S2426
Physics - Radiomics, functional and biological imaging, and outcome prediction
ESTRO 2026
10.1016/j.radonc.2022.109456. Keywords: radiation necrosis, segmentation, brain metastasis
Poster Discussion 1186
Piecewise dose-response relationship between mean heart dose and risk of cardiac troponin elevation Mei Chen 1,2 , Xiaodong Zhang 3 , Ting Xu 4 , Zhongxing Liao 1 1 Thoracic Radiation Oncology, University of Texas MD Anderson Cancer Center, Houston, USA. 2 Shanghai Institute of Applied Physics, University of Chinese Academy of Sciences, Shanghai, China. 3 Radiation Physics, University of Texas MD Anderson Cancer Center, Houston, USA. 4 Radiation Oncology, University of Texas MD Anderson Cancer Center, Houston, USA Purpose/Objective: Rising concerns over radiation-induced cardiac toxicity have driven widespread adoption of cardiac-sparing techniques in non-small-cell lung cancer (NSCLC) radiotherapy, yet whether dose reduction translates into fewer subclinical cardiac injuries remains unclear. This study aims to investigate the dose-response relationship between mean heart dose (MHD) and the risk of cardiac troponin elevation before and after implementing cardiac-sparing strategies. Material/Methods: We retrospectively reviewed data from 160 NSCLC patients treated pre-practice change and 91 NSCLC patients treated post-practice change. Both cohorts were enrolled in prospective clinical trials.The study endpoint was during-treatment elevation of high- sensitivity cardiac troponin (hs-cTnT) by ≥ 5 ng/L from baseline. Both conventional (cLR) and piecewise logistic regression (pLR) models were developed to evaluate the dose-response relationship between MHD and hs-cTnT elevation. Univariable analysis was performed, followed by multivariable analysis adjusted for clinical covariates. Covariate selection employed 1000 bootstrap resampling iterations with stepwise logistic regression to enhance model robustness and prevent overfitting.The MHD cut-point in the pLR model was determined via grid search to minimize the Akaike Information Criterion (AIC). Overall performance of the model was evaluated using AIC, and discriminative ability was quantified using the area under the receiver operating curve (AUC). Likelihood ratio test (LRT) was performed to compare the goodness-of-fit of two types of models. Results: Despite 37.2% MHD reduction (10.86 ± 8.74 Gy vs 6.82 ± 6.05 Gy, p< 0.001), the hs-cTnT elevation rates were unchanged (26.9% vs 26.4% p = 1.000). MHD was
Conclusion: The risk of RN is dose-, and volume-dependent; therefore, RN risk analyses should not be limited to only V12Gy but expanded across the full dose spectrum. While higher doses carry higher RN risk, low dose dose-volumes, such as V5Gy, may confer negligible risks. Therefore, the dosimetry of coplanar MR - Linac [6] SRT for BM may be acceptable, because it delivers identical target coverage with a modest low - dose bath compared to non-coplanar VMAT linac- based SRT. References: [1] B. Vellayappan, Front. Oncol., vol. 8, Sep. 2018, doi: 10.3389/fonc.2018.00395.[2] G. Minniti, Radiat. Oncol., vol. 6, no. 1, p. 48, May 2011, doi: 10.1186/1748-717X-6-48.[3] M. T. Milano, Int. J. Radiat. Oncol. Biol. Phys., vol. 110, no. 1, pp. 68–86, May 2021, doi: 10.1016/j.ijrobp.2020.08.013.[4] B. J. Blonigen, Int. J. Radiat. Oncol. Biol. Phys., vol. 77, no. 4, pp. 996–1001, Jul. 2010, doi: 10.1016/j.ijrobp.2009.06.006.[5] C. Gaser, GigaScience, vol. 13, p. giae049, Jan. 2024, doi: 10.1093/gigascience/giae049.[6] E. Seravalli, Radiother. Oncol., vol. 179, p. 109456, Feb. 2023, doi:
Made with FlippingBook - Share PDF online