Comparison of two established approaches for generating Hounsfield look‐up tables for CT‐based SPR prediction
Medical Physics · 2026
DOI: 10.1002/mp.70236 ↗At a glance
Vergleichsstudie in Medical Physics, 2026. Written by Minoru Nakao and 4 co-authors. Institution: Hiroshima High-Precision Radiotherapy Cancer Center, Hiroshima, Japan. Major MEDLINE topics: Tomography, X-Ray Computed, Radiotherapy Planning, Computer-Assisted, Image Processing, Computer-Assisted. 29 references. Funded by: Japan Society for the Promotion of Science.
Abstract (original language)
Background
To ensure consistency in treatment planning across institutions, inter‐center variations in Hounsfield look‐up tables (HLUTs) have been evaluated using different methods in Japan and Europe. In Japan, the Hiroshima High‐Precision Radiotherapy Cancer Center (HIPRAC) developed the computed tomography number calibration audit (HIPRAC‐CTCA) based on a stoichiometric method, while in Europe, the European Particle Therapy Network (EPTN) implemented a method using validated tissue‐equivalent inserts. Comparing these approaches is essential to improving consistency and reliability.
Purpose
This study aims to enhance the reliability of the HIPRAC method by comparing it with the EPTN consensus guide‐based HLUT generation methods used in multi‐institutional evaluations.
Methods
HLUTs were generated using the HIPRAC and EPTN methods for both head and body phantom configurations. The HIPRAC‐based HLUT was created by measuring CT numbers of tough lung and tough bone inserts, applying the stoichiometric method, and calculating the theoretical HLUTs for representative tissues based on the International Commission on Radiological Protection Publication 110. The EPTN‐based HLUT was generated by measuring the CT numbers of validated tissue‐equivalent inserts, calculating the CT numbers of human tissues using the stoichiometric method, and determining the HLUT through piecewise linear regression between CT numbers and stopping‐power ratios (SPRs) for both tissue‐equivalent inserts and human tissues. Δ SPR values were calculated by subtracting the EPTN‐based HLUT from the HIPRAC‐based HLUT and were categorized into lung, soft tissue, and bone groups. Representative brain tumor and prostate cancer cases were analyzed to evaluate spot‐wise range shifts (Δ R 80 ) when applying the two HLUTs.
Results
Both methods showed good agreement in lung, soft tissue, and bone, with all Δ SPR values within ± 2% and mean differences ≤ 1%. In cortical bone, Δ SPR reached approximately 2%–3%, leading to minor range shifts. In the clinical case analysis, mean Δ R 80 were −0.28 mm for the brain tumor case and −0.45 mm for the prostate cancer case, indicating that proton range differences between the two HLUTs were negligibly small.
Conclusions
The established HLUT generation methods for CT‐based SPR prediction developed in Japan and Europe were compared and showed good agreement across all tissue types. Range differences in clinical cases were minor, and Δ SPR variations in cortical bone corresponded to negligible proton range deviations. These findings indicate that the European and Japanese approaches are practically equivalent and support the validity and reliability of the HIPRAC‐CTCA method.
Subjects
MeSH terms (MEDLINE)
Author keywords
Authors and institutions
- Hiroshima High-Precision Radiotherapy Cancer Center, Hiroshima, Japan.
- Department of Radiation Oncology, Graduate School of Biomedical & Health Sciences, Hiroshima University, Hiroshima, Japan.
- OncoRay-National Center for Radiation Research in Oncology, Faculty of Medicine and University Hospital Carl Gustav Carus, TUD Dresden University of Technology, Helmholtz-Zentrum Dresden-Rossendorf, Dresden, Germany.
- Helmholtz-Zentrum Dresden-Rossendorf, Institute of Radiooncology - OncoRay, Dresden, Germany.
- Department of Radiotherapy and Radiation Oncology, Faculty of Medicine and University Hospital Carl Gustav Carus, TUD Dresden University of Technology, Dresden, Germany.
- Department of Radiation Oncology, Mass General Brigham and Harvard Medical School, Boston, Massachusetts, USA.
Funding
- Japan Society for the Promotion of Science
Cite
APA
Nakao, M., Schwengfelder, J., Richter, C., Peters, N., & Ozawa, S. (2026). Comparison of two established approaches for generating Hounsfield look‐up tables for CT‐based SPR prediction. Medical Physics. https://doi.org/10.1002/mp.70236
Vancouver
Nakao M, Schwengfelder J, Richter C, Peters N, Ozawa S. Comparison of two established approaches for generating Hounsfield look‐up tables for CT‐based SPR prediction. Medical Physics. 2026. doi:10.1002/mp.70236
BibTeX
@article{Nakao2026,
author = {Nakao, Minoru and Schwengfelder, Julian and Richter, Christian and Peters, Nils and Ozawa, Shuichi},
title = {Comparison of two established approaches for generating Hounsfield look‐up tables for CT‐based SPR prediction},
journal = {Medical Physics},
year = {2026},
doi = {10.1002/mp.70236}
}
EUVIMED