Spontaneous abortion, commonly known as miscarriage, is a significant concern during early pregnancy. Histopathological examination of tissue samples is a widely used method to diagnose and classify tissue phenotypes found in products of conception (POC) after spontaneous abortion. Background: Histopathological examination is subjective and dependent on the skill and experience of the examiner. In recent years, artificial intelligence (AI)-based techniques have emerged as a promising tool in medical imaging, offering the potential to revolutionize tissue phenotyping and improve the accuracy and reliability of the histopathological examination process. The goal of this study was to investigate the use of AI techniques for the detection of various tissue phenotypes in POC after spontaneous abortion and evaluate the accuracy and reliability of these techniques compared to traditional manual methods. Methods: We present a novel publicly available dataset named HistoPoC, which is believed to be the first of its kind, focusing on spontaneous abortion (miscarriage) in early pregnancy. A diverse dataset of 5666 annotated images was prepared from previously diagnosed cases of POC from Atia General Hospital, Karachi, Pakistan, for this purpose. The digital images were prepared at 10× through a camera-connected microscope by a consultant histopathologist. Results: The dataset's effectiveness was validated using several deep learning-based models, demonstrating its applicability and supporting its use in intelligent diagnostic systems. Conclusions: The insights gained from this study could illuminate the causes of spontaneous abortion and guide the development of novel treatments. Additionally, this study could contribute to advancements in the field of tissue phenotyping and the wider application of deep learning techniques in medical diagnostics and treatment.
Keywords: artificial intelligence; deep learning; spontaneous abortion; tissue phenotyping.