RadiLogic: Bringing AI into Radiology

Facing the Covid-19 threat in early 2020, Institute for Infocomm Research (I²R) together with Tan Tock Seng Hospital (TTSH) and Institute of High Performance Computing (IHPC), developed RadiLogic to detect Covid-19 pneumonia from chest X-ray (CXR) images. It prioritises screening and isolation of suspected subjects with pneumonia symptoms. Following RadiLogic's development, I²R further improved technologies to enhance pneumonia detection and identify positive COVID-19 cases through CXR images.

RadiLogic enables high throughput screening for COVID-19 pneumonia, triaging patients based on CXR findings before RT-PCR diagnosis, optimising hospital resources.

Pneumonia, a common lung infection symptom, poses diagnostic challenges. RadiLogic's AI capabilities detect COVID-19 pneumonia in under 3 seconds with a GPU workstation, achieving an accuracy area under the curve (AUC) of over 94% in real-world deployments. This technology significantly reduces turn-around times and improves detection accuracy, with an AUC of 96.5% in an enhanced version.

RadiLogic's potential extends to disease detection in CXRs for clinical and research purposes, facilitating disease progression monitoring and evaluating drug effectiveness in treating lung infections.

The RadiLogic team's efforts have been recognised with prestigious awards such as the SCS IT Leader Award 2021, National Healthcare Innovation and Productivity (NHIP) Medals 2021, National Award Covid-19 (Commendation Medal) 2022, and Covid Resilience Medals 2023.