Oncospace

Oncospace is a software tool designed to enhance radiotherapy treatment planning by providing patient-specific dosimetric objectives for organs at risk, aiming to improve plan quality and operational efficiency. It utilizes machine learning for prostate cases and statistical database comparisons of anatomical geometry for other sites to predict achievable dose levels, thereby assisting clinicians in optimizing treatment plans within their existing Treatment Planning Systems without supplanting their ultimate clinical judgment.

Oncospace is a software system designed to assist radiation oncologists and medical dosimetrists in planning radiotherapy treatments for patients with conditions in the prostate, thoracic, pancreas, or head & neck regions. It functions as an adjunct to existing Treatment Planning Systems (TPS), employing machine learning algorithms for prostate cases and statistical or database lookup methods for the other specified anatomical sites. The core methodology involves predicting patient-specific dosimetric goals for Organs at Risk (OARs) by comparing an individual patient's anatomical geometry to a database of previously approved treatment plans. This enables users to select and customize treatment planning protocols, including dose prescriptions, delivery methods, and OAR objectives, and facilitates the automated initiation of plan optimization on a TPS by supplying these individualized parameters. The primary value of Oncospace lies in its capability to individualize OAR dose goals, moving beyond population-based protocol criteria to potentially identify opportunities for enhanced OAR dose reduction or to prevent overly aggressive sparing that might compromise target coverage or other OARs. This approach is intended to shorten the iterative process of treatment plan adjustment. The software provides a user interface for evaluating plans against both protocol-based and its predicted goals, although final clinical judgment in accepting these goals and assessing overall plan quality remains with the healthcare professional. Oncospace does not directly interface with treatment delivery machines. Performance validation, conducted using retrospective clinical data, demonstrated that the system reduced the average number of iterations required to achieve a clinically viable and deliverable radiation treatment plan by nearly two-thirds.

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 static machine learning models for prostate plan predictions compare to the statistical/database lookup methods used for thoracic, pancreas, and head & neck cases in terms of predictive accuracy and clinical utility?; What is the size, diversity, and quality control process for the database of approved treatment plans that Oncospace uses to ensure robust and generalizable OAR dose goal predictions across different patient populations?; Could the system's methodology for predicting OAR dose goals be adapted or extended in future versions to incorporate functional imaging data or other patient-specific biological markers beyond anatomical geometry?; What is the typical learning curve for radiation therapy professionals to effectively integrate Oncospace into their existing workflow, and how does it impact overall treatment planning time beyond just the reduction in optimization iterations?; Are there specific complex anatomical scenarios or patient subgroups where Oncospace demonstrates a markedly more significant advantage in OAR sparing or achieving optimal target coverage compared to traditional planning methods?