1031 - Using Virtual Monoenergetic Images and Electron Density (Rho) Images from Photon-Counting CT (PCCT) for Radiation Therapy (RT) Dose Calculation: A Comparison Study

X. Wu¹, P. Wohlfahrt², J. Shah³, Q. Wu¹, Q. J. J. Wu¹, and Y. Sheng¹; ¹Duke University Medical Center, Durham, NC, ²Siemens Healthineers, Forchheim, Germany, ³Siemens Healthineers, Cary, NC

Purpose/Objective(s): Photon-counting CT (PCCT) offers rich anatomical information due to better energy discrimination and higher spatial resolution than conventional energy-integrating CT. PCCT generates virtual monoenergetic images (VMI) as well as electron density (Rho) images which conventional CT simulator is not capable of. Both images are viable options for dose calculation in radiation therapy treatment planning however the benchmark performance of accuracy has not been established. Therefore, this study aims to assess the accuracy of dose calculation using PCCT images from a NAEOTOM Alpha from Siemens Healthineers. Materials/

Methods: A total of 29 patients receiving abdominal contrast enhanced PCCT scans were included in this study. Virtual monoenergetic images (VMI) of 70keV (equivalent to 120 kVp SECT) as well as Rho images were derived from venous phase. A Hounsfield look-up table (HLUT) converting 70keV VMI CT numbers to relative electron density was established following the ESTRO consensus guidelines. The HLUT for the Rho image equals a direct conversion table. Both HLUTs were entered in the Eclipse Treatment Planning System (Varian Medical Systems, Palo Alto, CA). All patients were planned for a 9-field intensity modulated radiation therapy (IMRT) simultaneous integral boost (SIB) regimen (25/33Gy) using an in-house deep-learning AI tool. Each patient’s plan was calculated on both VMI and Rho image with corresponding HLUT and dose differences were compared. Eleven dose-volume histogram (DVH) dose metrics related to PTV25 and PTV33 (mean, D98, D95, D0.1cc), plus C-Loop (mean, D0.2cc, D0.1cc) were included.
Results: The results revealed minimal difference between PCCT-derived 70keV VMI and Rho images, with DVH metric percentage differences predominantly falling under 1% range (25/29). The maximum difference was -1.60% (0.61/38.37 Gy for PTV33 D0.1cc of patient #20). The largest dose variation was observed in patient #13, with average DVH metric difference of -1.32% and standard deviation of 0.12%. Overall isodose line remained similar between two images for each patient.
Conclusion: The result demonstrated good agreement in dose calculation between 70keV VMI and Rho images. This benchmark performance demonstrates that PCCT can be used for dose calculation and makes it a competitive alternative for radiation therapy treatment planning given its superior anatomical information. Table. Distribution of DVH metric difference between Rho and VMI image for all patients, including 100, 75, 50, 25, 0 percentiles.

	PTV25				PTV33				C-Loop
	Mean	D98	D95	D0.1cc	Mean	D98	D95	D0.1cc	Mean	D0.2cc	D0.1CC
100	-0.25%	-0.20%	-0.30%	-0.28%	-0.27%	-0.23%	-0.35%	-0.25%	-0.18%	-0.22%	-0.22%
75	-0.67%	-0.67%	-0.69%	-0.51%	-0.70%	-0.71%	-0.76%	-0.66%	-0.49%	-0.54%	-0.53%
50	-0.85%	-0.89%	-0.84%	-0.75%	-0.84%	-0.93%	-0.93%	-0.78%	-0.57%	-0.71%	-0.71%
25	-1.01%	-1.10%	-1.05%	-0.95%	-1.04%	-1.16%	-1.11%	-1.01%	-0.66%	-0.88%	-0.90%
0	-1.32%	-1.46%	-1.40%	-1.40%	-1.43%	-1.43%	-1.44%	-1.60%	-1.18%	-1.35%	-1.35%