Elpixel Launches Enhanced EIRL Brain Aneurysm Detection Model Reducing False Positives

Elpixel Unveils New Model for EIRL Brain Aneurysm Detection

Elpixel Inc., headquartered in Chiyoda, Tokyo, has recently announced the launch of an updated version of the EIRL Brain Aneurysm detection model. This new model significantly reduces the rate of false positives, addressing a critical concern in the medical imaging community. The approval for this enhancement was granted based on Japan's Pharmaceuticals and Medical Devices Act, taking effect from June 12, 2025.

This innovative product is a software medical device designed specifically for detecting potential brain aneurysms from head MRA (Magnetic Resonance Angiography) images. Initially released in 2019 as the first AI-based medical device to receive approval in the field of radiological imaging in Japan, the EIRL series has since undergone a series of updates to improve its performance.

In June 2021, Elpixel announced a significant enhancement that improved detection accuracy by reducing the average number of false positives detected per case. The current model represents the second major upgrade, incorporating additional learning data and a refined algorithm to maintain sensitivity while ensuring an average of fewer than one false positive per case.

Background on Performance Enhancement

The EIRL series began with the EIRL Brain Aneurysm software, introduced in October 2019 to detect potential brain aneurysms from MRA images. Since then, Elpixel has rolled out nine products across three additional medical fields: head, chest, and colon. These products have seen widespread adoption in over 1,000 medical facilities ranging from university hospitals to clinics, contributing to more than 12 million total analysis cases, demonstrating significant support for diagnostic imaging across Japan.

As this technology becomes increasingly implemented in clinical practice, feedback from medical professionals has highlighted key areas for improvement, particularly concerning the incidence of false positives in the EIRL Brain Aneurysm model. This feedback propelled the recent focus on enhancing the product's precision and reliability, addressing a critical barrier for healthcare providers in diagnosis.

Enhanced Features of the New Model

The latest version of EIRL Brain Aneurysm now detects up to four potential aneurysm points per case, achieving a sensitivity rate of 92.3%. The updated model's performance guarantees that the average number of false positives remains below one, significantly improving upon earlier versions. These enhancements are expected to facilitate more accurate readings by physicians, ultimately benefiting patient care.

Detection Cases

Examples of cases where the new model demonstrates reduced false positive rates compared to the previous version include:

- Detection instances indicating diminished false positives with the new model.
- Specific cases highlighting the successful identification of basilar artery issues.

Note: Images in detection examples may differ from actual product displays.

About Elpixel Inc.

Elpixel is committed to pioneering advancements in healthcare AI, driven by the vision of a healthier future for all. Through strategic collaborations with partners, they are focused on innovative product development addressing challenges in prevention, diagnosis, drug discovery, and surgical assistance. Their leading solutions include the EIRL AI for diagnostic imaging and IMACEL for drug development applications.

For more details, visit Elpixel's corporate website or their official blog.

The Future of Imaging AI Technology

As healthcare modalities advance, the workload for medical imaging continues to grow. Elpixel's AI-powered algorithms aim to optimize the analysis of various medical images—from X-rays to MRIs—enhancing the efficiency and accuracy of diagnoses. By the end of June 2025, Elpixel's EIRL series will have been adopted in over a thousand medical facilities across all 47 prefectures, achieving over 12 million total analysis cases, including trials.