CMDE of NMPA Releases Technical Review Points for Clinical Trials of Implanted Brain-Computer Interface Hand Motor Function Compensation Systems

CMDE of NMPA released the Technical Review Points for Clinical Trials of Implanted Brain-Computer Interface Hand Motor Function Compensation Systems on October 22, 2026 (Notice No. 29 of 2026), aiming to further standardize technical review and unify review standards for this product category. The document is intended to guide applicants in preparing registration submissions. It is not mandatory as a regulation and should be used in compliance with applicable laws, regulations, normative documents, and mandatory medical device standards, while taking into account the specific characteristics of the product.

The document states that the product is generally intended for patients with quadriplegia caused by cervical spinal cord injury, using a pneumatic glove to assist compensatory hand grasping function. Eligible patients are generally 18 to 60 years old, with C2–C6 cervical spinal cord injury graded ASIA A–C, mainly complete A-level injury with an A-level proportion of ≥50%, disease confirmed for more than 1 year and stable for at least 6 months after standardized treatment, unable to perform grasping, and with some remaining upper-arm function. Under the current medical device classification catalogue, the classification code is 12-00 and the management category is Class III. The review points provide general requirements for products with epidural implantation, local field potential signal decoding, a relatively low channel count (e.g., fewer than 64 channels), and implantation in the hand knob area. They also clarify non-applicable cases, including non-implantable non-invasive products, deeper implantable products such as subdural, intracortical, deep brain, or intravascular products, products implanted outside the hand knob area, products that can control other external devices or achieve other clinical functions, and products for other patient populations.

For clinical trials, the document emphasizes that these products are high-risk clinical products and that clinical evaluation should be conducted through clinical trials. Domestic trials should comply with GCP requirements. It is generally recommended to conduct a feasibility trial before a confirmatory trial, and the number of feasibility trials for the same indication and same clinical function should generally be no fewer than 4 cases. A single-arm objective performance criteria design may be considered. The primary endpoint may be the BCI-assisted ARAT grasping response rate at 6 months after implantation, and the MCID should generally not be lower than 2 points. In the sample size example, with an objective performance criterion of 50%, expected response rate of 75%, one-sided α=0.025, power of 0.8, and a 10% dropout rate, the total sample size is recommended to be no fewer than 32 cases. Follow-up is recommended at 6 months after implantation or longer. Post-market follow-up studies are also required, with a recommended follow-up period of no less than 2 years after implantation, and the post-market study report should be submitted at the time of renewal registration.

For instructions for use, in addition to regulatory information, the document requires a clinical trial summary, intended use, postoperative training settings, requirements for determining the implanted brain area and intraoperative implantation location, motor imagery training paradigm for EEG decoding, adverse events and complications, algorithm tuning/calibration and retraining requirements, magnetic resonance conditional safety performance, and specific use environments.