Exploring Clinical Trade-Offs for Brain Tumor Cases with a Probabilistic Clinical Target Volume
G Buti1,2*, N Shusharina2, A Ajdari2, E Sterpin1,3, T Bortfeld2, (1) UCLouvain, Institute of Experimental and Clinical Research, Center of Molecular Imaging, Radiotherapy and Oncology (MIRO), 1200 Brussels, BE, (2) Massachusetts General Hospital and Harvard Medical School, Department of Radiation Oncology, Division of Radiation Biophysics, 02114 Boston, USA (3) KU Leuven, Department of Oncology, Laboratory of Experimental Radiotherapy, 3000 Leuven, BE.
Presentations
TU-A-202-4 (Tuesday, 7/12/2022) 7:30 AM - 8:30 AM [Eastern Time (GMT-4)]
Room 202
Purpose: For decades, ICRU reports have stated that the ultimate treatment planning method should take probabilities of microscopic tumor infiltration into account. This study addresses this gap by demonstrating how a novel probabilistic CTV concept –clinical target distribution (CTD)– can be used to navigate clinical trade-offs. In particular, the impact of the underlying tumor infiltration assumption on the tumor control probability (TCP) of patients with dose-constrained critical organs, is investigated.
Methods: The therapeutic effect of the dose is modeled by a step function to approximate the TCP function. Here, the dose has either a curative or non-curative effect. This gives rise to a ‘all-or-nothing’ dosing strategy where a subvolume of the CTD should either receive the prescription, or no dose at all. We consider two models of tumor infiltration: independent tumor islets, and circumferential tumor fronts. For each tumor infiltration model, a library of treatment plans is generated where the tradeoffs between CTD coverage and OAR sparing can be explored interactively. The methods are tested for a glioblastoma patient (60Gy prescription) with the focus to spare the contra-lateral hippocampus and brainstem.
Results: In the independent tumor islet model, the maximum TCP (95.8%) is reached by dosing 98% of the CTD. To spare the hippocampus, up to 30% of CTD can be underdosed without significant loss of TCP (TCP=92.3%) or target coverage (D98=60.3Gy). In contrast, in the circumferential growth model, the correlation of the voxels indicates that a uniform expansion of the GTV should be dosed uniformly. No gain in TCP was achieved by extending the high-dose region beyond a 6mm expansion (TCP=17.5%). The 6mm margin represents the largest expansion that can be dosed without compromising dose-homogeneity.
Conclusion: This study designs probabilistic planning approaches for clinical cases considering tumor propagation, which is normally a hidden assumption in the target definition process.
Funding Support, Disclosures, and Conflict of Interest: Research was supported by the Belgian F.R.S.-FNRS and the National Cancer Institute of the USA under grant number R01CA266275, and by the Therapy Imaging Program funded by the Federal Share of program income earned by Massachusetts General Hospital on C06CA059267, Proton Therapy Research and Treatment Center.
Keywords
Not Applicable / None Entered.
Taxonomy
Not Applicable / None Entered.
Contact Email