Artificial intelligence (AI), particularly deep learning (DL), has found growing applications in the field of medical imaging and healthcare. A substantial portion of the research related to DL has concentrated on retrospectively assessing model performance using validated datasets with known ground-truth labels. Few studies have taken the next step to investigate how DL assistance influences the diagnostic abilities of sonologists, and even fewer have explored the most effective ways in which DL can assist in clinical diagnosis.
In the current study, a multi-reader, crossover randomized controlled trial (RCT) was conducted, involving the recruitment of 36 sonologists. They were tasked with interpreting fetal neurosonographic images and videos both without the assistance of the PAICS system and with the aid of PAICS in two different modes. The primary objective was to assess the effectiveness of PAICS in supporting the diagnosis of fetal intracranial malformations and to compare it with other auxiliary diagnostic methods.
The findings of this research highlight that the two image and video reading modes, augmented by the deep learning capabilities of the PAICS system, substantially enhance the accuracy of CNS malformation classification. This suggests that the system holds significant promise in enhancing the diagnostic performance of sonologists when it comes to detecting fetal intracranial malformations.
During the course of the research, a total of 734 fetuses with abnormal intracranial findings and 19,709 normal fetuses underwent scanning. However, 254 fetuses with abnormal findings and 19,631 normal fetuses were excluded due to issues like image quality or redundancy. Ultimately, 709 original images and videos (549 images and 160 videos) from 558 fetuses met the inclusion criteria and were included in the study.
The trial findings suggest that PAICS has the potential to enhance the diagnostic performance of sonologists in identifying fetal intracranial malformations from neurosonographic data, whether utilized concurrently or in a secondary mode. It’s worth noting that the concurrent use of PAICS proved to be more effective for all readers. Further research in real clinical settings, with a larger number of cases, is warranted to thoroughly assess the assistance provided by PAICS in the detection of congenital intracranial malformations.
Check out the Paper. All Credit For This Research Goes To the Researchers on This Project. Also, don’t forget to join our 32k+ ML SubReddit, 40k+ Facebook Community, Discord Channel, and Email Newsletter, where we share the latest AI research news, cool AI projects, and more.
If you like our work, you will love our newsletter..
We are also on Telegram and WhatsApp.
Janhavi Lande, is an Engineering Physics graduate from IIT Guwahati, class of 2023. She is an upcoming data scientist and has been working in the world of ml/ai research for the past two years. She is most fascinated by this ever changing world and its constant demand of humans to keep up with it. In her pastime she enjoys traveling, reading and writing poems.