Radiomics Diagnostic Tool Based on Deep Learning for Colposcopy Image Classification

Abstract

Background: Colposcopy imaging is widely used to diagnose, treat and follow-up on premalignant and malignant lesions in the vulva, vagina, and cervix. Thus, deep learning algorithms are being used widely in cervical cancer diagnosis tools. In this study, we developed and preliminarily validated a model based on the Unet network plus SVM to classify cervical lesions on colposcopy images. Methodology: Two sets of images were used: the Intel & Mobile ODT Cervical Cancer Screening public dataset, and a private dataset from a public hospital in Ecuador during a routine colposcopy, after the application of acetic acid and lugol. For the latter, the corresponding clinical information was collected, specifically cytology on the PAP smear and the screening of human papillomavirus testing, prior to colposcopy. The lesions of the cervix or regions of interest were segmented and classified by the Unet and the SVM model, respectively. Results: The CAD system was evaluated for the ability to predict the risk of cervical cancer. The lesion segmentation metric results indicate a DICE of 50%, a precision of 65%, and an accuracy of 80%. The classification results’ sensitivity, specificity, and accuracy were 70%, 48.8%, and 58%, respectively. Randomly, 20 images were selected and sent to 13 expert colposcopists for a statistical comparison between visual evaluation experts and the CAD tool (p-value of 0.597). Conclusion: The CAD system needs to improve but could be acceptable in an environment where women have limited access to clinicians for the diagnosis, follow-up, and treatment of cervical cancer; better performance is possible through the exploration of other deep learning methods with larger datasets.

Computed-Assisted Diagnosis in Diagnostic Cervical Cancer Imaging : A Ten Years Systematic Review

“…Gaona, YJ et al (2022) 7 showed that synthetic images improve the colposcopy diagnosis performance of our AI tool for the classification of cervix images. We propose a combined Deep Learning method (Unet) plus Machine Learning (SVM) to obtain the best index of identification, classification, and prediction of cervix abnormalities through the processing of images.…”

Colposcopic Image Segmentation Based on Feature Refinement and Attention

“…Contrary to most previous studies that conducted preprocessing on colposcopic images [44]- [46], this study advocates for end-to-end segmentation of the region of interest on unaltered colposcopic images, enabling clinicians to make diagnoses based on authentic image data. Given the often blurred borders and complex gradients in raw colposcopic images, high-resolution information is essential for segmenting lesion contours and refining edge details.…”

“…Subsequent research has extensively explored various models for medical image segmentation based on U-net. For example, Bai et al [44] segmented colposcopic images using the U-Net model by replacing the fully connected layer with a convolutional layer and employing a deconvolution structure for data upsampling; Zhang et al [45] substituted the convolutional layer in U-Net with a pooling layer and added a dropout layer for colposcopic image segmentation; Yuliana et al [46] employed U-Net to segment ROI in colposcopic images. Additionally, Liu et al [8] compared U-Net, FCN, and SEGNet [47] in CIN image segmentation tasks, finding that U-Net delivers precise edge segmentation.…”