OncoStudio (OS-01)

The OncoStudio (OS-01) is a medical image processing system designed to enhance radiation oncology treatment planning by providing deep-learning-based automatic contouring of organs at risk from CT images. Its core science relies on advanced deep learning algorithms trained on diverse datasets, enabling efficient and accurate initial segmentation of anatomical regions for clinical use and supported by rigorous software validation.

OncoSoft. Co., Ltd.'s OncoStudio (OS-01), a Class II medical image management and processing system, has been deemed substantially equivalent to legally marketed predicate devices, allowing its commercialization. This regulatory clearance mandates adherence to general controls, including annual registration, device listing, good manufacturing practices, and truthful labeling, alongside specific requirements for adverse event reporting and compliance with the Unique Device Identification System. The OncoStudio system functions as standalone software that employs deep-learning for automatic contouring of organs at risk from DICOM-RT CT images, primarily for initial contouring in radiation oncology treatment planning across anatomical regions like the Head & Neck, Thorax, Abdomen, and Pelvis. It also supports manual contouring and general image management functions such as receiving, transmitting, storing, retrieving, displaying, and processing medical images. The device's safety and effectiveness were substantiated through comprehensive non-clinical software validation and performance evaluations utilizing diverse global datasets.

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. What specific deep-learning algorithms or architectures are employed by OncoStudio for automatic contouring, and how do they compare to conventional methods in terms of accuracy and efficiency?; Beyond initial contouring, how does the OncoStudio system integrate with existing radiation therapy planning systems, and what are its capabilities for real-time adjustments or adaptive planning?; What was the size, diversity, and ethnic representation of the global datasets used for non-clinical testing, and how did the system perform across different pathologies or image qualities?; Are there any documented instances or predictions of potential biases in the deep-learning model's contouring output for specific patient demographics or complex anatomical variations?; What post-market surveillance or clinical validation studies are planned or underway to further assess the long-term clinical utility, inter-observer variability reduction, and patient outcomes associated with OncoStudio's use?