Click here to

Session: Multi-Disciplinary BLUE RIBBON [Return to Session]

Clinical Implementation of DECT for Proton Therapy with Reduced Range Uncertainty Margins: Reproducibility and Predicted Range Differences

S O'Reilly1*, L Yin2, A Kassaee1, Y Chang1, C Cheng3, J Zou1, T Li1, A Lin1, J Lukens1, R Lustig1, G Kurtz1, M Alonso-Basanta1, L Dong1, BK Teo1, (1) University of Pennsylvania, Philadelphia, PA, (2) Johns Hopkins Proton Therapy Center, Washington, DC, (3) Rutgers Cancer Institute of New Jersey, New Brunswick, NJ


SU-I400-BReP-F2-2 (Sunday, 7/10/2022) 4:00 PM - 5:00 PM [Eastern Time (GMT-4)]

Exhibit Hall | Forum 2

Purpose: To describe the first clinical implementation of dual energy computed tomography (DECT) for calculation of proton stopping power ratios (SPRs) in a commercial treatment planning system (TPS) for plan optimization and treatment with reduced range uncertainty (RU) margins. Single energy CT (SECT) based plans typically use a 3.5% RU during optimization; however, studies have shown that DECT has less RU and could allow for reduced margin.

Methods: Between September 2020 and March 2022, 180 head and neck (H&N) and 272 central nervous system (CNS) proton therapy patients were scanned with DECT. Modified CT images were created using SPR-derived from electron density and effective atomic number images and converted into a CT with a HU-SPR calibration curve. The modified CT was imported into the TPS to create clinically used robust plans optimized using 3mm setup and 2% RU margins. An equivalent 120kVp SECT was created by mixing the DECT images and used in forward calculation as a pre-treatment safety check. SPR differences between SECT and DECT images were calculated for adipose, muscle, brain and dense bone for 90 patients. A subset of patients (n=20) who had at least two DECT-based SPR images generated during their treatment course were analyzed to confirm SPR reproducibility.

Results: DECT-predicted SPR was 2.98% lower for adipose, 0.02% lower for brain, 0.68% higher for muscle, and 1.37% lower for dense bone compared to SECT-predicted. SPR-generated CT scans were found to be reproducible with average SPR differences of <0.1% for adipose, brain and muscle and <0.2% for dense bone between scans. Small dosimetric improvements were seen in brainstem, optic structures, mandible, submandibular and oral cavity doses owing to more accurate SPR and smaller RU used.

Conclusion: DECT-based SPR-CT treatment planning with a reduced RU has been successfully implemented using a commercial TPS for CNS and H&N patients.

Funding Support, Disclosures, and Conflict of Interest: M Alonso-Basanta: Travel fees for presentations paid by IBA. A Lin: 1) Galera Therapeutics - Advisory board and consultant, 2) IBA - Advisory board.


Dual-energy Imaging, Protons


TH- External Beam- Particle/high LET therapy: Dual energy/spectral CT-based stopping power mapping

Contact Email