ESTRO 2022 - Abstract Book

S769

Abstract book

ESTRO 2022

In Fig.1, we plotted four arbitrary levels of responsive pts, according to ERI values, from 1 to 35. As an example, numbers at the arrows border represent the dose corresponding to a 40% rate of pCR. The % values on the dashed and dotted lines represent the expected improvement from our current protocol (EQD2=54 Gy) to the dose planned to be delivered in a new dose escaleted protocol, delivering EQD2=65.3 Gy to the residual tumor, assessed at half-therapy. For pts with favorable ERI values, high pCR rates are expected (between 67% and 88% for ERI ranging between 15 and 1). On the other hand, the gain remains limited for poorly responding pts: for pts with ERI=35, doses >80-85Gy should be delivered to reach similar pCR rates of responding pts. Conclusion The suggested Logit model incorporating the individual response to RCT for rectal cancer, based on early regression, should be considered as a promising tool for therapy personalization both in identifying pts as candidates to avoid surgery and in better defining the best treatment intensification approach aimed to increase pCR substantially.

MO-0881 Dose patterns associated to radiation induced esophagitis in locally advanced NSCLC patients

G. Palma 1 , S. Monti 1 , T. Xu 2 , R. Mohan 3 , Z. Liao 2 , L. Cella 1

1 National Research Council, Institute of Biostructures and Bioimaging, Naples, Italy; 2 The University of Texas MD Anderson Cancer Center, Department of Radiation Oncology, Houston, USA; 3 The University of Texas MD Anderson Cancer Center, Department of Radiation Physics, Houston, USA Purpose or Objective Radiation-induced esophagitis (RE) is a common dose-limiting complication associated with concurrent chemoradiation therapy for Non-Small-Cell Lung Cancer (NSCLC). A wide range of esophageal dosimetric parameters have been described as predictive of RE; including the maximum dose to 2 cm 3 and generalized equivalent uniform dose [Paximadis et al. 2018]. Aim of our study is to characterize the risk of RE for NSCLC patients enrolled in the prospective trial NCT00915005 and to investigate the dose patterns associated with RE. Materials and Methods We retrospectively analyzed locally advance NSCLC patients treated with IMRT or Passive Scattering Proton Therapy to a prescribed dose of 66/74 Gy in conventional 2 Gy daily fractionation with concurrent chemotherapy. For each patient, RE was graded according to the National Cancer Institute’s CTCAE v. 4. Dose maps were converted into biologically effective dose (BED – α / β =10 Gy). The relationships between non dosimetric factors and RE were tested by Pearson’s χ 2 -test Mann- Whitney U test, when appropriate. A Voxel-Based Analysis (VBA) was performed to assess local dose differences between patients with and without RE of grade ≥ 2. To this end, planning CTs and BED maps were spatially normalized to a common anatomical reference after masking the gross tumor volume. The generalized linear model (GLM) was then designed to include dose maps and each non-dosimetric variables selected by a Least Absolute Shrinkage and Selection Operator (LASSO) regularized GLM of RE [Palma et al. 2020]. A non-parametric permutation test of the maximum threshold-free cluster- enhanced statistic accounting for multiple comparisons was performed and the significance p - maps were generated.

Results

Of the identified 202 patients, 91 (45%) developed RE of grade ≥ 2 (G2); 26 of 91 (29%) developed G3 RE. There were no

cases of grade 4 or 5 toxicity. Of these 202 patients, 29 had incomplete or corrupted CT/dose info. In the reduced cohort

(173 patients), a comparable RE incidence was found (44% for G2 RE). None of the clinical factors, including tumor height, significantly correlated with RE, hence the multivariable VBA was designed by including only dosimetric variables in the GLM. The VBA identified regions of significant association between BED and G2 RE along a roughly cylindrical region tightly encompassing the thoracic esophagus (Figure).