CorelineSoft (CEO Jin-guk Kim), a medical artificial intelligence (AI) specialist, recently participated within the American Society of Cardiovascular CT 2024 and presented 4 research papers.
“Recently, there was a rise within the number of educational papers presented by medical AI firms,” said Kwak Ji-wan, head of technology at Coreline Soft. “The paper trend has definitely modified in comparison with the past.”
Up to now, the main focus was on proving AI’s performance, comparable to its accuracy meeting certain benchmarks, but now, there are a lot of cases where the advantages to patients or long-term performance are compared through ‘Leader Studies’.
Which means the medical community can also be taking note of practicality beyond reviewing the introduction of AI. Accordingly, Coreline Soft’s technology department emphasized that it’s specializing in the ‘construction process’ to enhance the convenience of introducing medical AI.
“The U.S., Germany, and Japan could be considered major markets, however the U.S. is the most important medical AI market,” said Kwak Ji-wan, head of technology. “It’s most significant to arrange an onboarding process that matches the national system.”
For instance, U.S. healthcare facilities place a high priority on data security. Even when using AI solutions on a trial basis, it takes a mean of three months to arrange documentation.
He emphasized that it will be significant to exhibit long-term usability and convenience, especially with the intention to realize AI solutions after testing.
The secret is ‘early diagnosis and rapid treatment.’ AI’s utility arises when it might probably detect diseases that would otherwise be missed with as little capital and energy as possible.
For this purpose, the primary differentiating factor that Coreline Soft has recommend is the ‘interface (UI)’.
Most existing AI solutions don’t provide a separate ‘editing viewer’ to users, but moderately receive results after inputting data.
Nonetheless, CorelineSoft provides 3D graphics for commentary, modification, and manipulation through the viewer. This can also be a component that may complement the ‘black box’ problem of AI.
This results in smooth research. Using the viewer, you may check AI interpretation results, images, and videos in 3D, and you may as well enlarge or separate each part. It makes it easier to research various solutions comparable to the brain, abdomen, and muscles.
The second is ‘kernel conversion’, a technology that minimizes shooting costs and time while maximizing the opportunity of disease detection.
The difference between the ‘heart’ and the ‘lungs’ is representative. Within the case of the lungs, it is feasible to take pictures while the patient is intentionally respiration, but the guts can’t be intentionally controlled by the patient. Since CT scans are taken in numerous modes, it’s not possible to cover each areas in a single chest scan.
Presently, it was reported that by applying AI-based kernel conversion technology to the captured images, the CT shooting time and value could be drastically reduced.
“Within the U.S., one among the largest tax expenditure areas is heart disease,” said Kwak. “The flexibility to detect heart abnormalities early with only a lung scan is appealing to the U.S., where the screening rate (the share of people who find themselves eligible for screening actually undergoing testing) is low.”
In this manner, ‘accidental disease discovery’ through AI not only brings more treatment and profit-generating opportunities to hospitals, but additionally has the structure to attenuate tax expenditures on the national level by lowering the national incidence rate.
Coreline Soft can also be specializing in such accidental discoveries. Within the elderly, injuries on account of muscle loss often result in lack of energy and even death. Presently, we’re advancing technology in order that we will analyze subcutaneous fat, visceral fat, and psoas muscles through images or video recordings, and, if mandatory, even conduct early screening for osteoporosis.
And the reason is that each one technologies come from ‘segmentation’. The goal is to expand the technology to many areas and diseases based on the principle of ‘minimum shooting, maximum segmentation’.
“Consequently of this effort, the variety of requests for demos has increased noticeably since last yr,” said Kwak Ji-wan, head of the technology division.
Along with requests from research professors, there have also been requests from American medical imaging textbook firms to make use of the pictures. The medical market places importance on the sustainability of firms and solutions, so it’s explained that continuous participation in academic conferences and research presentations have been effective.
“Up to now, I participated in academic societies that I had partial contact with, comparable to the Society of Cardiology and the Society of Radiology, but now I actually have more say within the SCCT society that focuses on cardiac CT imaging,” he said. SCCT is a spot to exhibit aortic dissection, aortic aneurysm, pulmonary hypertension, aortic valve calcium measurement, and introduce technologies which might be in preparation for approval by the U.S. Food and Drug Administration (FDA).
CorelineSoft said that nine of its AI-based product lines have received FDA approval, essentially the most amongst domestic medical AI firms.
“The technology department’s role is to attach actual technology consumers with AI solutions,” said Kwak Ji-wan, head of the technology department. “To this end, we’ve got prepared optimized product linkage scenarios for every country or hospital system, and have also developed widget-style solutions for the convenience of medical staff.”
Lastly, he emphasized that he would “maximize the convenience of medical AI solutions through ‘agile processes’ comparable to accepting feedback from the medical field and constantly striving for management.”
Reporter Jang Se-min semim99@aitimes.com
