ESTRO 2022 - Abstract Book

S403

Abstract book

ESTRO 2022

The median follow-up was 8.5 years, median pre-SRT PSA 0.54 ng/mL, median SRT EQD2 dose 72.6 Gy; 336/795 pts received PNI. The most clinically significant prognostic grouping identifying the high-risk cohort was found to be “pT3b and/or ISUP 4-5 Class”, as opposed to pT2/pT3a and ISUP 1-3 (low-risk group). Fits on the training cohort were successful with consistent best-fit parameters of: α eff=0.26/0.23Gy-1, C=10 7 , B=0.40/0.97, λ =0.87/0.41 for the low-risk (n=417) and high-risk (n=111) groups, respectively. The calibration was satisfactory (slope=0.79, R 2 =0.82). Performances were confirmed in the validation group (slope=0.91, R 2 =0.91). Results suggested an optimal SRT EQD2 dose in the range of 70 and 74 Gy for low- and high- risk groups, respectively. The estimated bRFS recovery obtainable with PNI corresponding to these dose levels is shown in the Figure: >5% for PSA>1 and 0.15 ng/mL in low-risk and high-risk pts, respectively, and >10% for high-risk pts with a pre- SRT PSA>0.25 ng/mL.

Conclusion An explainable one-size-fits-all equation satisfactorily predicts long-term bRFS after SRT. Model parameters were obtained by fitting a large multi-centric cohort, and the model was independently validated. Results suggest that a significant (and rapidly increasing for rising pre-SRT PSA values) impact of PNI may be expected in patients with pT3b and/or ISUP4-5 disease at prostatectomy. The added value of PNI is, on the contrary, negligible in the low-risk cohort for pre-SRT values up to 1 ng/mL. A calculation tool for individual estimates is available.

OC-0458 Combined radiomics and dosiomics predicts radiation pneumonitis : a model with external validation

Z. Zhang 1,2 , L. Wee 1 , L. Zhao 2 , Z. Wang 3 , A. Dekker 3

1 MAASTRO, Radiation Oncology, Maastricht, The Netherlands; 2 Tianjin Medical University Cancer Institute and Hospital, Radiation Oncology, Tianjin, China; 3 MAASTRO, Radiation Oncology , Maastricht, The Netherlands Purpose or Objective Radiation pneumonitis (RP) is one of the common side effects of radiotherapy in the thoracic region. Radiomics and dosiomics quantifies information implicit within medical images and radiotherapy dose distributions. In this study we demonstrated the prognostic potential of radiomics, dosiomics and clinical features for RP prediction. Materials and Methods Radiomics, dosiomics and clinical parameters were obtained on 314 retrospectively-collected and 35 prospectively-enrolled patients diagnosed with lung cancer between 2013 to 2019. A clinical risk score (C-score), radiomics risk score (R-score) and dosiomics risk score (D-score) were calculated based on logistic regression after feature selection. Seven models were built using different combinations of R-score, D-score, and C-score to evaluate their added prognostic power. Over- optimism was evaluated by bootstrap resampling from the training set, and the prospectively-collected cohort was used as the external test set. Model calibration and decision-curve characteristics of the best-performing models were evaluated. For ease of further evaluation, nomograms were constructed for selected models.