Abstract 666: A phase 1 study to detect adverse events after SBRT and immunotherapy by electronic nose in advanced NSCLC

Abstract It is increasingly understood that cancers can be recognized by the immune system and that inflammation relates to response. Combining stereotactic radiotherapy (SBRT) to increase release of cancer cell antigens with an anti-CTLA-4 and PD-L1 inhibitor may lead to increased response rates. Due to tumor heterogeneity it may be important to irradiate the primary tumor (addressing trunk mutations), instead of its metastasis (branch mutations). No solid predictive biomarker for response and/or immunotherapy related adverse events (TRAEs) is available. Electronic nose (eNose) technology measures the complete mixture of volatile organic compounds in exhaled breath and can detect lung cancer based on pattern recognition of the breath profile. We hypothesize that eNose is able to predict TRAEs and response in patients with NSCLC. Methods: In 3 sequential cohorts, immunotherapy regimes combined with SBRT were studied in stage IIIB/IV NSCLC patients progressing on chemotherapy. All patients were irradiated on the primary tumor (1x20 Gy on 9cc) 1 week after the 1st dose of immunotherapy. The 1st cohort (n=3) received durvalumab. The 2nd and 3rd cohort (both n=6) received a combination of durvalumab and tremelimumab followed by durvalumab monotherapy. Duplicate eNose measurements were performed by using the SpiroNose that contains 7 metal oxide semiconductor sensors. TRAEs were categorized using NCI CTCAE version 4.3. Descriptive statistics were used to summarize baseline characteristics and TRAEs. The relationship between breath profiles and response and TRAEs was analyzed with advanced signal processing, ambient correction and Mann-Whitney U test. Linear discriminant and receiver operating characteristics analysis followed. Pearson correlation and regression assessed duration of response. Findings: Fifteen patients were included as described above. Baseline characteristics of the groups were comparable. Median progression free survival was 2 months, overall survival 10 months (immature). No statistical difference was found in (duration of) response. There was 1 low grade TRAE (CTC 1-2) in cohort 1 and 10 in cohort 2/3. High grade TRAEs (CTC 3) were only present in cohort 2/3 (n=3) and 1 patient discontinued treatment. There was one dose limiting toxicity. At baseline, 12/15 patients performed eNose measurements; 7 with TRAEs and 5 without. The TRAE group had a significant higher sensor 7 signal compared to those without TRAE (p=.042). The cross-validated accuracy for detecting TRAE was 67%. The ROC-AUC was .857 [.638-1]. Interpretation: eNose breath profiles at baseline may predict which patients will develop TRAEs. The small sample size increases the risk of overfitting, therefore another cohort of 34 immuno-monotherapy patients is being analyzed (15 with and 19 without TRAEs, pending). We demonstrated no new safety data. Disclosure: This study was sponsored by a research grant from AstraZeneca. Citation Format: M Benthe Muntinghe-Wagenaar, Hanneke Kievit, Lucie B. Hijmering-Kappelle, Birgitta I. Hiddinga, J Fred Ubbels, Robin Wijsman, Mechteld F. Brasz, Gitte Slingers, Rianne de Vries, Milou M. Schuurbiers, Harry J. Groen, Huib A. Kerstjens, Anthonie J. van der Wekken, Michel M. van den Heuvel, T Jeroen Hiltermann. A phase 1 study to detect adverse events after SBRT and immunotherapy by electronic nose in advanced NSCLC [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 666.