ESTRO 2024 - Abstract Book

S3465

Physics - Dose prediction, optimisation and applications of photon and electron planning

ESTRO 2024

Fredén, E., Tilly, D., & Ahnesjö, A. (2022). PO-1713 DVH based evaluation of dose accumulation in an adaptive MR linac workflow. Radiotherapy and Oncology, 170 (2022), S1512-S1514.

223

Digital Poster

Knowledge ‐ based dose volume histogram prediction for lung tumor volumetric-modulated arc therapy

Johann Brand, Juliane Szkitsak, Stefan Speer

Universitätsklinikum Erlangen, Friedrich-Alexander-Universität Erlangen-Nürnberg, Department of Radiation Oncology, Erlangen, Germany

Purpose/Objective:

Volumetric-modulated arc therapy (VMAT) treatment planning is a compromise between a sufficient coverage of the planning target volume (PTV) and a simultaneous sparing of organs at risk (OARs). Regardless, treatment plan quality depends not only on patient specific factors including the type and stage of cancer, the location, and on size of PTV and OARs, but also on the planner’s level of experience, preferences, and the amount of time the planner invests in the plan. In clinical routine, planners often are content with plans, which fulfil certain criteria, but have the potential of further sparing of the surrounding OARs. Especially in the case of lung tumors, it is difficult to decide whether it is possible or worth spending more time to further spare surrounding organs. Therefore, this work aims towards a knowledge ‐ based structure dependent automated DVH prediction module for lung tumors.

Material/Methods:

The DVH prediction module is based on comparing geometric relationships between the PTV and the surrounding OARs of a new patient plan based on prior patient plans stored in a database created specifically for this purpose. As a basis treatment plan and structure data from 102 lung cancer cases, each treated in 28 fractions and 180 cGy/fx, were collected. To access and simplify the patient specific spatial information between a PTV and each OAR, a two dimensional metric named overlap volume histogram (OVH) [1] was used. Due to rotational symmetry of the OVH and coplanarity of VMAT treatment planning in our clinic, the OVH cannot appropriately describe the spatial correlation of PTV and OAR in cranio-caudal direction. Hence, it is thus complemented by so-called overlap z histogram (OZH) which quantify overlap of PTV and the corresponding OAR in that direction. OVH and OZH were calculated for the heart, spinal cord, esophagus, left lung, right lung, and total lung of all 102 lung cancer cases. To create a new patient specific plan, a set of four achievable DVHs is predicted by identifying plans in the database with similar OVH and OZH. To find plans with simultaneously similar OVH and OZH, for each OAR an individual method was developed to combine the information of the two histograms to achieve the best forecast results. The methods focus either more on OZH, OZH or on both equally. To evaluate the prediction capability, a leave-one-out approach was applied. The clinically achieved DVHs for each plan of the database were compared to the predicted DVHs received by the described method applied on the remaining 101 plans. Differences between the predicted and achieved DVHs were analyzed for each OAR and prediction method separately.