The U.S. Food and Drug Administration (FDA) has released a draft guidance document on premarket submissions for robotically assisted surgical devices.  

The draft guidance defines robotically assisted surgical devices as teleoperated, software-controlled systems that integrate robotic technologies and subassemblies. These devices are designed to assist surgeons in positioning and controlling surgical instruments to safely and effectively perform certain surgical procedures. They use a leader/follower control system architecture and may include automated functions that enhance user capabilities.  

Components of robotically assisted surgical devices typically include an operator/surgeon console, a bedside/patient subassembly, and/or an operating room staff subassembly.  

The guidance does not address autonomous robots that perform procedures independently of qualified surgeons.  

Recommendations in the draft guidance focus on nonclinical testing, clinical data, and labeling, with the aim of facilitating more consistent and efficient review of premarket submissions for these devices.  

The draft guidance recommends conducting nonclinical testing for design verification and validation before in vivo testing, with a focus on functionality, system control behavior, motion performance, visualization and display, and other factors. In vivo testing may include clinical or nonclinical data to demonstrate that the device can perform its intended surgical functions.  

The agency emphasized that labeling, if applicable, should clarify that authorization applies only to the performance of surgical procedures and that disease-related outcomes were not evaluated as part of the device review.  

AI and machine learning algorithms may also be incorporated into these devices, with examples including image segmentation, instrument tracking, and critical structure identification. The guidance recommends that each feature undergo a corresponding risk assessment. For AI or machine learning features specifically, developers are encouraged to discuss their proposals with the agency and consult other FDA guidance on AI to address performance validation, Predetermined Change Control Plans, data management, and other considerations. 

Public comments on the draft guidance are due by November 24, 2026.  

(For more on FDA guidance documents on AI, also read https://ascoai.org/articles/2026/08/fda-considering-approach-to-regulating-generative-ai-enabled-medical-devices-seeking-public-feedback/.)  

ASCO AI in Oncology is published by Conexiant under a license arrangement with the American Society of Clinical Oncology, Inc. (ASCO^®). The ideas and opinions expressed in ASCO AI in Oncology do not necessarily reflect those of Conexiant or ASCO. For more information, see Policies.