ESTRO 2020 Abstract book

S157 ESTRO 2020

and the best NTCP model were selected using Akaike weights (relative likelihoods). Differences between model parameters were tested for significance using the log likelihood ratio test by applying the TD 50 and m from the 2 distributions jointly or separately to the delivered dose. Results 80 symptomatic RIRF out of 8472 ribs (0.94%) were observed with a median time to fracture of 24.4 months and a median FU of 24.7 months. D 0 was the best RIRF predictor for both dose distributions. NTCP models of planned and delivered dose (Figure 1) presented Akaike weights of 0.28 and 0.72 respectively. The NTCP model based on planned dose showed a non-significantly higher TD 50 than for delivered dose, while m remained the same for both distributions. The model based on delivered dose gave a better fit (higher Akaike weights), which was not significant (Table 1).

Purpose or Objective Radiation therapy (RT) of thoracic cancers may cause acute severe radiation dermatitis (RD) with a profound impact on the quality of a patient's life. In general, modern RT techniques have achieved a reduction in RD incidence. However, the risk of a potential increase of skin toxicity has long been considered a peculiar drawback of proton therapy. Aim of this study was to develop normal tissue complication probability (NTCP) models for severe RD in thoracic cancer patients treated with Intensity- Modulated RT (IMRT) or Passive Scattering Proton Therapy (PSPT). Material and Methods We analyzed 166 Non-Small-Cell Lung Cancer (NSCLC) patients prospectively treated at a single institution with IMRT (103 patients) or PSPT (63 patients). All patients were treated to a prescribed dose of 60 to 74 Gy in conventional daily fractionation with concurrent chemotherapy. Median patient age was 66 years (range: 33–85 years). RD was scored according to CTCAE v3 scoring system. For each patient, the epidermis structure (skin) was automatically defined by an in house developed segmentation algorithm. The absolute dose-surface histogram (DSH) of the skin were extracted and normalized using the Body Surface Area (BSA) index as scaling factor. Patient and treatment- related characteristics were analyzed. Two different NTCP approaches were adopted: the pure dosimetric Lyman- Kutcher-Burman (LKB) model recast for DSH and the multivariable (MV) logistic model allowing for the inclusion of non-dosimetric predictors. Models were internally validated by Leave-One-Out (LOO) method. Model performances were evaluated by the area under the receiver operator characteristic curve (ROC-AUC) and the calibration plot parameters. Results By the end of RT, 118 of 166 (71%) patients developed acute RD of any degree. Fifteen of 166 (9%) patients developed severe dermatitis (grade 3) at a median time of 41 days (range: 28-51). RT technique did not impact RD incidence. Total gross tumor volume (GTV) size was the only non dosimetric variable significantly correlated with severe RD ( p =0.027). MV modelling for severe RD resulted in a single variable model including S 20Gy , the relative skin surface receiving more than 20 Gy (OR= 31.4, 95%CI 7.5- 131.7). The cut off for S 20Gy was 1,1% of the BSA. LKB model parameters were TD50 = 9.5 Gy [5.9-18.4], m = 0.24 [0.17- 0.35], n =0.62 [0.36-0.92] Both NTCP models showed comparable good discrimination (LOO ROC-AUC = 0.8 (95%CI 0.6-0.9) and calibration at cross validation (Figure). Conclusion Despite skin toxicity has long been considered a potential limiting factor in the clinical use of PSPT, no significant differences in RD incidence was found between RT modalities. Once externally validated, the availability of NTCP models for robust prediction of severe RD may advance treatment planning optimization. In addition, severe acute skin reactions may be prodromal of consequential skin late effects. Prediction and, possibly, prevention of these acute reactions may also avoid late effects.

Conclusion NTCP models of symptomatic RIRFs based on the delivered dose did not significantly improve compared to models based on planned dose. Both dose distributions can be used for risk prediction in the current planned dose-based NTCP models. PH-0285 NTCP models for severe radiation induced dermatitis after thoracic radiation therapy G. Palma 1 , S. Monti 1 , M. Conson 2 , T. Xu 3 , S. Hahn 3 , M. Durante 4 , R. Mohan 5 , Z. Liao 3 , L. Cella 1 1 Institute of Biostructure and Bioimaging-CNR, National Research Council of Italy, Napoli, Italy ; 2 Federico II University School of Medicine, Department of Advanced Biomedical Sciences, Napoli, Italy ; 3 MD Anderson Cancer Center, Department of Radiation Oncology, Houston, USA ; 4 GSI Helmholtz Centre for Heavy Ion Research, Department of Biophysics, Darmstadt, Germany ; 5 MD Anderson Cancer Center, Department of Radiation Physics, Houston, USA