Ensure safety and accuracy. AI / deep learning autosegmentation technologies are emerging and promising. Widespread implementation is imminent. Verifying the accuracy of autosegmentation in an objective and rigorous manner is paramount.
Make data-driven purchasing decisions. Use StructSure to generate accuracy data to see how well different algorithms reproduce the standard of contouring quality that you expect. Request from the competing vendors to have a trial license of their autosegmentation solutions, select your own datasets and standards against which to compare, then use StructSure as the objective "referee" to help you determine which products give you the expected output.
Implement an efficient and reproducible validation process. Just like you must validate and commission dose calculation algorithms, you will also need to validate and commission your autosegmentation model(s). You will need to do this not only initially, but also as the models evolve and new models are added (i.e., for different body sites and imaging modalities). The StructSure process is efficient, comprehensive, and can be automated and reproduced via batch analyses.
Communicate better with your autosegmentation vendor(s). For any body sites, models, or anatomical structures that do not pass your validation, or that need improvement, StructSure produces a wealth of data about the errors and differences vs. expected results. These objective measures can be shared with your autosegmentation vendor(s) to ensure clear, but comprehensive, communication and to avoid debates based on subjective observations.
Low cost. StructSure is low cost and can even be paid for on an "as needed" basis via SaaS pricing.