Machine Learning Based Prediction of Squamous Cell Carcinoma in Ex Vivo Confocal Laser Scanning Microscopy
Image classification with convolutional neural networks (CNN) offers an unprecedented opportunity to medical imaging. Regulatory agencies in the USA and Europe have already cleared numerous deep learning/machine learning based medical devices and algorithms. While the field of radiology is on the fo...
| Main Authors: | , , , , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2021-11-01
|
| Series: | Cancers |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-6694/13/21/5522 |
| _version_ | 1797512750704885760 |
|---|---|
| author | Cristel Ruini Sophia Schlingmann Žan Jonke Pinar Avci Víctor Padrón-Laso Florian Neumeier Istvan Koveshazi Ikenna U. Ikeliani Kathrin Patzer Elena Kunrad Benjamin Kendziora Elke Sattler Lars E. French Daniela Hartmann |
| author_facet | Cristel Ruini Sophia Schlingmann Žan Jonke Pinar Avci Víctor Padrón-Laso Florian Neumeier Istvan Koveshazi Ikenna U. Ikeliani Kathrin Patzer Elena Kunrad Benjamin Kendziora Elke Sattler Lars E. French Daniela Hartmann |
| author_sort | Cristel Ruini |
| collection | DOAJ |
| description | Image classification with convolutional neural networks (CNN) offers an unprecedented opportunity to medical imaging. Regulatory agencies in the USA and Europe have already cleared numerous deep learning/machine learning based medical devices and algorithms. While the field of radiology is on the forefront of artificial intelligence (AI) revolution, conventional pathology, which commonly relies on examination of tissue samples on a glass slide, is falling behind in leveraging this technology. On the other hand, ex vivo confocal laser scanning microscopy (ex vivo CLSM), owing to its digital workflow features, has a high potential to benefit from integrating AI tools into the assessment and decision-making process. Aim of this work was to explore a preliminary application of CNN in digitally stained ex vivo CLSM images of cutaneous squamous cell carcinoma (cSCC) for automated detection of tumor tissue. Thirty-four freshly excised tissue samples were prospectively collected and examined immediately after resection. After the histologically confirmed ex vivo CLSM diagnosis, the tumor tissue was annotated for segmentation by experts, in order to train the MobileNet CNN. The model was then trained and evaluated using cross validation. The overall sensitivity and specificity of the deep neural network for detecting cSCC and tumor free areas on ex vivo CLSM slides compared to expert evaluation were 0.76 and 0.91, respectively. The area under the ROC curve was equal to 0.90 and the area under the precision-recall curve was 0.85. The results demonstrate a high potential of deep learning models to detect cSCC regions on digitally stained ex vivo CLSM slides and to distinguish them from tumor-free skin. |
| first_indexed | 2024-03-10T06:05:06Z |
| format | Article |
| id | doaj.art-f55b4b793c9c407f97c8de4c2c434cba |
| institution | Directory Open Access Journal |
| issn | 2072-6694 |
| language | English |
| last_indexed | 2024-03-10T06:05:06Z |
| publishDate | 2021-11-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Cancers |
| spelling | doaj.art-f55b4b793c9c407f97c8de4c2c434cba2023-11-22T20:36:29ZengMDPI AGCancers2072-66942021-11-011321552210.3390/cancers13215522Machine Learning Based Prediction of Squamous Cell Carcinoma in Ex Vivo Confocal Laser Scanning MicroscopyCristel Ruini0Sophia Schlingmann1Žan Jonke2Pinar Avci3Víctor Padrón-Laso4Florian Neumeier5Istvan Koveshazi6Ikenna U. Ikeliani7Kathrin Patzer8Elena Kunrad9Benjamin Kendziora10Elke Sattler11Lars E. French12Daniela Hartmann13Department of Dermatology and Allergy, University Hospital, LMU Munich, 80337 Munich, GermanyDepartment of Dermatology and Allergy, University Hospital, LMU Munich, 80337 Munich, GermanyMunich Innovation Labs GmbH, 80336 Munich, GermanyDepartment of Dermatology and Allergy, University Hospital, LMU Munich, 80337 Munich, GermanyMunich Innovation Labs GmbH, 80336 Munich, GermanyM3i Industry-in-Clinic-Platform GmbH, 80336 Munich, GermanyM3i Industry-in-Clinic-Platform GmbH, 80336 Munich, GermanyM3i Industry-in-Clinic-Platform GmbH, 80336 Munich, GermanyDepartment of Dermatology and Allergy, University Hospital, LMU Munich, 80337 Munich, GermanyDepartment of Dermatology and Allergy, University Hospital, LMU Munich, 80337 Munich, GermanyDepartment of Dermatology and Allergy, University Hospital, LMU Munich, 80337 Munich, GermanyDepartment of Dermatology and Allergy, University Hospital, LMU Munich, 80337 Munich, GermanyDepartment of Dermatology and Allergy, University Hospital, LMU Munich, 80337 Munich, GermanyDepartment of Dermatology and Allergy, University Hospital, LMU Munich, 80337 Munich, GermanyImage classification with convolutional neural networks (CNN) offers an unprecedented opportunity to medical imaging. Regulatory agencies in the USA and Europe have already cleared numerous deep learning/machine learning based medical devices and algorithms. While the field of radiology is on the forefront of artificial intelligence (AI) revolution, conventional pathology, which commonly relies on examination of tissue samples on a glass slide, is falling behind in leveraging this technology. On the other hand, ex vivo confocal laser scanning microscopy (ex vivo CLSM), owing to its digital workflow features, has a high potential to benefit from integrating AI tools into the assessment and decision-making process. Aim of this work was to explore a preliminary application of CNN in digitally stained ex vivo CLSM images of cutaneous squamous cell carcinoma (cSCC) for automated detection of tumor tissue. Thirty-four freshly excised tissue samples were prospectively collected and examined immediately after resection. After the histologically confirmed ex vivo CLSM diagnosis, the tumor tissue was annotated for segmentation by experts, in order to train the MobileNet CNN. The model was then trained and evaluated using cross validation. The overall sensitivity and specificity of the deep neural network for detecting cSCC and tumor free areas on ex vivo CLSM slides compared to expert evaluation were 0.76 and 0.91, respectively. The area under the ROC curve was equal to 0.90 and the area under the precision-recall curve was 0.85. The results demonstrate a high potential of deep learning models to detect cSCC regions on digitally stained ex vivo CLSM slides and to distinguish them from tumor-free skin.https://www.mdpi.com/2072-6694/13/21/5522squamous cell carcinomaex vivo confocal laser scanning microscopyreflectance confocal microscopyfluorescence confocal microscopydigital pathologydigital staining |
| spellingShingle | Cristel Ruini Sophia Schlingmann Žan Jonke Pinar Avci Víctor Padrón-Laso Florian Neumeier Istvan Koveshazi Ikenna U. Ikeliani Kathrin Patzer Elena Kunrad Benjamin Kendziora Elke Sattler Lars E. French Daniela Hartmann Machine Learning Based Prediction of Squamous Cell Carcinoma in Ex Vivo Confocal Laser Scanning Microscopy Cancers squamous cell carcinoma ex vivo confocal laser scanning microscopy reflectance confocal microscopy fluorescence confocal microscopy digital pathology digital staining |
| title | Machine Learning Based Prediction of Squamous Cell Carcinoma in Ex Vivo Confocal Laser Scanning Microscopy |
| title_full | Machine Learning Based Prediction of Squamous Cell Carcinoma in Ex Vivo Confocal Laser Scanning Microscopy |
| title_fullStr | Machine Learning Based Prediction of Squamous Cell Carcinoma in Ex Vivo Confocal Laser Scanning Microscopy |
| title_full_unstemmed | Machine Learning Based Prediction of Squamous Cell Carcinoma in Ex Vivo Confocal Laser Scanning Microscopy |
| title_short | Machine Learning Based Prediction of Squamous Cell Carcinoma in Ex Vivo Confocal Laser Scanning Microscopy |
| title_sort | machine learning based prediction of squamous cell carcinoma in ex vivo confocal laser scanning microscopy |
| topic | squamous cell carcinoma ex vivo confocal laser scanning microscopy reflectance confocal microscopy fluorescence confocal microscopy digital pathology digital staining |
| url | https://www.mdpi.com/2072-6694/13/21/5522 |
| work_keys_str_mv | AT cristelruini machinelearningbasedpredictionofsquamouscellcarcinomainexvivoconfocallaserscanningmicroscopy AT sophiaschlingmann machinelearningbasedpredictionofsquamouscellcarcinomainexvivoconfocallaserscanningmicroscopy AT zanjonke machinelearningbasedpredictionofsquamouscellcarcinomainexvivoconfocallaserscanningmicroscopy AT pinaravci machinelearningbasedpredictionofsquamouscellcarcinomainexvivoconfocallaserscanningmicroscopy AT victorpadronlaso machinelearningbasedpredictionofsquamouscellcarcinomainexvivoconfocallaserscanningmicroscopy AT florianneumeier machinelearningbasedpredictionofsquamouscellcarcinomainexvivoconfocallaserscanningmicroscopy AT istvankoveshazi machinelearningbasedpredictionofsquamouscellcarcinomainexvivoconfocallaserscanningmicroscopy AT ikennauikeliani machinelearningbasedpredictionofsquamouscellcarcinomainexvivoconfocallaserscanningmicroscopy AT kathrinpatzer machinelearningbasedpredictionofsquamouscellcarcinomainexvivoconfocallaserscanningmicroscopy AT elenakunrad machinelearningbasedpredictionofsquamouscellcarcinomainexvivoconfocallaserscanningmicroscopy AT benjaminkendziora machinelearningbasedpredictionofsquamouscellcarcinomainexvivoconfocallaserscanningmicroscopy AT elkesattler machinelearningbasedpredictionofsquamouscellcarcinomainexvivoconfocallaserscanningmicroscopy AT larsefrench machinelearningbasedpredictionofsquamouscellcarcinomainexvivoconfocallaserscanningmicroscopy AT danielahartmann machinelearningbasedpredictionofsquamouscellcarcinomainexvivoconfocallaserscanningmicroscopy |