Automated Skin Lesion Classification on Ultrasound Images
The growing incidence of skin cancer makes computer-aided diagnosis tools for this group of diseases increasingly important. The use of ultrasound has the potential to complement information from optical dermoscopy. The current work presents a fully automatic classification framework utilizing fully...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2021-07-01
|
| Series: | Diagnostics |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2075-4418/11/7/1207 |
| _version_ | 1827687753794453504 |
|---|---|
| author | Péter Marosán-Vilimszky Klára Szalai András Horváth Domonkos Csabai Krisztián Füzesi Gergely Csány Miklós Gyöngy |
| author_facet | Péter Marosán-Vilimszky Klára Szalai András Horváth Domonkos Csabai Krisztián Füzesi Gergely Csány Miklós Gyöngy |
| author_sort | Péter Marosán-Vilimszky |
| collection | DOAJ |
| description | The growing incidence of skin cancer makes computer-aided diagnosis tools for this group of diseases increasingly important. The use of ultrasound has the potential to complement information from optical dermoscopy. The current work presents a fully automatic classification framework utilizing fully-automated (FA) segmentation and compares it with classification using two semi-automated (SA) segmentation methods. Ultrasound recordings were taken from a total of 310 lesions (70 melanoma, 130 basal cell carcinoma and 110 benign nevi). A support vector machine (SVM) model was trained on 62 features, with ten-fold cross-validation. Six classification tasks were considered, namely all the possible permutations of one class versus one or two remaining classes. The receiver operating characteristic (<i>ROC</i>) area under the curve (<i>AUC</i>) as well as the accuracy (<i>ACC</i>) were measured. The best classification was obtained for the classification of nevi from cancerous lesions (melanoma, basal cell carcinoma), with <i>AUC</i>s of over 90% and <i>ACC</i>s of over 85% obtained with all segmentation methods. Previous works have either not implemented FA ultrasound-based skin cancer classification (making diagnosis more lengthy and operator-dependent), or are unclear in their classification results. Furthermore, the current work is the first to assess the effect of implementing FA instead of SA classification, with FA classification never degrading performance (in terms of <i>AUC</i> or <i>ACC</i>) by more than 5%. |
| first_indexed | 2024-03-10T09:42:28Z |
| format | Article |
| id | doaj.art-d1c7ecd3574e4ef39dcd5eb61e6cb171 |
| institution | Directory Open Access Journal |
| issn | 2075-4418 |
| language | English |
| last_indexed | 2024-03-10T09:42:28Z |
| publishDate | 2021-07-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Diagnostics |
| spelling | doaj.art-d1c7ecd3574e4ef39dcd5eb61e6cb1712023-11-22T03:34:19ZengMDPI AGDiagnostics2075-44182021-07-01117120710.3390/diagnostics11071207Automated Skin Lesion Classification on Ultrasound ImagesPéter Marosán-Vilimszky 0Klára Szalai 1András Horváth 2Domonkos Csabai 3Krisztián Füzesi 4Gergely Csány 5Miklós Gyöngy 6Faculty of Information Technology and Bionics, Pázmány Péter Catholic University, Práter u. 50/A, 1083 Budapest, HungaryDepartment of Dermatology, Venereology and Dermatooncology, Semmelweis University, Mária u. 41, 1085 Budapest, HungaryFaculty of Information Technology and Bionics, Pázmány Péter Catholic University, Práter u. 50/A, 1083 Budapest, HungaryDermus Kft., Sopron út 64, 1116 Budapest, HungaryDermus Kft., Sopron út 64, 1116 Budapest, HungaryDermus Kft., Sopron út 64, 1116 Budapest, HungaryFaculty of Information Technology and Bionics, Pázmány Péter Catholic University, Práter u. 50/A, 1083 Budapest, HungaryThe growing incidence of skin cancer makes computer-aided diagnosis tools for this group of diseases increasingly important. The use of ultrasound has the potential to complement information from optical dermoscopy. The current work presents a fully automatic classification framework utilizing fully-automated (FA) segmentation and compares it with classification using two semi-automated (SA) segmentation methods. Ultrasound recordings were taken from a total of 310 lesions (70 melanoma, 130 basal cell carcinoma and 110 benign nevi). A support vector machine (SVM) model was trained on 62 features, with ten-fold cross-validation. Six classification tasks were considered, namely all the possible permutations of one class versus one or two remaining classes. The receiver operating characteristic (<i>ROC</i>) area under the curve (<i>AUC</i>) as well as the accuracy (<i>ACC</i>) were measured. The best classification was obtained for the classification of nevi from cancerous lesions (melanoma, basal cell carcinoma), with <i>AUC</i>s of over 90% and <i>ACC</i>s of over 85% obtained with all segmentation methods. Previous works have either not implemented FA ultrasound-based skin cancer classification (making diagnosis more lengthy and operator-dependent), or are unclear in their classification results. Furthermore, the current work is the first to assess the effect of implementing FA instead of SA classification, with FA classification never degrading performance (in terms of <i>AUC</i> or <i>ACC</i>) by more than 5%.https://www.mdpi.com/2075-4418/11/7/1207skin ultrasoundcomputer visioncomputer-aided diagnosisskin lesion classification |
| spellingShingle | Péter Marosán-Vilimszky Klára Szalai András Horváth Domonkos Csabai Krisztián Füzesi Gergely Csány Miklós Gyöngy Automated Skin Lesion Classification on Ultrasound Images Diagnostics skin ultrasound computer vision computer-aided diagnosis skin lesion classification |
| title | Automated Skin Lesion Classification on Ultrasound Images |
| title_full | Automated Skin Lesion Classification on Ultrasound Images |
| title_fullStr | Automated Skin Lesion Classification on Ultrasound Images |
| title_full_unstemmed | Automated Skin Lesion Classification on Ultrasound Images |
| title_short | Automated Skin Lesion Classification on Ultrasound Images |
| title_sort | automated skin lesion classification on ultrasound images |
| topic | skin ultrasound computer vision computer-aided diagnosis skin lesion classification |
| url | https://www.mdpi.com/2075-4418/11/7/1207 |
| work_keys_str_mv | AT petermarosanvilimszky automatedskinlesionclassificationonultrasoundimages AT klaraszalai automatedskinlesionclassificationonultrasoundimages AT andrashorvath automatedskinlesionclassificationonultrasoundimages AT domonkoscsabai automatedskinlesionclassificationonultrasoundimages AT krisztianfuzesi automatedskinlesionclassificationonultrasoundimages AT gergelycsany automatedskinlesionclassificationonultrasoundimages AT miklosgyongy automatedskinlesionclassificationonultrasoundimages |