Abstract
Aim: In this study, we aimed to find out the usefulness of artificial intelligence for the histological diagnosis of products of conception by identifying the chorionic villi in the tissue specimens.
Materials and Methods: A total of 400 anonymized digital images were acquired, which were divided into two groups. Group 1 included 200 images containing chorionic villi while Group 2 was comprised of 200 images of decidual tissue. Two variants of two deep learning computer vision algorithms VGG-16, VGG-19, and Resnet-18, Resnet-34 had been applied for the evaluation and analysis of the digital pathology image.
Results: The application of deep learning computer vision algorithms, VGG-16, VGG-19, Resnet-18, and Resnet-34 revealed the diagnostic accuracy of 95.3%, 99.4%, 98.1 %, 98.2% and F1-Score of 0.989, 0.989, 0.984 and 0.989 on test data respectively.
Discussion: The specimens of products of conception are quite frequently received for histopathological evaluation by the pathology laboratory. A careful histological examination is required for the identification of chorionic villi in the submitted specimens. The present study revealed that the application of artificial intelligence could be valuable assistance to the histopathologists for the microscopic examination of biopsy specimens. The diagnostic accuracy achieved in the present study is quite close to the reported figures for the diagnosis of lung and prostatic cancer with the help of deep learning in other series. The present study revealed that computer vision-based system may be an effective adjunct tool for the histopathological detection of chorionic villi.