| K-number | K242524 |
| Device name | SyMRI |
| Applicant | Syntheticmr AB (Publ.) |
| Product code | LNH |
| Device class | Class II |
| Decision date | Dec 6, 2024 |
| Decision | Substantially Equivalent |
| Regulation | 892.1000 |
SyMRI is a post-processing software that analyzes MR imaging data to generate parametric maps (R1, R2 relaxation rates, and proton density) and automatically segments and quantifies brain tissues. It can generate multiple contrast-weighted images (T1W, T2W, FLAIR, etc.) from the parametric maps with adjustable contrast after acquisition, intended for head imaging in clinical settings with physician interpretation.
SyMRI uses the same fundamental algorithm as the predicate to calculate parametric maps and employs identical segmentation algorithms for volumetric quantification (white matter, gray matter, CSF, myelin correlated, etc.). Both devices support M2D-MDME and 3D-QALAS acquisition sequences and operate as automated software pipelines on standard hardware. The key enhancement is that synthetic contrast-weighted images from 3D-QALAS data can now be used clinically for diagnosis, whereas the predicate could only use M2D-MDME-derived images clinically.
Not stated in this summary.
The subject device maintains identical core quantification and segmentation algorithms as the predicate, with the same pre-defined performance criteria for accuracy and precision demonstrated through non-clinical verification testing using standard phantoms (NIST/ISMRM Model 130). The clinical enhancement—enabling diagnostic use of 3D synthetic images—was validated by a prospective multi-center, multi-reader study (189 subjects, 5 radiologists) showing that synthetic 3D images are non-inferior to conventional 3D images in sensitivity, specificity, and diagnostic accuracy across diverse brain pathologies, with lower artifact prevalence. This functional expansion does not introduce new safety risks and demonstrates equivalent or superior diagnostic capability.
View the full FDA submission: accessdata.fda.gov