| K-number | K244023 |
| Device name | MAGENTIQ-COLO (ME-APDS) |
| Applicant | Magentiq Eye, Ltd. |
| Product code | QNP |
| Device class | Class II |
| Decision date | Jan 24, 2025 |
| Decision | Substantially Equivalent |
| Regulation | 876.1520 |
MAGENTIQ-COLO (ME-APDS) is computer-aided detection software that assists endoscopists during colonoscopy by automatically detecting and highlighting regions of video consistent with colorectal polyps in real time. The system processes video from endoscopy cameras using deep machine learning algorithms and displays the results on a separate screen; it is intended as an adjunct to standard colonoscopy for screening and surveillance, not as a diagnostic tool or replacement for tissue sampling.
Both the subject device and predicate use a Computer-Aided Polyp Detection (CADe) engine to process video frames and indicate presence and location of detected lesions in real time during colonoscopy. The subject device is a modification that includes updated labeling identifying additional compatible endoscopy systems and modifications to hardware and software, but the AI algorithm remains identical to the predicate device.
IEC 60601-1-2:2014 [Including AMD 1:2021] and TR 60601-4-2 Edition 1.0 2016-05 for EMC testing; IEC 60601-1 Edition 3.2 2020-08 CONSOLIDATED VERSION for electrical safety testing.
The device has identical intended uses, indications, and technological principles as the predicate; the AI algorithm is unchanged. Non-clinical testing (pixel-level image quality, software validation, EMC, electrical safety) all showed expected results. Standalone performance testing on 212 colonoscopy videos with 702 polyps demonstrated similar or superior detection rates (99.7% recall on histology-verified polyps) with acceptable false positive rates (FPPF threshold 0.0333 achieved). The minor hardware and software modifications with additional compatible endoscopy systems do not alter the intended use or raise new safety/effectiveness issues; clinical validation from the predicate supports the subject device.
View the full FDA submission: accessdata.fda.gov