ESTRO 2023 - Abstract Book

S535

Sunday 14 May 2023

ESTRO 2023

Results Our AI model yielded an area under the receiver operating characteristic curve (AUC) of 0.714 (95% CI: 0.693 – 0.734) on the hold-out dataset after bootstrapping with 50 iterations. The AUC for the standard dose and high dose subsets were 0.696 (95% CI: 0.685 – 0.721) and 0.721 (95% CI: 0.705 – 0.738), respectively. The AUC for our model was significantly higher (p < 0.001) than the logistic regression model trained using only clinical and dosimetric variables (AUC: 0.597 ; 95% CI: 0.582 – 0.621). The overall AUC of our model went up to 0.747 (95% CI: 0.726 – 0.768) by ensemble voting using the top five bootstrap models (ranked by AUC) . Conclusion We developed an AI model to predict post-treatment cardiac toxicity for NSCLC patients receiving chemo-RT. The model significantly outperformed conventional logistic regression analysis. The predictive accuracy may be improved by expanding the training dataset and developing a model specific to patients treated with standard RT doses. Future work aims to: (1) Expand the model to predict oesophageal and pulmonary toxicities; (2) Externally validate predictions on a real-world dataset, initially on a cohort being collected at University College London Hospitals. MO-0643 Dose to heart substructures is associated with grade ≥ 3 cardiac events after SABR in stage I–II NSCLC M. Cerrato 1 , F. Menegatti 1 , I. Bonavero 1 , C. Casale 1 , C. Grossi 1 , S. Badellino 1 , R. Parise 1 , U. Ricardi 1 , M. Levis 1 1 University of Turin, Department of Oncology, Turin, Italy Purpose or Objective The aim of this prospective study is to investigate for any existing association between the dose delivered to heart substructures and OS, non-tumor-related survival, and the development of cardiac toxicity, in early stage (ES) NSCLC patients treated with SABR. Materials and Methods Patients with PS ECOG score 0-1 and no previous radiation treatment to the mediastinum were included. Details on pre- existing cardiovascular risk factors and comorbidities, ongoing pharmacological therapies and lung function tests were collected at baseline. Cardiac substructures (chambers, and coronary arteries) were prospectively contoured in order to collect dosimetric data. Cardiac events (CEs) were evaluated with the CTCAE 5.0 grading. Cox regression model was used to assess associations between covariates and CEs. The Fine and Gray competing risk regression model was used to predict the cumulative incidence of grade (G) ≥ 3 CEs considering non-cardiac death as a competing risk. Results From March 2019 to September 2022, 104 patients were enrolled with a median age of 76 years. The median overall survival (OS) was of 34 months (1-year OS: 83%; 2 year-OS 65%). Death was documented in 28 (27%) patients, in 14 (50%) for non- neoplastic causes. 51 (49%) patients had a pre-existing cardiac disease. Overall, 21 patients (20%) developed at least one CE during the follow-up and 15 patients had a first CE of G ≥ 3 at a median of 4 months (Interquartile Range: 2.5 to 20 months). Cox multivariable analysis shows that pre-existing cardiac disease (hazard ratio [HR]: 5.86; p= 0.008), mean right atrium (RA) dose (HR: 1.28, p= 0.01) and descendent aorta dose (10 cc) (HR: 1.08, p= 0.017) were associated with the development of a G ≥ 3 CE. Patients with pre-existing heart conditions had a higher cumulative incidence of the first G ≥ 3 CE (12 months- cumulative incidence was 11% and 3.1 %, respectively; p=0.001) ( Figure 1 ) and a decreased G ≥ 3 Cardiac Event-Free Survival (p=0.005) ( Figure 2 ) compared to patients without any cardiac comorbidity. In competing-risk multivariable analysis, pre-existing cardiac diseases (HR: 4.32; p= 0.046), mean RA dose (HR:1.22; p=0.007) and diuretic therapy (HR: 2.57, p= 0.040) showed a statistically significant association with G ≥ 3 CEs. However, the development of a G ≥ 3 CE did not affect OS (p=0.4). Heart doses were not associated with non-tumor-related survival or OS.