Artificial intelligence solution to classify pulmonary nodules on CT - 26/11/20
Bienvenue sur EM-consulte, la référence des professionnels de santé.
Article gratuit.
Connectez-vous pour en bénéficier!
Highlights |
• | An algorithm was created to detect and classify pulmonary nodules in two categories based on their volume greater than 100 mm3 or not, using machine learning and deep learning techniques. |
• | A fully functional pipeline using 3D U-NET, 3D Retina-UNET and classifier stage with a support vector machine algorithm was developed, resulting in high capabilities for pulmonary nodule classification. |
• | The developed pipeline, from a database from different hospitals and with different data acquisition protocols, has very satisfactory performance. |
Abstract |
Purpose |
The purpose of this study was to create an algorithm to detect and classify pulmonary nodules in two categories based on their volume greater than 100 mm3 or not, using machine learning and deep learning techniques.
Materials and method |
The dataset used to train the model was provided by the organization team of the SFR (French Radiological Society) Data Challenge 2019. An asynchronous and parallel 3-stages pipeline was developed to process all the data (a data “pre-processing” stage; a “nodule detection” stage; a “classifier” stage). Lung segmentation was achieved using 3D U-NET algorithm; nodule detection was done using 3D Retina-UNET and classifier stage with a support vector machine algorithm on selected features. Performances were assessed using area under receiver operating characteristics curve (AUROC).
Results |
The pipeline showed good performance for pathological nodule detection and patient diagnosis. With the preparation dataset, an AUROC of 0.9058 (95% confidence interval [CI]: 0.8746–0.9362) was obtained, 87% yielding accuracy (95% CI: 84.83%–91.03%) for the “nodule detection” stage, corresponding to 86% specificity (95% CI: 82%–92%) and 89% sensitivity (95% CI: 84.83%–91.03%).
Conclusion |
A fully functional pipeline using 3D U-NET, 3D Retina-UNET and classifier stage with a support vector machine algorithm was developed, resulting in high capabilities for pulmonary nodule classification.
Le texte complet de cet article est disponible en PDF.Keywords : Lung cancer, Pulmonary nodule, Support vector machine, Deep learning, Machine learning.
Abbreviations : 2D, 3D, AI, AUC, AUROC, CAD, CNN, CPU, CT, FN, FP, GPU, HU, LIDC, R-CNN, RFE, ROC, SVM
Plan
Vol 101 - N° 12
P. 803-810 - décembre 2020 Retour au numéro
Article précédent
- Artificial intelligence to predict clinical disability in patients with multiple sclerosis using FLAIR MRI
- P. Roca, A. Attye, L. Colas, A. Tucholka, P. Rubini, S. Cackowski, J. Ding, J.-F. Budzik, F. Renard, S. Doyle, E.L. Barbier, I. Bousaid, R. Casey, S. Vukusic, N. Lassau, S. Verclytte, F. Cotton, OFSEP Investigators, Steering Committee, Investigators, Imaging group, B. Brochet, R. Casey, F. Cotton, J. De Sèze, P. Douek, F. Guillemin, D. Laplaud, C. Lebrun-Frenay, L. Mansuy, T. Moreau, J. Olaiz, J. Pelletier, C. Rigaud-Bully, B. Stankoff, S. Vukusic, R. Marignier, M. Debouverie, G. Edan, J. Ciron, A. Ruet, N. Collongues, C. Lubetzki, P. Vermersch, P. Labauge, G. Defer, M. Cohen, A. Fromont, S. Wiertlewsky, E. Berger, P. Clavelou, B. Audoin, C. Giannesini, O. Gout, E. Thouvenot, O. Heinzlef, A. Al-Khedr, B. Bourre, O. Casez, P. Cabre, A. Montcuquet, A. Créange, J.-P. Camdessanché, J. Faure, A. Maurousset, I. Patry, K. Hankiewicz, C. Pottier, N. Maubeuge, C. Labeyrie, C. Nifle, R. Ameli, R. Anxionnat, A. Attye, E. Bannier, C. Barillot, D. Ben Salem, M.-P. Boncoeur-Martel, F. Bonneville, C. Boutet, J.-C. Brisset, F. Cervenanski, B. Claise, O. Commowick, J.-M. Constans, P. Dardel, H. Desal, Vincent Dousset, F. Durand-Dubief, J.-C. Ferre, E. Gerardin, T. Glattard, S. Grand, T. Grenier, R. Guillevin, C. Guttmann, A. Krainik, S. Kremer, S. Lion, N. Menjot de Champfleur, L. Mondot, O. Outteryck, N. Pyatigorskaya, J.-P. Pruvo, S. Rabaste, J.-P. Ranjeva, J.-A. Roch, J.C. Sadik, D. Sappey-Marinier, J. Savatovsky, J.-Y. Tanguy, A. Tourbah, T. Tourdias
- Comparison of breast density assessment between human eye and automated software on digital and synthetic mammography: Impact on breast cancer risk
- M. Le Boulc’h, A. Bekhouche, E. Kermarrec, A. Milon, C. Abdel Wahab, S. Zilberman, N. Chabbert-Buffet, I. Thomassin-Naggara
Bienvenue sur EM-consulte, la référence des professionnels de santé.