FDG-PET/CT for discrimination between malignant necrotic lymph nodes and benign cystic lesions in the neck

Objectives Patients presenting with cystic lesions in the neck without obvious signs of malignancy constitute a diagnostic challenge. They may be a sign of underlying malignancy. We examined the diagnostic performance of (18)F-fluorodeoxyglucose (FDG)-PET/CT to identify malignant etiologies.

Methods We retrospectively included patients referred from the Department of Head and Neck Surgery from January 2011 until December 2012 for FDG-PET/CT scans because of a lateral neck cyst. Scan results were compared to clinical workup findings by histopathology and follow up. A scan was considered positive if the cystic lesion was FDG-avid and/or if a primary tumor was identified. This finding was considered true positive if confirmed by biopsy and false positive if not confirmed by biopsy and/or follow-up. A scan without abnormal FDG-uptake was considered false or true negative depending on whether follow-up revealed malignancy or not.

Results The study comprised of 32 patients, 10 females and 22 males. Median age was 56 years (range 44-82). Fifteen were diagnosed with cancer during follow-up. Median follow-up was 29 months (range 11-42). PET/CT suggested malignancy in 23 patients (14 true positive, 9 false positive) and showed no malignancy in 9 (8 true negative, 1 false negative). The sensitivity, specificity, accuracy, positive and negative predictive values (NPV) were 93% [70%-99%], 47% [26%-69%], 69% [51%-82%], 61% [41%-78%], and 89% [57%-98%], respectively (Wilson score based 95% confidence intervals in brackets). The primary tumor was identified by PET/CT in 10 out of the 15 patients with cancer.

Conclusions Malignant cystic lymph nodes may present with a thin metabolically active periphery and, hence, mimic benign cysts by FDG-PET/CT. Still, we found that FDG-PET/CT could reliably rule out malignancy in cystic neck lesions (NPV 89%), albeit with a high frequency of false positive scans, requiring further diagnostic work-up.