Oral Scientific SessionsValidation of Temporal Change in Quantitative PET Metrics as Predictors of Recurrence in Early Stage Lung Cancer Patients Treated With SABR

To validate previously reported prognostic quantitative PET metrics which change over time, specifically standard uptake value velocity (SUVV) and metabolic tumor growth rate (MTGR), as predictors for tumor recurrence and survival. The charts of patients with early stage lung cancer who were enrolled and treated on an ongoing clinical trial of individualized lung tumor stereotactic ablative radiotherapy (iSABR) between January 2012 and April 2016 were retrospectively reviewed under an institutional review board-approved protocol. All patients had a diagnosis of primary lung cancer and underwent PET/CT scanning at the time of staging and at treatment simulation. Dynamic metrics, specifically the change in lesion SUVmax between scans per unit time (SUVV) and the change in metabolic tumor volume (MTGR) were measured. Metabolic tumor volume was measured using a gradient-based segmentation algorithm. Previously reported SUVV threshold (0.08 U/day), which was predictive of survival, was tested on the validation cohort. Pooled analysis of the study cohort and the previously reported training cohort was performed via univariate and multivariate Cox proportional hazards regression to evaluate the association of clinical variables on recurrence, progression free survival (PFS), and overall survival (OS). The validation cohort contained 60 lesions in 60 patients and the pooled analysis evaluated 156 lesions in 144 patients. Demographics of the validation cohort were similar to the training cohort, including median age (75 vs 77 years), histology (68 vs 56% adenocarcinoma, 18 vs 30% squamous cell carcinoma), and T-stage (63% vs 69% T1). In the validation cohort, 53 patients (88%) had SUVV below the threshold; survival in this subgroup was not different from the training subgroup with SUVV less than 0.08 (P = 0.06). In the pooled analysis, mean SUVV was 0.023 U/day (SD 0.176) and mean MTGR was 0.063 ml/day (SD 0.167). Patients underwent SABR delivering a median 50 Gy (range 20-66 Gy) in 4 fractions (range 1-8 fractions) with median equivalent biological effective dose 105 Gy (range 60-180 Gy). Local, regional, and distant recurrence rates were 8% (N = 5), 15% (N = 9), and 17% (N = 10) in the validation cohort, respectively, after median 20 months follow-up. SUVV threshold of 0.08 U/day remained significantly associated with survival in the pooled analysis, with a median OS benefit of 9 months (P = 0.04) for slower SUVV. MTGR less than 0.04 ml/day was associated with improved survival (median 41 vs 22 months, P = 0.01). In multivariate models, SUVV was associated with regional (HR 3.1, P = 0.03) and distant recurrence (HR 3.6, P = 0.01). Increasing SUVV and MTGR are poor prognostic features, associated with worse overall survival. High SUVV has increased risk of regional and distant failure. Threshold values for these metrics are reported and may serve as benchmarks for risk stratification.