ESTRO 2024 - Abstract Book

S3699

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

ESTRO 2024

Material/Methods:

A cohort of 30 patients with advanced H&N cancer and treated with concomitant chemoradiotherapy was considered in this study. Each patient was imaged pre-treatment both with static FMISO PET/Computed Tomography (CT) and FDG PET/CT. The PET images were rigidly co-registered with the planning CT in the RayStation TP System (TPS, research version 10B).

The treatment plans were computed through a workflow assisted via scripting in the RayStation TPS.

The automated pipeline comprised three main stages:

I.

Biologically-based dose prescription

II.

Treatment plan optimization Treatment plan evaluation

III.

Biologically-based dose prescription

Oxygen distributions were extracted from FMISO PET images based on a previously developed sigmoidal conversion function of uptake into partial pressure of oxygen levels (pO 2 ) [1]. Maps of clonogenic cell density per voxel were derived from FDG PET images, assuming a linear conversion. This information was synergistically used to determine a patient-tailored dose prescription using a previously developed dose painting by contours framework [1, 2] A hypoxic target volume (HTV) within the clinical target volume (CTV) was segmented on pO 2 maps by thresholding at 10 mmHg. Four target volumes were defined for planning: planning target volume (PTV), CTV, gross tumor volume (GTV), and HTV.

B. Plan optimisation.

Patient-specific prescription doses, Organs At Risk (OARs), and auxiliary planning structures were used as input for a class solution template. The template considered OAR-specific parameters and priorities based on user defined clinical goals. Objective function weights were automatically updated and treatment optimization iterated until target coverage was achieved.

C. Dosimetric and radiobiological plan evaluation

Various dose metrics were evaluated, including Dmean, D95%, and D2% for target volumes, as well as the conformity index (CI) and homogeneity index (HI). Additionally, dose metrics related to OAR dose constraints were assessed. Dose values were also reported as Equivalent Dose in 2 Gy (EQD2). A radiobiological evaluation of the treatment plans in terms of Target Control Probability (TCP) and Normal Tissue Complication Probability (NTCP) in selected OARs [3] was also performed. TCP calculations were performed by considering the planned dose distribution and the initial underlying radiosensitivity as determined from the tumor oxygenation and clonogenic cell density [2].

Results obtained with the template-based automated planning were compared with manual planning for ten selected patient cases with an HTV>1.2 ml.

Results: