Prognostic value of PET parameters in patients with pleomorphic lung cancer: Results from a single institution.

ABSTRACT

OBJECTIVES: Pleomorphic lung carcinoma (PLC) is a rare histotype of non-small cell lung cancer (NSCLC) characterized by aggressive clinical course, poor response to therapy and poor prognosis. Therefore, aim of our study is to analyze with 18F-FDG PET/CT a subset of patients affected by PLC to evaluate their metabolic characteristics in terms of SUVmax, MTV and TLG, in order to correlate them with overall survival (OS) and disease-free survival (DFS). MATERIAL AND METHODS: We retrospectively analyzed 49 consecutive patients with histologically defined PLC occurred to our Institution between 2003 and 2014. All patients underwent F18-FDG PET-CT before surgery and primary tumor was automatically segmented using an isocontour threshold method. SUV threshold for tumor segmentation was defined as the 41 % of lesion SUVmax. Total volume of the segmented VOI (MTV, centimeters cubed) and average SUV (SUVavg, grams per milliliter) in the segmented VOI were measured. RESULTS: In our population men were significantly more affected than women (427). According to Youden criteria, SUVmax, MTV41 and TLG41 best cut-off values to predict 2-year mortality were, 18.95, 27.89 and 290.45, respectively, with TLG41 showing best specificity (85 %) and positive predictive value (82.4 %). As concerning 2-year recurrence, SUVmax, MTV41 and TLG41 best cut-off values were 10.08, 27.89 and 134.85, with SUVmax showing best sensitivity (96.7 %) and negative predictive value (85.7 %). ROC curves confirmed that SUVmax, MTV41 and TLG41 were equally accurate to predict 2-year mortality and 2-year recurrence in our population. CONCLUSION: Metabolic biomarkers such as SUVmax, MTV and TLG can be used as a prognostic index for disease progression, recurrence and death in patients with PLC, independently from other clinical/pathological prognostic elements.