Convolutional neural networks for accurate real-time diagnosis of oral epithelial dysplasia and oral squamous cell carcinoma using high-resolution in vivo confocal microscopy
Abstract Oral cancer detection is based on biopsy histopathology, however with digital microscopy imaging technology there is real potential for rapid multi-site imaging and simultaneous diagnostic analysis. Fifty-nine patients with oral mucosal abnormalities were imaged in vivo with a confocal lase...
Saved in:
| Main Authors: | , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2025-01-01
|
| Series: | Scientific Reports |
| Subjects: | |
| Online Access: | https://doi.org/10.1038/s41598-025-86400-5 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832585785451216896 |
|---|---|
| author | Rishi S. Ramani Ivy Tan Lindsay Bussau Lorraine A. O’Reilly John Silke Christopher Angel Antonio Celentano Lachlan Whitehead Michael McCullough Tami Yap |
| author_facet | Rishi S. Ramani Ivy Tan Lindsay Bussau Lorraine A. O’Reilly John Silke Christopher Angel Antonio Celentano Lachlan Whitehead Michael McCullough Tami Yap |
| author_sort | Rishi S. Ramani |
| collection | DOAJ |
| description | Abstract Oral cancer detection is based on biopsy histopathology, however with digital microscopy imaging technology there is real potential for rapid multi-site imaging and simultaneous diagnostic analysis. Fifty-nine patients with oral mucosal abnormalities were imaged in vivo with a confocal laser endomicroscope using the contrast agents acriflavine and fluorescein for the detection of oral epithelial dysplasia and oral cancer. To analyse the 9168 images frames obtained, three tandem applied pre-trained Inception-V3 convolutional neural network (CNN) models were developed using transfer learning in the PyTorch framework. The first CNN was used to filter for image quality, followed by image specific diagnostic triage models for fluorescein and acriflavine, respectively. Images were categorised based on a histopathological diagnosis into 4 categories: no dysplasia, lichenoid lesions, low-grade dysplasia and high-grade dysplasia/oral squamous cell carcinoma (OSCC). The quality filtering model had an accuracy of 89.5%. The acriflavine diagnostic model performed well for identifying lichenoid (AUC = 0.94) and low-grade dysplasia (AUC = 0.91) but poorly for identifying no dysplasia (AUC = 0.44) or high-grade dysplasia/OSCC (AUC = 0.28). In contrast, the fluorescein diagnostic model had high classification performance for all diagnostic classes (AUC range = 0.90–0.96). These models had a rapid classification speed of less than 1/10th of a second per image. Our study suggests that tandem CNNs can provide highly accurate and rapid real-time diagnostic triage for in vivo assessment of high-risk oral mucosal disease. |
| format | Article |
| id | doaj-art-82ce7086272e4866a77a6140b7d01eac |
| institution | Kabale University |
| issn | 2045-2322 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | Scientific Reports |
| spelling | doaj-art-82ce7086272e4866a77a6140b7d01eac2025-01-26T12:31:42ZengNature PortfolioScientific Reports2045-23222025-01-0115111310.1038/s41598-025-86400-5Convolutional neural networks for accurate real-time diagnosis of oral epithelial dysplasia and oral squamous cell carcinoma using high-resolution in vivo confocal microscopyRishi S. Ramani0Ivy Tan1Lindsay Bussau2Lorraine A. O’Reilly3John Silke4Christopher Angel5Antonio Celentano6Lachlan Whitehead7Michael McCullough8Tami Yap9Melbourne Dental School, University of MelbourneMelbourne Dental School, University of MelbourneOptiscan Imaging LtdWalter and Eliza Hall InstituteWalter and Eliza Hall InstitutePeter MacCallum Cancer CentreMelbourne Dental School, University of MelbourneWalter and Eliza Hall InstituteMelbourne Dental School, University of MelbourneMelbourne Dental School, University of MelbourneAbstract Oral cancer detection is based on biopsy histopathology, however with digital microscopy imaging technology there is real potential for rapid multi-site imaging and simultaneous diagnostic analysis. Fifty-nine patients with oral mucosal abnormalities were imaged in vivo with a confocal laser endomicroscope using the contrast agents acriflavine and fluorescein for the detection of oral epithelial dysplasia and oral cancer. To analyse the 9168 images frames obtained, three tandem applied pre-trained Inception-V3 convolutional neural network (CNN) models were developed using transfer learning in the PyTorch framework. The first CNN was used to filter for image quality, followed by image specific diagnostic triage models for fluorescein and acriflavine, respectively. Images were categorised based on a histopathological diagnosis into 4 categories: no dysplasia, lichenoid lesions, low-grade dysplasia and high-grade dysplasia/oral squamous cell carcinoma (OSCC). The quality filtering model had an accuracy of 89.5%. The acriflavine diagnostic model performed well for identifying lichenoid (AUC = 0.94) and low-grade dysplasia (AUC = 0.91) but poorly for identifying no dysplasia (AUC = 0.44) or high-grade dysplasia/OSCC (AUC = 0.28). In contrast, the fluorescein diagnostic model had high classification performance for all diagnostic classes (AUC range = 0.90–0.96). These models had a rapid classification speed of less than 1/10th of a second per image. Our study suggests that tandem CNNs can provide highly accurate and rapid real-time diagnostic triage for in vivo assessment of high-risk oral mucosal disease.https://doi.org/10.1038/s41598-025-86400-5Oral cancerEarly diagnosisDigital microscopyDeep learning |
| spellingShingle | Rishi S. Ramani Ivy Tan Lindsay Bussau Lorraine A. O’Reilly John Silke Christopher Angel Antonio Celentano Lachlan Whitehead Michael McCullough Tami Yap Convolutional neural networks for accurate real-time diagnosis of oral epithelial dysplasia and oral squamous cell carcinoma using high-resolution in vivo confocal microscopy Scientific Reports Oral cancer Early diagnosis Digital microscopy Deep learning |
| title | Convolutional neural networks for accurate real-time diagnosis of oral epithelial dysplasia and oral squamous cell carcinoma using high-resolution in vivo confocal microscopy |
| title_full | Convolutional neural networks for accurate real-time diagnosis of oral epithelial dysplasia and oral squamous cell carcinoma using high-resolution in vivo confocal microscopy |
| title_fullStr | Convolutional neural networks for accurate real-time diagnosis of oral epithelial dysplasia and oral squamous cell carcinoma using high-resolution in vivo confocal microscopy |
| title_full_unstemmed | Convolutional neural networks for accurate real-time diagnosis of oral epithelial dysplasia and oral squamous cell carcinoma using high-resolution in vivo confocal microscopy |
| title_short | Convolutional neural networks for accurate real-time diagnosis of oral epithelial dysplasia and oral squamous cell carcinoma using high-resolution in vivo confocal microscopy |
| title_sort | convolutional neural networks for accurate real time diagnosis of oral epithelial dysplasia and oral squamous cell carcinoma using high resolution in vivo confocal microscopy |
| topic | Oral cancer Early diagnosis Digital microscopy Deep learning |
| url | https://doi.org/10.1038/s41598-025-86400-5 |
| work_keys_str_mv | AT rishisramani convolutionalneuralnetworksforaccuraterealtimediagnosisoforalepithelialdysplasiaandoralsquamouscellcarcinomausinghighresolutioninvivoconfocalmicroscopy AT ivytan convolutionalneuralnetworksforaccuraterealtimediagnosisoforalepithelialdysplasiaandoralsquamouscellcarcinomausinghighresolutioninvivoconfocalmicroscopy AT lindsaybussau convolutionalneuralnetworksforaccuraterealtimediagnosisoforalepithelialdysplasiaandoralsquamouscellcarcinomausinghighresolutioninvivoconfocalmicroscopy AT lorraineaoreilly convolutionalneuralnetworksforaccuraterealtimediagnosisoforalepithelialdysplasiaandoralsquamouscellcarcinomausinghighresolutioninvivoconfocalmicroscopy AT johnsilke convolutionalneuralnetworksforaccuraterealtimediagnosisoforalepithelialdysplasiaandoralsquamouscellcarcinomausinghighresolutioninvivoconfocalmicroscopy AT christopherangel convolutionalneuralnetworksforaccuraterealtimediagnosisoforalepithelialdysplasiaandoralsquamouscellcarcinomausinghighresolutioninvivoconfocalmicroscopy AT antoniocelentano convolutionalneuralnetworksforaccuraterealtimediagnosisoforalepithelialdysplasiaandoralsquamouscellcarcinomausinghighresolutioninvivoconfocalmicroscopy AT lachlanwhitehead convolutionalneuralnetworksforaccuraterealtimediagnosisoforalepithelialdysplasiaandoralsquamouscellcarcinomausinghighresolutioninvivoconfocalmicroscopy AT michaelmccullough convolutionalneuralnetworksforaccuraterealtimediagnosisoforalepithelialdysplasiaandoralsquamouscellcarcinomausinghighresolutioninvivoconfocalmicroscopy AT tamiyap convolutionalneuralnetworksforaccuraterealtimediagnosisoforalepithelialdysplasiaandoralsquamouscellcarcinomausinghighresolutioninvivoconfocalmicroscopy |