Malignancy Risk in 18F-FDG-Avid Thyroid Incidentalomas: Controversies and Limitations

Abstract

Introduction: The prevalence of malignancy in thyroid incidentalomas (TI) discovered on 18F-FDG-PET or PET/CT varies between 0% and 63.6%. The pooled malignancy rate according to three systematic reviews is 33-35%. The 2015 American Thyroid Association (ATA) guidelines recommend that such nodules, when one centimeter or larger in size, should undergo further investigation with thyroid ultrasound (US) and fine-needle aspiration (FNA) cytology. Objectives: The objective of our study was to determine the rate of malignancy amongst TI discovered incidentally on 18F-FDG-PET or PET/CT, examine their clinicopathologic characteristics, and assess the usefulness of maximum standardized uptake values (SUVmax) in differentiating benign and malignant lesions. Methods: We performed an electronic medical record search looking at all 18F-FDG-PET or PET/CT reports during the study period of 12/01/2015 to 05/31/2019 that included the keyword ‘thyroid’ in the impression. Exclusion criteria included a history of thyroid disease or malignancy, known lesion(s) detected on previous clinical or radiological examinations and diffuse radiotracer uptake. Of the 476 reports reviewed, 136 cases were included in the study. Results: Common indications included initial staging or restaging of lymphoma (diffuse large B-cell, mantle-cell, T-cell types) (27.9%), lung adenocarcinoma (18.4%), head and neck cancer (16.9%) and breast cancer (11%). Fifty-eight (42.6%) patients had metabolically inactive lesions; five (8.6%) underwent further investigation with thyroid US and 3 subsequently with FNA (5%). All 3 had benign cytology. Seventy-seven (56.6%) patients had metabolically active lesions and 25 (32.5%) underwent imaging with thyroid US. Twelve (15.6%) had FNA; eight (66.7%) had benign cytology, two (16.7%) revealed atypia of undetermined significance and two (16.7%) were malignant. Biopsy for the two patients with malignant cytology showed follicular cell neoplasm of oncocytic hurtle cell type, and invasive follicular carcinoma with focal insular and papillary features and extensive capsular and vascular invasion. The mean SUVmax in malignant vs benign lesions was 9.05 and 6.41 respectively. Conclusion: The malignancy rate was 2.6% amongst all patients with 18F-FDG-avid TI and 8% amongst patients with metabolically active lesions who were investigated with thyroid US+/- FNA. This is significantly lower than malignancy rates previously reported in the literature. The evident inhomogeneity in the literature is likely multifactorial and may be explained in part by a dissimilarity among studies, and an informed decision by some to avoid invasive testing in the context of poor prognosis from underlying non-thyroidal cancer. Research is needed to determine the cohort of patients who could potentially benefit from further evaluation and treatment.