Interpretable pap smear cell representation for cervical cancer screening

TL;DR

This paper presents an explainable deep learning method for cervical cell screening that detects abnormalities without training on abnormal samples, improving interpretability and addressing class imbalance issues.

Contribution

Introduces a one-class variational autoencoder approach for pap smear analysis that localizes abnormalities and discriminates cancerous from normal cells without using abnormal training data.

Findings

Achieves high AUC scores for SCC and HSIL detection.

Enhances clustering metrics like V-measure and homogeneity.

Effectively isolates abnormal regions without additional training.

Abstract

Screening is critical for prevention and early detection of cervical cancer but it is time-consuming and laborious. Supervised deep convolutional neural networks have been developed to automate pap smear screening and the results are promising. However, the interest in using only normal samples to train deep neural networks has increased owing to class imbalance problems and high-labeling costs that are both prevalent in healthcare. In this study, we introduce a method to learn explainable deep cervical cell representations for pap smear cytology images based on one class classification using variational autoencoders. Findings demonstrate that a score can be calculated for cell abnormality without training models with abnormal samples and localize abnormality to interpret our results with a novel metric based on absolute difference in cross entropy in agglomerative clustering. The best model that discriminates squamous cell carcinoma (SCC) from normals gives 0.908 +- 0.003 area under operating characteristic curve (AUC) and one that discriminates high-grade epithelial lesion (HSIL) 0.920 +- 0.002 AUC. Compared to other clustering methods, our method enhances the V-measure and yields higher homogeneity scores, which more effectively isolate different abnormality regions, aiding in the interpretation of our results. Evaluation using in-house and additional open dataset show that our model can discriminate abnormality without the need of additional training of deep models.