MCS flagged 11 eye diseases on fundus photos in China's review

In the maker's trial that China's CMDE reviewed, the MCS fundus software met its targets for 11 eye diseases, with 91.18% sensitivity and 91.56% specificity for diabetic retinopathy.

  • Not rated an organization

    The accuracy was measured in the maker's pre-market trial, and the report does not say that the software's prompts guided any patient's care.

Depth runs from 1, a convenience or time saved, to 5, a life or a livelihood preserved.

Status
Intended benefit.
Evidence

foundstudiedreportedclaimed

An independent regulatory document supports the benefit.

Compared with

Targets set for each disease from a pilot trial and the literature, with central reading as the gold standard. Measured.

Without the system
Target of 82.0% sensitivity for diabetic retinopathy
With it
91.18% [86.48%,94.35%] sensitivity for diabetic retinopathy
The system’s part
Not isolated. The trial measured the software's own readings against central reading. It did not compare doctors' diagnoses with and without the software. The trial was the maker's own. The report says that the software uses a deep convolutional neural network to score image quality and classify diseases.
Not yet checked
LimitsCosts and harmsLearns from dataLater outcome

Download data Cite Suggest a correction

Benefit type
Improved Task Accuracy
System
MCS / MCSLocal (眼底病变眼底图像辅助诊断软件)
Made by
Beijing Zhiyuan Huitu Technology Co., Ltd. (北京致远慧图科技有限公司)
Used by
Beijing Zhiyuan Huitu Technology (registration trial with 2,778 fundus photographs at four institutions)
Where
Four trial institutions that the report does not name, for a registration in China
Stage
Tested in a laboratory
Date
, when the source was published
Kind
Aggregate result
Broad benefit
Identifying Objects, Actions, and Events
Collections
Scrutiny and findings
Last revised
1 November 2026

Sources

  1. Regulatory document, by authors with no tie to the system.

    Center for Medical Device Evaluation (CMDE), National Medical Products Administration, China. 医疗器械产品注册技术审评报告：眼底病变眼底图像辅助诊断软件 (technical review report for the registration of fundus disease fundus image diagnosis-aid software), acceptance no. CQZ2200072, applicant 北京致远慧图科技有限公司. 11 June 2024. Section 三、临床评价概述 (clinical evaluation).

    特异度 95%置信区间下限均高于目标值， 点估计值均高于预期值
    认为该产品的上市为适用人群带来的受益大于风险。

Details

AI system

Reported configuration

Release V2 (MCS cloud server version and MCSLocal local server version)

Workflow

The software first scored the quality of each photograph and analyzed only those that passed. A single deep-learning framework estimated the probability of each disease and gave up to three prompts.

Follow-up

Conditions for replication

Photographs are single color images of the posterior pole of an adult's eye that show the whole optic disc and macula. Photographs come from the specified fundus camera models. Doctors confirm or rule out disease with other examinations and do not decide on the prompts alone.

Entry details

Sector

Healthcare.

Purpose

Beijing Zhiyuan Huitu Technology built MCS to tell doctors qualified in fundus examination whether a patient's fundus photograph has a suspected abnormality and which of 11 eye diseases the patient may have.

30 more questions not yet checked

model, provider, role, configuration, inputs and outputs, autonomy, human oversight, AI contribution, beneficiaries, reach, distribution and access, duration, every measurement question, every money question, ownership and control, dependencies, cross-border connections, geopolitical implications, replication attempts, availability, institutional decision, grounds, action and result, success criterion, every evidence question, observed or intended, alternative assessments, related entries, other identifiers, category mappings.

Cite this record

Peter H. Vartanian. (2026). MCS flagged 11 eye diseases on fundus photos in China's review (AIBD-1094). Artificial Intelligence Benefits Database, edition of 1 November 2026. https://petervartanian.xyz/aibd/records/AIBD-1094.html (archived at https://doi.org/10.7910/DVN/AADSII, version 7.0)

BibTeX Data Field definitions