Prognostic implication of metabolic parameters measured by 18F FDG PET/CT in patients of cervical cancer.

Objectives: To correlate metabolic parameters in FDG PET/CT with event free survival (EFS) in cervical cancer patients and to correlate the primary tumor metabolic parameters with nodal disease to identify a cut off to predict nodal metastases.

Methods: This retrospective observational study was approved by the institutional review board and institutional ethics committee. Waiver for informed consent was obtained. 80 eligible patients with cervical cancer who underwent baseline FDG PET/CT from January 2005 to December 2009 were retrospectively reviewed. The metabolic parameters (SUVmax, SUVmean, Metabolic Tumour Volume [MTV], and Total Lesion Glycolysis [TLG]) of the primary tumor, nodal disease and whole body were measured. Continuous PET parameters were stratified using optimal cut offs derived from ROC curve analysis; survival curves were estimated using the Kaplan-Meier method and were compared using the log-rank test. The Cox proportional hazards model was used in multivariate analysis to identify independent prognostic factors. Association of tumor metabolic parameters with presence of nodes was evaluated using Mann Whitney U Test. Cut offs to predict nodal metastases were calculated using ROC curves.

Results: Median follow-up was 94.5 months (5-134). 28 patients had events (either progression or recurrence), 54 patients were alive, 20 were dead, and 6 patients had lost to follow up with disease at the end of study. In univariate analysis, Tumor TLG, Nodal SUV max, Nodal SUV mean, Nodal MTV, Nodal TLG, Whole body MTV and Whole body TLG were significantly associated with worse EFS with p values of 0.030, 0.002, 0.00, 0.002, 0.002, 0.009, and 0.002 respectively. In the multivariate analysis, only Nodal SUV mean had statistically significant association with EFS (p value=0.025 and HR=14.036). Using Mann Whitney U test it was observed that Tumor MTV and Tumor TLG had a strong association with presence of metastatic nodes (p=0.001, 0.001 respectively). Optimum cut offs in Tumor TLG and Tumor MTV to predict different level of nodes (nodes+/-, pelvic nodes+/-, para aortic nodes+/-) are 225.52, 244.52, 306.93 and 21.92, 23.38, 30.24 respectively.

Conclusions: Staging FDG PET/CT prognosticates cervical cancer patients at baseline by mapping the metabolic disease extent. Volume based metabolic parameters such as MTV and TLG can provide a more complete estimation of the true volume and biological aggressiveness; can contribute a quantitative prognostic measure to the FIGO staging system. Nodal SUV mean was an independent significant prognostic factor in predicting event free survival in cervical cancer patients. MTV and TLG of primary tumor help to predict the extent of nodal involvement thus having the potential to identify high risk patients who can be treated and followed up aggressively.