Abstract Book

S1167

ESTRO 37

2 University of Illinois at Chicago- Chicago- Illinois- USA, Computer Science, Chicago, USA 3 University of Minnesota of Public Health- Minneapolis- Minnesota- USA., Biostatistics, Minneapolis, USA 4 University of Iowa- Iowa City- IA- USA, Electrical & Computer Engineering, Iowa City, USA Purpose or Objective The aim of this work is to investigate the applicability of combining serially derived radiomic features with standard clinical variables into a xerostomia probability computational tool. Material and Methods Patients undergoing RT for oropharyngeal cancer with daily CT-on-rails imaging were reviewed. Xerostomia status at 6-months post-treatment was retrieved as well as chief clinical variables, (e.g. age, smoking status, T- category, chemotherapy, and dose received). A total of 437 scans were analyzed. Ipsilateral parotid glands were contoured on baseline CT and propagated to daily CT images using deformable vector fields generated for IGRT until mid-treatment time point. A total of 145 radiomic features were selected from the categories: intensity direct, neighborhood intensity difference, grey-level co- occurrence matrix, grey-level run length and shape. Spearman correlation was used to reduce the 145 features to 5 features based on a cutoff of 0.7. These features included: ‘LocalEntropyStd’, ‘LocalStdStd’, ‘Compactness2’, ‘Volume’ and ‘Contrast’. These were then used to build and evaluate three distinct types of models (1) using only the baseline (BL) value of the radiomic feature, (2) the ratio between the mid-RT value of the radiomics feature to its initial value at BL, and (3) a functional principal component analysis (FPCA) model that leverages the structure of the temporal trajectory in the evolution of the radiomic features from BL to mid-RT. Afterwards, we ran logistic regression was to explore the capacity of relevant clinical variables to predict xerostomia at 6-month post-RT; either alone or in combination with one of the radiomic-derived models. Results 28 patients were included. At 6 months, xerostomia was reported as follows: minimal or mild (70.4%) vs moderate-severe (29.6%). The corresponding Receiver operating characteristics area under the curve (ROC AUCs) and confidence intervals (CIs) for various predictive models plotted and depicted in Figure & Table 1 . Clinical only model showed worse AUC when compared with any composite clinical/radiomics models. Combination radiomics model included input from all 3 radiomics models: baseline, delta & FPCA. Noteworthy, each of the 3 individual radiomics models performed worse than the clinical model with corresponding AUCs of: 0.59, 0.64, and 0.72, respectively. This suggests that the functional approach yields a superior ROC compared to using either the baseline radiomics feature or using the delta ratio between the mid and initial time points. Table 1. Color legend, AUC and confidence intervals of various models ROC curves

TE2=3.5, 4.4, 8.8, 9.5ms. To acquire the data, TR value as 8.6,9, 13.4, 14.1ms when possible, a scan-time constraint below 12min and a flip angle FA=12° were chosen. N small spheres ( r=4mm ), selected in bones and soft-tissues regions, were used for CNR estimation. Synthetic CT (sCT) was generated by an automatic segmentation of DTE into bones, soft-tissues, and air and then setting bulk Hounsfield units to the voxels. It was compared with the reference CT. CBCT and kv images (lateral, anteroposterior) of the head phantom were acquired using the Linear Accelerator for the assessment of MR-only IGRT. The results of 3D/3D and 2D/2D auto- matching, using the TPS Eclipse TM v11.0 (Varian Medical System), were compared between sCT and CT or between derived DRRs. Results The optimal SNR for good head region coverage was obtained for an isotropic resolution 1.3x1.3x1.3mm 3 , a matrix size of 186x186x100, NEX=2, and RBW=62.5 KHz with respect of chemical shift artefacts below 1mm. Figure 1.d shows variation of CNR according to TR and TE2 values. The scan-time increased with TR values. TE2=4.4ms, TR=9ms gave the best CNR/scan-time compromise with CNR=4.66 and scan-time=7:25min and was used for studying the clinical feasibility. sCT had a mean absolute error MAE=113HU and a Dice coefficient in bones DI bone =0.8. Figure 1.a, 1.b, 1.c show the results of the fusion of registered sCT or DRR sCT with the linac’s images. The largest deviations of sCT/CBCT registration parameters were <1.1mm and <0.7° for translation and rotation parameters respectively. For couch shifts in vertical, longitudinal and lateral directions, the largest deviations were <0.8mm after DRR sCT /kV images automatching.

Conclusion 3D-UTE-Cones were optimized and tested for clinical implementation of MRI-only IGRT. In future work, the feasibility will be demonstrated on real patients with brain or H&N tumor. EP-2121 Serial Parotid Gland Radiomic-based Model Predicts Post-Radiation Xerostomia in Oropharyngeal Cancer H. Elhalawani 1 , A.S.R. Mohamed 1 , A. Kanwar 1 , A. Dursteler 1 , C.D. Rock 1 , S.E. Eraj 1 , M. Meheissen 1 , S. Volpe 1 , P. Yang 1 , R. Granberry 1 , R. Ger 1 , X. Fave 1 , L. Zhang 1 , J. Yang 1 , G.E. Marai 2 , D. Vock 3 , G. Canahuate 4 , D. Mackin 1 , L. Court 1 , G.B. Gunn 1 , A. Rao 1 , C.D. Fuller 1 1 The University of Texas- MD ANderson Cancer Center, Radiation Oncology, Houston, USA

Figure 1. ROC curve for radiomics features kinetics till mid-RT paired with key clinical variables