(Sunday, 7/12/2020) [Eastern Time (GMT-4)]
Room: AAPM ePoster Library
Purpose: clinical radiotherapy investigations become more data-driven, relying on large cohorts of data, the efficient analysis of multi-patient samples becomes critical for machine learning applications, automated planning validation, and outcome studies. The purpose of this project was to develop a data extraction tool using an Application Programming Interface (API) for ClearCheck (Radformation) that allows for the efficient batch processing of large numbers of datasets from the Eclipse TPS (Varian).
Methods: is an integrated commercial plan evaluation tool for Eclipse that automatically extracts and evaluates plan data, including checks of dose constraints, prescription, and structures. The ClearCheck API is a C# class that allows the user to call ClearCheck's functionality automatically. We created a stand-alone executable script that uses the ClearCheck API to extract treatment planning data in batch and export them as JavaScript Object Notation (JSON) files. We then used MATLAB to parse the JSON files and analyze the data for two studies: (1) classifying breast patients by treatment technique and (2) comparing DVH data for planned versus adapted treatments of prostate patients. Patients were initially identified using a diagnosis code search
Results: ClearCheck API-driven framework enabled us to mine the treatment planning data for 1800+ patients who received radiotherapy for breast cancer from 2010-2018, and efficiently classify them by treatment technique, laterality, and patient position. This framework also allowed us to easily extract the DVH data for 500+ prostate plans in a comparison of multiple adaptive treatment strategies.
Conclusion: the ClearCheck API, we were able to readily extract and analyze vast amounts of treatment planning data for two studies requiring hundreds to thousands of datasets. Tools such as the ClearCheck API can help researchers overcome the challenges of data curation and streamline the aggregation of datasets large enough to answer many open questions in radiotherapy.
Funding Support, Disclosures, and Conflict of Interest: This work was supported by the Agency for Healthcare Research and Quality (R01HS025440) and an internal grant from the UCSD CPRM. We received technical support from Radformation and Varian Medical Systems. KLM reports income for personal consulting and speaker's honoraria from Varian. KS is the cofounder and CEO of Radformation.
Keywords
Not Applicable / None Entered.
Taxonomy
Not Applicable / None Entered.
Contact Email
