SIS System (Version 5.6.0)

The SIS System (Version 5.6.0) is an advanced medical imaging software that utilizes machine learning and image processing to generate patient-specific 3D anatomical models of deep brain structures like the STN and GPi/GPe from clinical MR images. Its purpose is to enhance visualization and localization for neurosurgical planning, particularly for procedures such as Deep Brain Stimulation lead placement, by leveraging deep learning models trained on ultra-high resolution 7T MRI data.

The SIS System (Version 5.6.0) by Surgical Information Sciences is a software-only medical device designed for viewing, presentation, and documentation of medical imaging, with specific modules for image processing, image fusion, and intraoperative functional planning. Its primary function is to aid medical professionals, such as surgeons, neurologists, and radiologists, in visualizing and locating critical brain structures, specifically the subthalamic nuclei (STN) and globus pallidus externa and interna (GPe and GPi), during neurological procedures. The system is particularly indicated for surgical procedures requiring precise identification of anatomical structures from images, including Deep Brain Stimulation (DBS) lead placement, where its 3D output can be integrated with stereotactic image-guided surgery systems. The system employs machine learning and advanced image processing techniques to enhance standard clinical MR images. It generates patient-specific 3D anatomical models of targeted brain structures using pre-trained deep learning neural network models. These models are trained with ultra-high resolution 7T Magnetic Resonance images serving as ground truth, allowing the software to predict the shape and position of a patient's specific brain structures from their own clinical MR scans. Additionally, the SIS System can identify and locate implanted leads from post-operative CT images and co-register them with the preoperatively modeled brain structures. Version 5.6.0 is an update to version 5.1.0, primarily clarifying its use in DBS lead placement, with the core image processing algorithm remaining unchanged. Performance data, including repeated software verification, validation, visualization accuracy, MRI-CT registration, and lead segmentation tests, confirmed that the updated device functions as specified and performs comparably to its predicate.

AI Regulatory Concerns

Powered by

Jev reads the full FDA letter and returns calibrated concern probabilities per regulatory category. These are AI estimates, not regulatory advice.

• Q1. How does the performance of the SIS System's deep learning models for STN and GPi/GPe segmentation compare to manual segmentation by experienced neuroradiologists in terms of accuracy and time efficiency?; What is the range of variability in segmentation accuracy when the SIS System is applied to MR images acquired from different scanner vendors or with varying acquisition parameters?; Can the SIS System adapt or be retrained for visualizing other deep brain structures beyond STN and GPi/GPe relevant for other neurological interventions?; What specific "minor updates to the backend of the software" were implemented in version 5.6.0, and do they have any potential impact on processing speed or compatibility with newer operating systems/hardware?; How does the system's accuracy in identifying implanted DBS leads on post-operative CT images vary with different lead types, implant artifacts, or image quality?