Using Multiple Dermoscopic Photographs of One Lesion Improves Melanoma Classification via Deep Learning: A Prognostic Diagnostic Accuracy Study
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2023
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866914271045091328 |
|---|---|
| author | Hekler, Achim Maron, Roman C. Haggenmüller, Sarah Schmitt, Max Wies, Christoph Utikal, Jochen S. Meier, Friedegund Hobelsberger, Sarah Gellrich, Frank F. Sergon, Mildred Hauschild, Axel French, Lars E. Heinzerling, Lucie Schlager, Justin G. Ghoreschi, Kamran Schlaak, Max Hilke, Franz J. Poch, Gabriela Korsing, Sören Berking, Carola Heppt, Markus V. Erdmann, Michael Haferkamp, Sebastian Drexler, Konstantin Schadendorf, Dirk Sondermann, Wiebke Goebeler, Matthias Schilling, Bastian Kather, Jakob N. Krieghoff-Henning, Eva Brinker, Titus J. |
| author_facet | Hekler, Achim Maron, Roman C. Haggenmüller, Sarah Schmitt, Max Wies, Christoph Utikal, Jochen S. Meier, Friedegund Hobelsberger, Sarah Gellrich, Frank F. Sergon, Mildred Hauschild, Axel French, Lars E. Heinzerling, Lucie Schlager, Justin G. Ghoreschi, Kamran Schlaak, Max Hilke, Franz J. Poch, Gabriela Korsing, Sören Berking, Carola Heppt, Markus V. Erdmann, Michael Haferkamp, Sebastian Drexler, Konstantin Schadendorf, Dirk Sondermann, Wiebke Goebeler, Matthias Schilling, Bastian Kather, Jakob N. Krieghoff-Henning, Eva Brinker, Titus J. |
| contents | Background: Convolutional neural network (CNN)-based melanoma classifiers face several challenges that limit their usefulness in clinical practice. Objective: To investigate the impact of multiple real-world dermoscopic views of a single lesion of interest on a CNN-based melanoma classifier.
Methods: This study evaluated 656 suspected melanoma lesions. Classifier performance was measured using area under the receiver operating characteristic curve (AUROC), expected calibration error (ECE) and maximum confidence change (MCC) for (I) a single-view scenario, (II) a multiview scenario using multiple artificially modified images per lesion and (III) a multiview scenario with multiple real-world images per lesion.
Results: The multiview approach with real-world images significantly increased the AUROC from 0.905 (95% CI, 0.879-0.929) in the single-view approach to 0.930 (95% CI, 0.909-0.951). ECE and MCC also improved significantly from 0.131 (95% CI, 0.105-0.159) to 0.072 (95% CI: 0.052-0.093) and from 0.149 (95% CI, 0.125-0.171) to 0.115 (95% CI: 0.099-0.131), respectively. Comparing multiview real-world to artificially modified images showed comparable diagnostic accuracy and uncertainty estimation, but significantly worse robustness for the latter.
Conclusion: Using multiple real-world images is an inexpensive method to positively impact the performance of a CNN-based melanoma classifier. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2306_02800 |
| institution | arXiv |
| publishDate | 2023 |
| record_format | arxiv |
| spellingShingle | Using Multiple Dermoscopic Photographs of One Lesion Improves Melanoma Classification via Deep Learning: A Prognostic Diagnostic Accuracy Study Hekler, Achim Maron, Roman C. Haggenmüller, Sarah Schmitt, Max Wies, Christoph Utikal, Jochen S. Meier, Friedegund Hobelsberger, Sarah Gellrich, Frank F. Sergon, Mildred Hauschild, Axel French, Lars E. Heinzerling, Lucie Schlager, Justin G. Ghoreschi, Kamran Schlaak, Max Hilke, Franz J. Poch, Gabriela Korsing, Sören Berking, Carola Heppt, Markus V. Erdmann, Michael Haferkamp, Sebastian Drexler, Konstantin Schadendorf, Dirk Sondermann, Wiebke Goebeler, Matthias Schilling, Bastian Kather, Jakob N. Krieghoff-Henning, Eva Brinker, Titus J. Image and Video Processing Computer Vision and Pattern Recognition Machine Learning Background: Convolutional neural network (CNN)-based melanoma classifiers face several challenges that limit their usefulness in clinical practice. Objective: To investigate the impact of multiple real-world dermoscopic views of a single lesion of interest on a CNN-based melanoma classifier. Methods: This study evaluated 656 suspected melanoma lesions. Classifier performance was measured using area under the receiver operating characteristic curve (AUROC), expected calibration error (ECE) and maximum confidence change (MCC) for (I) a single-view scenario, (II) a multiview scenario using multiple artificially modified images per lesion and (III) a multiview scenario with multiple real-world images per lesion. Results: The multiview approach with real-world images significantly increased the AUROC from 0.905 (95% CI, 0.879-0.929) in the single-view approach to 0.930 (95% CI, 0.909-0.951). ECE and MCC also improved significantly from 0.131 (95% CI, 0.105-0.159) to 0.072 (95% CI: 0.052-0.093) and from 0.149 (95% CI, 0.125-0.171) to 0.115 (95% CI: 0.099-0.131), respectively. Comparing multiview real-world to artificially modified images showed comparable diagnostic accuracy and uncertainty estimation, but significantly worse robustness for the latter. Conclusion: Using multiple real-world images is an inexpensive method to positively impact the performance of a CNN-based melanoma classifier. |
| title | Using Multiple Dermoscopic Photographs of One Lesion Improves Melanoma Classification via Deep Learning: A Prognostic Diagnostic Accuracy Study |
| topic | Image and Video Processing Computer Vision and Pattern Recognition Machine Learning |
| url | https://arxiv.org/abs/2306.02800 |