| K-number | K240680 |
| Device name | CDM Insights |
| Applicant | Oxford Brain Diagnostics, Ltd. |
| Product code | QIH |
| Device class | Class II |
| Decision date | Dec 6, 2024 |
| Decision | Substantially Equivalent |
| Regulation | 892.2050 |
CDM Insights is post-processing software that assists trained healthcare practitioners in analyzing MR brain images of adults over 45 years old. It performs automated segmentation and quantification of brain structures and white matter hyperintensities, compares results against normative data from healthy populations, and presents findings as numerical values and visual overlays in DICOM-encapsulated PDF reports.
CDM Insights uses identical technological principles, classification, and regulatory framework as the predicate AI-Rad Companion Brain MR. Key differences include: CDM presents percentiles rather than z-scores for comparison; supports diffusion-weighted imaging (DWI) analysis using single-compartment diffusion models; provides longitudinal comparison across multiple time points; and delivers output as DICOM-encapsulated PDF versus DICOM structured reports. Both perform brain morphometry segmentation, quantification, and deviation mapping with substantially equivalent functionality.
ISO 13485 (quality management), ISO 14971:2019 (risk management), IEC 62304 (software lifecycle), AAMI TIR57 and ANSI AAMI SW96 (cybersecurity), and FDA guidance documents including Design Control Guidance, General Principles of Software Validation, Content of Premarket Submissions for Device Software Functions, and Cybersecurity in Medical Devices.
Substantial equivalence is established through: (1) identical intended use, technological principle, classification, and regulatory framework as the predicate; (2) substantially equivalent core features including brain morphometry segmentation, quantification, and deviation mapping; (3) rigorous performance testing demonstrating segmentation accuracy (Dice scores ≥0.58 for all regions, exceeding predicate criterion), repeatability across 121 subjects, reproducibility across 1500+ subjects and multiple scanner models, and percentile accuracy validated on ~2000 independent test scans; and (4) comprehensive compliance with FDA-recognized consensus standards and guidance documents, with all pre-specified acceptance criteria successfully met.
View the full FDA submission: accessdata.fda.gov