Abstract 5398: Optimization generator app (OpGen) for radiotherapy treatment planning using case-based reasoning
Reza Reiazi, Surendra Prajapati, Abdallah S. R. Mohamed, Clifton David Fuller, Mohammad Salehpour · Cancer Research · 2023
Abstract Case-based reasoning (CBR) means adopting previous experiences (i.e, data from previous patients) to meet new demands (new patients). In this scenario, we developed an application named OpGen to generate optimization objective values for head and neck radiotherapy inverse treatment planning, using the previously treated patient with an acceptable outcome. OpGen will be helpful in radiotherapy treatment planning automation and efficiently reducing the required time to reach the target goals and improving the quality and safety of the treatment.OpGen has been developed in Python and has four main modules: 1- Analyzing radiotherapy treatment planning parameters of the previous patients in the database. We have developed (a) a powerful graphical user interface (GUI) that allows quantitative and qualitative analysis of the previous knowledge, including the type of frequently used objective parameter (Min Dose, Max Dose, Min DVH dose, etc.); (b) The range of values used for each specific objective type, and (c) Histogram of the number of patients vs. values used for a specific objective type. 2- Importing patients from treatment planning and retrieving a user-defined number of treatment planning information from the previously treated patients in the database corresponding to the given patient using the user-defined quantitative features (geometrical, distance or geo-distance) and retrieval algorithms (k-Nearest Neighbor and Random Forest). 3- Generating optimization objective values based on the following available options: (a) The most similar patient in the database; (b)- Average or median value over the number of patients retrieved from the database; and (c) Deep learning generated values using a model trained over 100 similar cases retrieved from the database. 4- Generate a portable database of the objective values and patient geometric information for any inter- or intra-institutional usage of the algorithm. OpGen is compatible with RayStation (RaySearch Laboratories™) treatment planning and can retrieve, analyze and generate treatment planning optimization parameters in less than 3 minutes. OpGen can potentially increase the quality and safety of radiotherapy treatment planning by improving the efficiency of the optimization process. Citation Format: Reza Reiazi, Surendra Prajapati, Abdallah Sherif Mohamed, Clifton David Fuller, Mohammad Salehpour. Optimization generator app (OpGen) for radiotherapy treatment planning using case-based reasoning. [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 5398.