Ajou University

"AI Assistant for Radiologists" Kakao Brain introduced last month a web-based radiology report Research Use Only (RUO) service powered by its large-scale artificial intelligence (AI), "KARA-CXR." Kakao Brain plans to expand the interactive functions supporting various imaging medical examinations through the advancement of large-scale multi-modal generation AI technology. SEOUL, South Korea, Feb. 22, 2024 /PRNewswire/ -- Kakao Brain has unveiled the research use only (RUO) web-based solution powered by "KARA-CXR" (short for "Kakao Brain AI for Radiology Assistant-CXR"), a large-scale AI algorithm that instantly generates preliminary radiology reports for doctor's review when chest X-ray images are uploaded onto the web. Through this user-friendly and seamless solution, medical researchers can experience various functions such as: receiving two types of AI interpretations of chest X-ray images together (a simple report and a detailed full report); ability to upload multiple DICOM files at once and receiving multiple reports simultaneously; and being able to share the AI interpretations within the research group or medical community. It also features compliant security measures that anonymize and protect any personal information contained in DICOM files. Kakao Brain started developing 'KARA-CXR' in 2022, collaborating with 11 leading university hospitals and large medical institutions. Partners include: Ewha Womans University Medical Center (Seoul, Mok-dong), Ajou University Hospital, Soonchunhyang Medical Center (Cheonan, Gumi, Seoul), Chungnam National University Hospital, Chungbuk National University Hospital, Keimyung University Dongsan Medical Center, Inha University Hospital, Dongguk University Ilsan Hospital, among others. 'KARA-CXR' has analyzed over 16 million chest X-ray images and readings. KARA-CXR is built on Kakao Brain's healthcare-specialized large-scale generative AI model, and the research team is preparing to publish the results in influential academic journals. Kakao Brain is preparing to obtain the European Union certification and seek approval from the US FDA and Korean Ministry of Food and Drug Safety. Kakao Brain is committed to the expansion of AI-powered radiology solution to address the inefficiency in the existing interpretation and reporting process, a shortage of radiologists, and to unburden healthcare professionals from the increasing workload. As an RUO tool, KARA-CXR is currently available for research use only and is not intended to diagnose, treat, cure, or prevent any disease. Kim Il-doo, CEO of Kakao Brain, stated, "As KARA-CXR was developed using high-quality data and the latest technology involving large-scale learning and generative AI, I hope that it will eventually play a great role in enhancing the workflow for the radiologists, not to mention improving the efficiency of chest x-ray image interpretation." He also added, "We are planning to go beyond generating AI interpretations of chest X-rays by expanding the algorithms to include interactive functions that support a variety of medical imaging tests, by utilizing the advanced large-scale multi-modal generative AI technology." Kakao Brain is a Korea-based global AI technology company. Founded in 2017, Kakao Brain aims to develop AI technology that will change people's lifestyle and embraces the challenge of asking the 'Unthinkable Question' to spark innovation, allowing everyone to live better lives. Kakao Brain has developed numerous innovative AI technologies and services designed to enhance people's quality of life through continuous technological innovation. These include the large-scale language model KoGPT and the image generation model Karlo. Additionally, Kakao Brain contributes to the development of the AI technology community by releasing 'Coyo,' a dataset consisting of approximately 740 million images and text. More information about Kakao Brain can be found on their official website: . SOURCE Kakao Brain

Aug. 29, 2022 13:00 UTC SAN DIEGO--(BUSINESS WIRE)-- Genome Insight, a cancer whole genome platform company, is launching a pilot program in Ajou University Medical Center to test the use of Whole Genome Sequencing (WGS) in routine cancer care setting. Cancer patients at various stages in the treatment journey will be eligible to receive WGS testing if the clinician notices potential clinical benefits for that patient. Genome Insight will provide its proprietary cancer WGS analysis and interpretation solution to Ajou University Medical Center. WGS captures a near complete overview of genomic characteristics of a tumor in one test. It enables discovery of all types of genomic mutations, including single nucleotide variants, copy number alterations, insertions/deletions, and structural variants, many of which are targets for precise diagnosis and treatment but hardly detected by common medical tests. Potential direct benefits of cancer WGS include an opportunity to test new cancer drugs in a clinical trial and the development of new biomarkers for precision oncology. Whilst the potential of WGS as a comprehensive cancer diagnostic tool has been demonstrated in a number of studies, it is yet to be used in routine diagnostic setting due in part to its high cost, long turn-around-time and technical challenges. This pilot program will give the opportunity for clinicians at Ajou University Medical Center to use WGS as needed to achieve the best treatment outcome. Genome Insight will be providing whole-genome data production, analysis and interpretation results in a timely manner that fits into the clinician’s practice. In addition, CancerVisionTM, Genome Insight’s genome browser optimized for clinicians, will be provided in clinics to explore each patient’s cancer genome space in streamlined graphical user interfaces. Genome Insight and Ajou University Medical Center will initially target more than 100 patients this year and will consider further extension of the program to benefit more patients. “This pilot program is meaningful in that WGS-based precision medicine has now become a true option for cancer patients. Through this collaboration with Ajou University Medical Center, we will gain real-world experience that will be crucial to improve our offerings to healthcare providers and patients.” said Young Seok Ju, founder of Genome Insight. “We are very pleased to be the first medical center in Korea to introduce WGS-based cancer care throughout the treatment journeys of our cancer patients. I hope this pilot program will serve as an opportunity for Ajou University Medical Center to advance our precision medicine capabilities and bring better care to the patients.” said Hae Shim Park, President of Ajou University Medical Center. About Genome Insight Genome Insight is a cancer whole genome platform company with the goal to accelerate use of whole genome sequencing (WGS) for accurate diagnosis and personalized treatment for cancer and rare diseases. Our proprietary GINS platform is an automated WGS pipeline and computer system that can rapidly generate meaningful interpretation insights for clinicians and patients, making it possible to use WGS in real-life clinical settings. We are also active in research using the power of WGS to advance novel therapeutic strategies in cancer and rare diseases. Our company HQ is in San Diego (US) with R&D offices in Seoul and Daejeon (Korea). To learn more, visit About Ajou University Medical Center Ajou University Medical Center, located in Suwon, South Korea, plays a pivotal role as a tertiary general medical center. It is known for its world-class achievements in the area of patient safety and medical service by receiving the JCI certification. In particular, it has 13 cancer centers, specialized treatment centers and clinics, 1,187 beds with best medical staff. Approximately 1,200,000 outpatients, 50,000 inpatients visit Ajou University Medical Center every year. It is ranked ‘first-class’ in cancer care (particularly in breast cancer, lung cancer, stomach cancer, large intestine cancer) by the Review & Assessment Service of the National Health Insurance Corporation.

Most of today's imaging sensors are based on CMOS technology and inorganic photodetectors that convert light signals into electric signals. Although photodetectors made from organic materials are attracting attention because they can help boost sensitivity, for example, it has proven difficult to fabricate high-performance organic photodetectors. Researchers now describe new green-light absorbing transparent organic photodetectors that are highly sensitive and compatible with CMOS fabrication methods. Incorporating these new photodetectors into organic-silicon hybrid image sensors could be useful for applications such as light-based heart-rate monitoring, fingerprint recognition and devices that detect the presence of nearby objects. Researchers have developed and demonstrated new green-light absorbing transparent organic photodetectors that are highly sensitive and compatible with CMOS fabrication methods. Incorporating these new photodetectors into organic-silicon hybrid image sensors could be useful for applications such as light-based heart-rate monitoring, fingerprint recognition and devices that detect the presence of nearby objects. Whether used in smartphones or scientific cameras, most of today's imaging sensors are based on CMOS technology and inorganic photodetectors that convert light signals into electric signals. Although photodetectors made from organic materials are attracting attention because they can help boost sensitivity, for example, it has proven difficult to fabricate high-performance organic photodetectors. "For organic photodetectors to be incorporated into mass-produced CMOS image sensors requires organic light absorbers that are easy to fabricate on large scales and can accomplish vivid image recognition and produce distinct images in the dark with a high frame rate," said Sungjun Park from Ajou University in the Republic of Korea, who co-led the research team. "We developed transparent green-sensitive organic photodiodes that can meet these requirements." The researchers describe the new organic photodetectors in Optica, Optica Publishing Group's journal for high-impact research. They also created a hybrid RGB imaging sensor by superposing the transparent green-absorbing organic photodetector onto a silicon photodiode with red and blue filters. "The green-selective light-absorbing organic layer used in these image sensors considerably reduced crosstalk between the different colored pixels thanks to the introduction of a mixed organic buffer layer," said research team co-leader Kyung-Bae Park from the Samsung Advanced Institute of Technology (SAIT) in the Republic of Korea. "This new design could allow high-performance organic photodiodes to become the main component for imaging modules and optoelectronic sensors used in a variety of applications." More practical organic photodetectors Most organic materials are not suitable for mass production because they cannot withstand the high temperatures used for post-processing or they become unstable during long-time use at moderate temperatures. To overcome this challenge the researchers focused on modifying the photodetector's buffer layer to improve stability, efficiency and detectivity. Detectivity is a measure of how well a sensor can detect weak signals. "We introduced a bathocuproine (BCP):C60 mixed buffer layer as an electron transporting layer," said Sungjun Park. "This gave the organic photodetectors exceptional characteristics, including higher efficiency and an extremely low dark current, which reduces noise." This photodetector can be placed on a silicon photodiode with red and blue filters to create a hybrid image sensor. The researchers showed that the new photodetectors exhibited a detectivity comparable to those of conventional silicon photodiodes. The detectors operated stably under temperatures above 150 °C for 2 hours and showed long-term operational stability at 85 °C for 30 days. The photodetectors also exhibited good color expression. Next, they plan to tailor the new photodetectors and hybrid image sensor for use in various applications such as mobile and wearable sensors (including CMOS image sensors), proximity sensors and fingerprint-on-display devices.