Professor Seung-soo Lee's team from the Department of Radiology at Seoul Asan Medical Center and Smart Care Works jointly conducted automatic image analysis using AI, and the results are expected to be clinically utilized in various fields as it drastically reduced time and enabled fast and accurate liver volume evaluation.

SmartCareWorks recently conducted a joint study with Professor Seung-Soo Lee's team from the Department of Radiology at Seoul Asan Medical Center using a deep learning algorithm to automatically segment the liver and spleen in portal phase CT images, and announced on the 8th that the deep learning algorithm dramatically reduced the time for CT volumetry.

This algorithm can segment the liver and spleen with an accuracy of dice score of 97% or higher (Korean J Radiol. (2020;21:987-997), it was installed as a GoCDSS module in GoWIX, a web PACS of Smart Care Works, and used for research.

In this study, we used this deep learning algorithm to measure the volume of the liver and spleen by analyzing CT images of a large study population (approximately 3,400 people).

This algorithm is expected to be used clinically in various fields.

Preoperative assessment of residual liver volume before hepatectomy or living donor liver transplantation is a necessary preoperative evaluation process for safe surgery that minimizes complications associated with surgery. It is predicted that rapid and accurate liver volume assessment will be possible by replacing existing manual or semi-automatic software-based CT volumetry.

In addition, SmartCareWorks' GoWIX, a Web PACS equipped with deep learning, performs image analysis through deep learning along with the transmission of CT image examinations, allowing doctors to review the liver and spleen segmentation results and images together during the image interpretation stage.

Therefore, if deep learning algorithms are well integrated into the imaging examination and interpretation process, Professor Seung-soo Lee said that in the near future, the size of the liver or spleen volume may be determined by comparing the volume values measured by CT or MRI with a reference standard, rather than by indirect methods such as subjective assessment or diameter measurement by the reader.

In addition, it appears that the volume of the liver and spleen can be used to predict the severity and prognosis of chronic liver disease patients. Against this background, this study presented personalized reference intervals for liver and spleen volumes considering the patient's age, sex, height, and weight in a large group of healthy prospective liver donors.

This is a result obtained for Koreans, and the value can be calculated using GoWIX, a web-based calculator developed by Smart Care Works, so it is a result that can be applied directly to treatment. Professor Lee Seung-su said.

Professor Seung-soo Lee's research using GoWIX of Smart Care Works was published in 'Radiology', and Smart Care Works CEO Jeong-beom Cheon stated that the journal's review of the research was quite favorable, and that he acknowledged the excellence of the research results to the extent that they directly quoted the paper's claim, saying, "Automatic measurement of liver and spleen volume using deep learning and determination of normal or abnormal based on volume will soon be used clinically."

This study, which began in 2016 as a government project (Project name: Development and commercialization of a decision support system based on image informatics using big data for intractable liver disease associated with metabolic syndrome), ended in March of this year, but Professor Seung-soo Lee's team and Smart Care Works announced that they plan to expand this research and pursue commercialization.