ESTRO 2020 Abstract book

S160 ESTRO 2020

dose to oral mucosa. Model performance was assessed by area under the receiver operating characteristics (AUC- ROC) curve. The EORTC HNCG-ROG 1219 DAHANCA validation cohort consisted of 192 randomized patients. After removing missing data from validation cohort, toxicity data and radiotherapy plans were available for 169 patients. Acute toxicity was defined based on CTCAE v.3.0. Results Sixty/169 patients developed grade 3 oral mucositis in the validation cohort. The mean predicted probability of toxicity was 32%, while the observed probability was 35.5%. The original logistic regression function was in the form of:

PH-0289 Temporary alopecia in photon cranial irradiation: Establishing the dose-response relation using MRI A. Van Lier 1 , C.G.M. Gadellaa-van Hooijdonk 1 , T. Coolen 1 , T. Nguyen 1 , E. Brand 1 , J.J.C. Verhoeff 1 , E. Seravalli 1 1 UMC Utrecht, Department of Radiation Oncology, Utrecht, The Netherlands Purpose or Objective To establish the dose-response relation for skin dose and temporary alopecia in patients treated with stereotactic or conventionally fractionated photon radiotherapy to intracranial lesions using MRI imaging, for implementation as soft constraint in the radiotherapy planning system. Material and Methods 46 patients receiving a follow-up MRI examination after cranial irradiation by photons (VMAT technique) (~ 90 days after last fraction) were screened for alopecia, of which 10 showed local temporary alopecia. For the reported analysis 5 patients with alopecia (3 stereotactic regime, 2 conventional fractionation) were found eligible for analysis, based on the following criteria: no prior cranial irradiation, stereotactic or conventional fractionation. Alopecic areas were marked with a gadolinium-filled line- marker after which the follow-up MRI scan was obtained according to local protocol. A T1-weighted 3D scan (1mm 3 resolution) was used to automatically delineate the scalp skin with a thickness of 3 mm, which is known to bear the scalp hair follicles. Skin delineations, which were divided in an alopecic volume and an unaffected hair-bearing volume using the line-marker, were rigidly registered to the planning CT scan and corresponding radiotherapy dose distribution (1mm 3 dose grid). Subsequently, the dose-per- voxel was obtained for the two volumes. Afterwards, a logistic probability curve (dose vs. effect) was fitted to the dose-per-voxel values to estimate the normal-tissue complication probability curve. Results 3D visualization showed that the affected skin delineation corresponded consequently to areas with a higher skin dose (see example in Figure 1). After combining voxel data of all patients, it was shown that the median dose in the alopecic volume was higher than for the unaffected regions (0.2 vs. 1.5 Gy for a single fraction delivered in a stereotactic regime and 5.7 vs. 16.9 Gy for conventionally fractionated radiotherapy in 28-30 fractions, for in total 44,262 and 16,423 voxels respectively). Using the logistic fit, a RD 50 (meaning 50% alopecia probability) was found at a physical dose of 3.2 Gy (single fraction), and 20.3 Gy (28- 30 fractions); RD 10 (10% alopecia probability) was found at 1.6 and 3.3 Gy, respectively (Figure 2). All fit parameters had a p-value << 0.01.

demonstrates a constant, β n

were the regression

where β 0

coefficients and x n were the predictive variables. The regression coefficients of the original model were: β 0 = - 1.8 and β 1 = 0.03 and the AUC was reported to be 0.62. Applying the original model to the validation dataset, the predicted NTCP was calculated for each single patient. Predicted probabilities were then arranged from lowest to highest probabilities and categorized into equally sized bins. The mean observed probabilities per each bin were plotted versus mean predicted probabilities to visualize calibration. Pearson’s correlation coefficient was 0.83, indicating a high correlation between predicted and observed probabilities. ROC curve was also plotted and AUC was 0.67 for the validation cohort.

Conclusion Mean dose to oral cavity was suggested to be a significant predictor of grade 3 oral mucositis for patients receiving IMRT with or without chemoradiation. The calibration plot demonstrated model’s external validity in an independent, plausibly related patient cohort. Considering a relatively low AUC, the future goal of our group would be to develop better discriminating models, by considering hidden, relevant predictors. References 1. Otter, Sophie, et al. "Evaluation of the risk of grade 3 oral and pharyngeal dysphagia using atlas-based method and multivariate analyses of individual patient dose distributions." International Journal of Radiation Oncology* Biology* Physics 93.3 (2015): 507-515.