18F-choline PET/CT in the study of primary hyperparathyroidism: Evaluation of the technique, visual and semi-quantitative analysis and correlation with other imaging techniques

Abstract

Objective

To assess the usefulness of performing a dual-time-point protocol in the acquisition of ¹⁸F-choline (¹⁸F-FCH) PET/CT in the pre-surgical localization of PHPT, and to demonstrate the impact of this imaging technique on the management and outcome-based surgical decision making, compared to other imaging techniques. To evaluate the diagnostic performance of the test to discriminate between pathological parathyroid gland and cervical lymph node, as well as to establish its correlation with other imaging techniques (scintigraphy, ultrasound, CT and MRI).

Patients and methods

We included 39 patients who underwent surgery for PHPT, in whom dual-time-point ¹⁸F-FCH PET/CT was performed. Metabolic index of parathyroid (P-SUVmax; P-SUVpeak), lymph node (N-SUVpeak), thyroid (T-SUVpeak) and mediastinum (M-SUVpeak) uptake were analyzed visually and semiquantitatively in both images. PET/CT results were correlated with ^99mTc-MIBI scintigraphy, ultrasound, MRI and CT.

Results

In 36 patients (92%), PET/CT was positive, localizing 38 pathological glands. The sensitivity (S) of PET/CT was 97% and positive predictive value (PPV) 94%. In the visual analysis, dual-time-point protocol was necessary in 61% of the cases. Correlation between PET/CT with MRI was 80%, with 4D-CT 50%, and with the other techniques <50%. P-SUVmax shows correlation with adenoma weight and size, and with presurgical PTH. The best cutoff point for SUVpeak to differentiate parathyroid vs. lymph node was 2.6 in early images (S = 70%; specificity = 75%; p = 0.007) and 0.86 for SUVpeak/T-SUVpeak index (S = 73%; specificity = 69%; p = 0.001).

Conclusion

¹⁸F-FCH PET/CT is an excellent preoperative localization technique in patients with HPTP with negative, doubtful or inconclusive imaging techniques, being of vital importance in guiding minimally invasive surgery. The dual-time-point protocol was necessary in more than half of the cases (61%). The SUVpeak cut-off points to discriminate between parathyroid gland and lymph nodes were statistically significant.