Related Experiment Video
Updated: Jan 14, 2026
Proton Therapy Delivery and Its Clinical Application in Select Solid Tumor Malignancies
Published on: February 6, 2019
Radiobiological Modeling for Acute Proctitis After Conventional and Hypofractionated Radiation Therapy for Prostate
Christian Am Jongen1, Ben Jm Heijmen1, Luca Incrocci1
1Department of Radiotherapy, Erasmus MC Cancer Institute, University Medical Center Rotterdam, Rotterdam, The Netherlands.
Purpose:
Acute proctitis (AP) is a common side effect of pelvic radiation therapy. Randomized studies showed increased risks after moderate hypofractionated radiation therapy with shorter overall treatment times (TTs). We developed normal tissue complication probability (NTCP) models for AP for a fixed TT of 4 weeks (W4), based on combined dose-fractionation data.
Methods And Materials:
Study patients treated with 20 fractions of 3.0/3.1 Gy, 39 fractions of 2 Gy (both 5 fractions/wk) or 19 fractions of 3.4 Gy (3 fractions/wk) to prostate ± vesicles were selected from 2 prospective studies (n = 321). We focused on observed toxicity in W4 (not the total acute period). Patient-reported symptoms and delivered dose data at W4 were collected (20 × 3.0/3.1 Gy, 20 × 2 Gy, 12 × 3.4 Gy). AP was scored based on 7 patient-reported symptoms (e.g. mucus loss). We fitted 3 logistic regression models based on the equivalent uniform dose (EUD) in biologically effective dose (BED) with diabetes and androgen deprivation therapy (ADT) as candidate clinical predictors. The EUD volume parameter, n, was estimated with maximum likelihood estimation (MLE). In model_10Gy, we assumed α/β = 10 Gy, whereas in model_α/β, α/β was estimated with MLE together with n. In model_TT, the impact of TT on NTCP was explored using Fowler's BED formula with TT correction. Backward elimination was applied for parameter selection and bootstrapping for internal validation.
Results:
AP incidence was 66%, 25%, and 24% after 20 × 3.0/3.1 Gy, 20 × 2 Gy, and 12 × 3.4 Gy, respectively. Model_10Gy contained EUD with n = 0.30 (95% confidence interval [CI], 0.15 to 0.45), and ADT (OR, 0.61) as predictors, with an area under the curve of 0.77. Model_α/β showed similar performance with an estimated α/β of 17 Gy (95% CI, 6 to >20 Gy). Model_TT estimated NTCP reductions of 4% to 20% when TT is extended by 1 week.
Conclusions:
We developed well-performing NTCP models for AP as a function of EUD and ADT and showed a good fit for the commonly assumed α/β of 10 Gy, and for both conventional and moderate hypofractionated schemes.
