Deep learning in CT colonography: differentiating premalignant from benign colorectal polyps

Wesp, Philipp ; Grosu, Sergio; Graser, Anno; Maurus, Stefan; Schulz, Christian; Knosel, Thomas; Fabritius, Matthias P.; Schachtner, Balthasar; Yeh, Benjamin M.; Cyran, Clemens C.; Ricke, Jens; Kazmierczak, Philipp M. und Ingrisch, Michael (2022): Deep learning in CT colonography: differentiating premalignant from benign colorectal polyps. In: European Radiology, Bd. 32, Nr. 7: S. 4749-4759

Abstract

Objectives To investigate the differentiation of premalignant from benign colorectal polyps detected by CT colonography using deep learning. Methods In this retrospective analysis of an average risk colorectal cancer screening sample, polyps of all size categories and morphologies were manually segmented on supine and prone CT colonography images and classified as premalignant (adenoma) or benign (hyperplastic polyp or regular mucosa) according to histopathology. Two deep learning models SEG and noSEG were trained on 3D CT colonography image subvolumes to predict polyp class, and model SEG was additionally trained with polyp segmentation masks. Diagnostic performance was validated in an independent external multicentre test sample. Predictions were analysed with the visualisation technique Grad-CAM++. Results The training set consisted of 107 colorectal polyps in 63 patients (mean age: 63 +/- 8 years, 40 men) comprising 169 polyp segmentations. The external test set included 77 polyps in 59 patients comprising 118 polyp segmentations. Model SEG achieved a ROC-AUC of 0.83 and 80% sensitivity at 69% specificity for differentiating premalignant from benign polyps. Model noSEG yielded a ROC-AUC of 0.75, 80% sensitivity at 44% specificity, and an average Grad-CAM++ heatmap score of >= 0.25 in 90% of polyp tissue. Conclusions In this proof-of-concept study, deep learning enabled the differentiation of premalignant from benign colorectal polyps detected with CT colonography and the visualisation of image regions important for predictions. The approach did not require polyp segmentation and thus has the potential to facilitate the identification of high-risk polyps as an automated second reader.

• Dokumententyp: Zeitschriftenartikel
• Fakultät: Medizin
• Themengebiete: 600 Technik, Medizin, angewandte Wissenschaften > 610 Medizin und Gesundheit
• ISSN: 0938-7994
• Sprache: Englisch
• DFG: Gefördert durch die Deutsche Forschungsgemeinschaft (DFG) - 491502892