D. Tong1, A. Sun2, A. Bezjak3, M. L. Yap4, X. Y. Ye5, and J. P. Bissonnette6; 1Princess Marget Cancer Centre, Toronto, ON, Canada, 2Radiation Medicine Program, Princess Margaret Cancer Center, University of Toronto, Toronto, ON, Canada, 3Radiation Medicine Program, Princess Margaret Cancer Centre, University Health Network, Toronto, ON, Canada, 4Liverpool and Macarthur Cancer Therapy Centres, sydney, NSW, Australia, 5Department of Biostatistics, Princess Margaret Cancer Centre, Toronto, ON, Canada, 6Radiation Medicine Program, Princess Margaret Cancer Centre, Toronto, ON, Canada
Purpose/Objective(s):Potentially curative chemoradiotherapy (CRT) is the treatment of choice in locally-advanced inoperable non-small cell lung cancer (NSCLC). Locoregional relapse is common but potentially salvageable in the absence of distant disease. 4DPET/4DCT image features during a course of CRT have been shown to predict clinical outcomes, two years post-CRT completion, including overall survival. The impact on five-year outcomes and the predictive power of these image features 3-month post-CRT are not clear. Materials/
Methods: In this prospective, REB-approved study, patients with LA-NSCLC receiving curative intent CRT had 4DPET/4DCT scans at 0, 2, 4, 7 weeks during CRT, and 3-month post-CRT. These patients were treated pre-Durvalumab approval (2010-12). Image features including V3SUV (volume represented by voxels with SUV>3), SUV50 (SUV>50% of SUVmax), Gross tumour volume on CT (GTV_CT) and their rates of change between timepoints were analyzed and correlated with clinical outcomes (locoregional relapse-free survival (LRRFS), overall survival (OS)) at 2 years and 5 years using mixed-effect models for longitudinal data. The predictive power of the features identified associated with the outcomes were further assessed by area under ROC curve using logistic regression. Results: Of 29 patients recruited, 25 completed all scans and were included for analysis. Median follow-up was 32 months. Median OS was 32 months and median LRRFS was 23.9 months. At 2 years, 15 were alive, 11 disease-free. At 5 years, 8 were alive and 6 disease-free. Five out of 6 that were relapse-free at 5 years remained disease-free at their last follow-up (median follow-up= 120 months). Multiple image features at weeks 0 and 2 reached statistical significance for OS and LRRFS at 2 years, including V3SUV_tumor and nodal disease volume(V3SUV_t+n), V3SUV_nodal disease volume(V3SUV_n) and GTV_CT_n(LRRFS only). V3SUV_t+n and V3SUV_n at week 2 of CRT most strongly predicted LRRFS (p=0.003, AUC 0.80) and OS (p=0.018, AUC 0.76) at 2 years. No features at 3-month post-CRT were predictive of outcomes. No features were predictive of 5-year LRRFS and OS. Conclusion: In this study of CRT without adjuvant immunotherapy, 5-year OS of 32% and LRRFS of 24% were observed, and the 3-month follow-up 4DPET/4DCT was not predictive of outcomes. 4DPET/4DCT at week 2 CRT has a potential to guide earlier salvage treatment in CRT non-responders.
