with a figure-of-merit of 95.99%, while the classifier accuracies for the three DN nets (DN-121, DN-169, and DN-201) were 98%, 98%, and 99% with a loss of ~0.003, ~0.0025, and ~0.002 using 50 epochs, respectively. The mean AUC for all three DN models was 0.99 (p < 0.0001). The COVLIAS 2.0-cXAI showed 80% scans for mean alignment index (MAI) between heatmaps and gold standard, a score of four out of five, establishing the system for clinical settings. Conclusions: The COVLIAS 2.0-cXAI successfully showed a cloud-based explainable AI system for lesion localization in lung CT scans.https://www.mdpi.com/2075-4418/12/6/1482COVID-19 lesionlung CTHounsfield unitsglass ground opacitieshybrid deep learningexplainable AI |
| spellingShingle | Jasjit S. Suri Sushant Agarwal Gian Luca Chabert Alessandro Carriero Alessio Paschè Pietro S. C. Danna Luca Saba Armin Mehmedović Gavino Faa Inder M. Singh Monika Turk Paramjit S. Chadha Amer M. Johri Narendra N. Khanna Sophie Mavrogeni John R. Laird Gyan Pareek Martin Miner David W. Sobel Antonella Balestrieri Petros P. Sfikakis George Tsoulfas Athanasios D. Protogerou Durga Prasanna Misra Vikas Agarwal George D. Kitas Jagjit S. Teji Mustafa Al-Maini Surinder K. Dhanjil Andrew Nicolaides Aditya Sharma Vijay Rathore Mostafa Fatemi Azra Alizad Pudukode R. Krishnan Ferenc Nagy Zoltan Ruzsa Mostafa M. Fouda Subbaram Naidu Klaudija Viskovic Mannudeep K. Kalra COVLIAS 2.0-cXAI: Cloud-Based Explainable Deep Learning System for COVID-19 Lesion Localization in Computed Tomography Scans Diagnostics COVID-19 lesion lung CT Hounsfield units glass ground opacities hybrid deep learning explainable AI |
| title | COVLIAS 2.0-cXAI: Cloud-Based Explainable Deep Learning System for COVID-19 Lesion Localization in Computed Tomography Scans |
| title_full | COVLIAS 2.0-cXAI: Cloud-Based Explainable Deep Learning System for COVID-19 Lesion Localization in Computed Tomography Scans |
| title_fullStr | COVLIAS 2.0-cXAI: Cloud-Based Explainable Deep Learning System for COVID-19 Lesion Localization in Computed Tomography Scans |
| title_full_unstemmed | COVLIAS 2.0-cXAI: Cloud-Based Explainable Deep Learning System for COVID-19 Lesion Localization in Computed Tomography Scans |
| title_short | COVLIAS 2.0-cXAI: Cloud-Based Explainable Deep Learning System for COVID-19 Lesion Localization in Computed Tomography Scans |
| title_sort | covlias 2 0 cxai cloud based explainable deep learning system for covid 19 lesion localization in computed tomography scans |
| topic | COVID-19 lesion lung CT Hounsfield units glass ground opacities hybrid deep learning explainable AI |
| url | https://www.mdpi.com/2075-4418/12/6/1482 |
| work_keys_str_mv | AT jasjitssuri covlias20cxaicloudbasedexplainabledeeplearningsystemforcovid19lesionlocalizationincomputedtomographyscans AT sushantagarwal covlias20cxaicloudbasedexplainabledeeplearningsystemforcovid19lesionlocalizationincomputedtomographyscans AT gianlucachabert covlias20cxaicloudbasedexplainabledeeplearningsystemforcovid19lesionlocalizationincomputedtomographyscans AT alessandrocarriero covlias20cxaicloudbasedexplainabledeeplearningsystemforcovid19lesionlocalizationincomputedtomographyscans AT alessiopasche covlias20cxaicloudbasedexplainabledeeplearningsystemforcovid19lesionlocalizationincomputedtomographyscans AT pietroscdanna covlias20cxaicloudbasedexplainabledeeplearningsystemforcovid19lesionlocalizationincomputedtomographyscans AT lucasaba covlias20cxaicloudbasedexplainabledeeplearningsystemforcovid19lesionlocalizationincomputedtomographyscans AT arminmehmedovic covlias20cxaicloudbasedexplainabledeeplearningsystemforcovid19lesionlocalizationincomputedtomographyscans AT gavinofaa covlias20cxaicloudbasedexplainabledeeplearningsystemforcovid19lesionlocalizationincomputedtomographyscans AT indermsingh covlias20cxaicloudbasedexplainabledeeplearningsystemforcovid19lesionlocalizationincomputedtomographyscans AT monikaturk covlias20cxaicloudbasedexplainabledeeplearningsystemforcovid19lesionlocalizationincomputedtomographyscans AT paramjitschadha covlias20cxaicloudbasedexplainabledeeplearningsystemforcovid19lesionlocalizationincomputedtomographyscans AT amermjohri covlias20cxaicloudbasedexplainabledeeplearningsystemforcovid19lesionlocalizationincomputedtomographyscans AT narendrankhanna covlias20cxaicloudbasedexplainabledeeplearningsystemforcovid19lesionlocalizationincomputedtomographyscans AT sophiemavrogeni covlias20cxaicloudbasedexplainabledeeplearningsystemforcovid19lesionlocalizationincomputedtomographyscans AT johnrlaird covlias20cxaicloudbasedexplainabledeeplearningsystemforcovid19lesionlocalizationincomputedtomographyscans AT gyanpareek covlias20cxaicloudbasedexplainabledeeplearningsystemforcovid19lesionlocalizationincomputedtomographyscans AT martinminer covlias20cxaicloudbasedexplainabledeeplearningsystemforcovid19lesionlocalizationincomputedtomographyscans AT davidwsobel covlias20cxaicloudbasedexplainabledeeplearningsystemforcovid19lesionlocalizationincomputedtomographyscans AT antonellabalestrieri covlias20cxaicloudbasedexplainabledeeplearningsystemforcovid19lesionlocalizationincomputedtomographyscans AT petrospsfikakis covlias20cxaicloudbasedexplainabledeeplearningsystemforcovid19lesionlocalizationincomputedtomographyscans AT georgetsoulfas covlias20cxaicloudbasedexplainabledeeplearningsystemforcovid19lesionlocalizationincomputedtomographyscans AT athanasiosdprotogerou covlias20cxaicloudbasedexplainabledeeplearningsystemforcovid19lesionlocalizationincomputedtomographyscans AT durgaprasannamisra covlias20cxaicloudbasedexplainabledeeplearningsystemforcovid19lesionlocalizationincomputedtomographyscans AT vikasagarwal covlias20cxaicloudbasedexplainabledeeplearningsystemforcovid19lesionlocalizationincomputedtomographyscans AT georgedkitas covlias20cxaicloudbasedexplainabledeeplearningsystemforcovid19lesionlocalizationincomputedtomographyscans AT jagjitsteji covlias20cxaicloudbasedexplainabledeeplearningsystemforcovid19lesionlocalizationincomputedtomographyscans AT mustafaalmaini covlias20cxaicloudbasedexplainabledeeplearningsystemforcovid19lesionlocalizationincomputedtomographyscans AT surinderkdhanjil covlias20cxaicloudbasedexplainabledeeplearningsystemforcovid19lesionlocalizationincomputedtomographyscans AT andrewnicolaides covlias20cxaicloudbasedexplainabledeeplearningsystemforcovid19lesionlocalizationincomputedtomographyscans AT adityasharma covlias20cxaicloudbasedexplainabledeeplearningsystemforcovid19lesionlocalizationincomputedtomographyscans AT vijayrathore covlias20cxaicloudbasedexplainabledeeplearningsystemforcovid19lesionlocalizationincomputedtomographyscans AT mostafafatemi covlias20cxaicloudbasedexplainabledeeplearningsystemforcovid19lesionlocalizationincomputedtomographyscans AT azraalizad covlias20cxaicloudbasedexplainabledeeplearningsystemforcovid19lesionlocalizationincomputedtomographyscans AT pudukoderkrishnan covlias20cxaicloudbasedexplainabledeeplearningsystemforcovid19lesionlocalizationincomputedtomographyscans AT ferencnagy covlias20cxaicloudbasedexplainabledeeplearningsystemforcovid19lesionlocalizationincomputedtomographyscans AT zoltanruzsa covlias20cxaicloudbasedexplainabledeeplearningsystemforcovid19lesionlocalizationincomputedtomographyscans AT mostafamfouda covlias20cxaicloudbasedexplainabledeeplearningsystemforcovid19lesionlocalizationincomputedtomographyscans AT subbaramnaidu covlias20cxaicloudbasedexplainabledeeplearningsystemforcovid19lesionlocalizationincomputedtomographyscans AT klaudijaviskovic covlias20cxaicloudbasedexplainabledeeplearningsystemforcovid19lesionlocalizationincomputedtomographyscans AT mannudeepkkalra covlias20cxaicloudbasedexplainabledeeplearningsystemforcovid19lesionlocalizationincomputedtomographyscans |