What happened
Infervision Medical Technology built IR-CT-PE, software that uses deep learning to segment the airways, pulmonary vessels, lung lobes and emboli on pulmonary CT angiography. It flags adults with suspected acute pulmonary embolism for triage. The Center for Medical Device Evaluation (CMDE) of China's National Medical Products Administration reviewed the company's application for Class III registration. The trial was the maker's own, a multicenter, single-arm study of 485 cases at two institutions. An expert panel's case judgments served as the reference. Sensitivity was 92.27% (95% CI 87.57% to 95.61%) and specificity 96.91% (95% CI 94.21% to 98.58%), both above targets of 80%. The software's median time to a result was 167.0 seconds. The hospitals' routine imaging reports took a median of 55431.0 seconds, about 15.4 hours. No software fault or adverse event occurred. On 19 February 2025, CMDE concluded that “该产品上市带来的受益大于风险” (the benefits of putting the product on the market outweigh the risks) and recommended approval. The report does not say whether the trial was prospective or whether doctors saw the prompts.
Sources
-
Regulatory document, by authors with no tie to the system.
软件的分诊灵敏度 92.27%(95%CI,87.57%-95.61%),区间下限大于目标值 80%
经综合评价,认为该产品上市带来的受益大于风险,综合剩余风险可接受。
Details
AI system
- Reported configuration
-
Release 1 (full version 1.0.3.0)
- Workflow
-
The software segmented the airways, pulmonary vessels, lung lobes and emboli with deep learning and gave a triage prompt for suspected acute pulmonary embolism. An expert panel's case judgments served as the reference.
Follow-up
- Conditions for replication
-
Patients are adults with suspected acute pulmonary embolism. Trained physicians use the prompt and do not base a clinical decision on it alone.
Entry details
- Sector
-
Healthcare.
- Purpose
-
Infervision Medical Technology built IR-CT-PE to flag adults with suspected acute pulmonary embolism on pulmonary CT angiography so that trained physicians can triage them.
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). IR-CT-PE flagged pulmonary embolism in a two-site trial in China (AIBD-1086). Artificial Intelligence Benefits Database, edition of 1 November 2026. https://petervartanian.xyz/aibd/records/AIBD-1086.html (archived at https://doi.org/10.7910/DVN/AADSII, version 7.0)