Optimal FDG PET/CT volumetric parameters for risk stratification in patients with locally advanced non-small cell lung cancer: results from the ACRIN 6668/RTOG 0235 trial

Abstract

Purpose: In recent years, multiple studies have demonstrated the value of volumetric FDG-PET/CT parameters as independent prognostic factors in patients with non-small cell lung cancer (NSCLC). We aimed to determine the optimal cut-off points of pretreatment volumetric FDG-PET/CT parameters in predicting overall survival (OS) in patients with locally advanced NSCLC and to recommend imaging biomarkers appropriate for routine clinical applications.

Methods: Patients with inoperable stage IIB/III NSCLC enrolled in ACRIN 6668/RTOG 0235 were included. Pretreatment FDG-PET scans were quantified using semiautomatic adaptive contrast-oriented thresholding and local-background partial-volume-effect-correction algorithms. For each patient, the following indices were measured: metabolic tumor volume (MTV), total lesion glycolysis (TLG), SUVmax, SUVmean, partial-volume-corrected TLG (pvcTLG), and pvcSUVmean for the whole-body, primary tumor, and regional lymph nodes. The association between each index and patient outcome was assessed using Cox proportional hazards regression. Optimal cut-off points were estimated using recursive binary partitioning in a conditional inference framework and used in Kaplan-Meier curves with log-rank testing. The discriminatory ability of each index was examined using time-dependent receiver operating characteristic (ROC) curves and corresponding area under the curve (AUC(t)).

Results: The study included 196 patients. Pretreatment whole-body and primary tumor MTV, TLG, and pvcTLG were independently prognostic of OS. Optimal cut-off points were 175.0, 270.9, and 35.5 cm³ for whole-body TLG, pvcTLG, and MTV, and were 168.2, 239.8, and 17.4 cm³ for primary tumor TLG, pvcTLG, and MTV, respectively. In time-dependent ROC analysis, AUC(t) for MTV and TLG were uniformly higher than that of SUV measures over all time points. Primary tumor and whole-body parameters demonstrated similar patterns of separation for those patients above versus below the optimal cut-off points in Kaplan-Meier curves and in time-dependent ROC analysis.

Conclusion: We demonstrated that pretreatment whole-body and primary tumor volumetric FDG-PET/CT parameters, including MTV, TLG, and pvcTLG, are strongly prognostic for OS in patients with locally advanced NSCLC, and have similar discriminatory ability. Therefore, we believe that, after validation in future trials, the derived optimal cut-off points for primary tumor volumetric FDG-PET/CT parameters, or their more refined versions, could be incorporated into routine clinical practice, and may provide more accurate prognostication and staging based on tumor metabolic features.

Keywords: Clinical trial; FDG-PET; Metabolic tumor volume; Non-small cell lung cancer; Risk stratification.

Appendix Figure 1

Area under the time-dependent (cumulative sensitivity/dynamic specificity) receiver operating characteristic curve (AUC(t)) corresponding to overall survival (OS) over the range of observed follow-up time, demonstrating the discriminatory abilities of whole-body (A) and primary tumor parameters (B).

Appendix Figure 2

(A) Scatterplot of whole-body SUVmean versus pvcSUVmean (B) Plot of percentage of difference between whole-body SUVmean and pvcSUVmean by tumor size (C) Scatterplot of whole-body TLG versus pvcTLG (D) Plot of percentage of difference between whole-body TLG and pvcTLG by tumor size.

Figure 1

Area under the time-dependent (cumulative sensitivity/dynamic specificity) receiver operating characteristic curves (AUC(t)) corresponding to overall survival (OS) over the range of observed follow-up time for whole-body and primary tumor MTV, TLG and pvcTLG. Appendix Figure 1 shows the AUC(t) for all parameters.

Figure 2

Kaplan-Meier (KM) curves corresponding to overall survival (OS) for prognostic whole-body and primary tumor FDG-PET/CT parameters (MTV, TLG, pvcTLG), dichotomized by their respective derived optimal cut-off points.

Figure 3

Area under the time-dependent (cumulative sensitivity/dynamic specificity) receiver operating characteristic curves (AUC(t)) corresponding to overall survival (OS) over the range of observed follow-up time, comparing TLG to pvcTLG and SUVmean to PVC SUVmean of both primary and whole-body parameters. Appendix Figure 1 shows the AUC(t) for all parameters.