Prognostic value of positron emission tomography-assessed tumor heterogeneity in patients with thyroid cancer undergoing treatment with radiopeptide therapy

Introduction: Peptide receptor radionuclide therapy (PRRT) is a treatment option for both iodine-refractory differentiated and advanced medullary thyroid cancer (TC). It requires over-expression of somatostatin receptor subtype II (SSTR) that can be non-invasively assessed by positron emission tomography (PET). Assessment of tumor heterogeneity is increasingly used as a tool for prognostication prediction. We investigated the potential of SSTR-PET to assess intraindividual tumor heterogeneity and thereby treatment response prior to PRRT. Methods: 12 patients with progressive radioiodine-refractory differentiated (1 papillary, 1 oxyphilic, 2 oncocytic, 4 follicular) or medullary (n = 4) TC were enrolled. SSTR-PET was performed at baseline. Conventional PET parameters and heterogeneity parameters were analyzed regarding their potential to predict progression-free (PFS, mean, 221 days) and overall survival (OS, mean, 450 days). Parameters of a subgroup of lesions (n = 23) were also correlated with morphological response according to modified RECIST criteria. Results: In patient-based analysis, all conventional parameters failed to predict PFS. Several textural parameters showed a significant capability to assess PFS. Thereby, Grey level non uniformity had the highest area under the curve (AUC, 0.93) in Receiver operating characteristics analysis followed by Contrast' (AUC, 0.89). In lesion-based analysis, only Entropy revealed potential to evaluate disease progression. OS could not be assessed by any parameter investigated. Conclusions: Tumor heterogeneity seems to be a predictor of response to PRRT in patients with iodine-refractory differentiated/advanced medullary thyroid cancer and outperforms conventional PET parameters like standardized uptake value. In a theranostic approach, assessment of textural parameters may help in selecting patients who might benefit from PRRT. (C) 2014 Elsevier Inc. All rights reserved.