Optimization of Parathyroid Imaging for Small Adenomas

Nadine Mallak; Abhishek Kashyap; Jeffrey Murguia; Sonia Sugg; Michael Graham

Objectives: Tc99m Sestamibi imaging for parathyroid adenoma (PA) has been used for years. More recently the indication for such imaging has shifted to patients with relatively modest changes in parathyroid hormone (PTH) and Calcium levels which has made detection of PAs more challenging. Our goal in this study is to optimize the methodology to maximize the probability of detection of these small adenomas. Methods: We retrospectively reviewed patients who underwent surgery for hyperparathyroidism preceded by a parathyroid scan at our institution between Feb 2014 and Dec 2015. Parathyroid imaging consisted of planar images following administration of Tc99m sestamibi (MIBI) at 5, 10, 15 and 90 min in addition to SPECT/CT at 15 min and SPECT at 90 min. 2 hrs later Tc99m pertechnetate (TcO4) was administered and planar and SPECT images were acquired. Images were interpreted independently by 2 reviewers blinded to the clinical history, PTH levels and prior imaging; a consensus was reached in case of disagreement. Planar MIBI images were interpreted initially and scored on a scale from 1 to 5 for the presence of PA (1: definitely negative, 2: probably negative, 3: equivocal, 4: probably positive, 5: definitely positive). Then SPECT/CT MIBI images were evaluated and scored the same way; then SPECT TcO4 images were compared to SPECT MIBI images and scored. Correlation was made with surgical and pathological results for the presence, location, and weight of hypercellular parathyroid glands, i.e. PA. Results: 98 patients were included, with a total of 141 PAs found at surgery. Planar images had a sensitivity of 16% (95% confidence interval: 0.1-0.22), specificity of 98% (0.97-1.00), PPV of 86% and a NPV of 68%; average weight of lesions correctly identified was 0.82 g, and of the lesions missed 0.39 g. Dual phase MIBI SPECT had a sensitivity of 48% (0.40-0.56), specificity of 94% (CI: 0.91-0.97), PPV of 82% and NPV of 76%; average weight of lesions correctly identified was 0.64 g, and of lesions missed 0.29 g. Dual tracer MIBI/TcO4 SPECT had sensitivity of 64% (0.56-0.72), specificity of 93% (0.89-0.96), PPV of 83% and NPV of 82%; average weight of lesions correctly identified was 0.58 g, and of lesions missed 0.23g. Although there is a trend to detect smaller lesions on dual phase MIBI SPECT compared to MIBI planar imaging, no statistical significance is seen (P = 0.19), and the same trend is seen between dual tracer MIBI/TcO4 SPECT and MIBI SPECT (P = 0.14). However, there is a significant difference between the size of the lesions detected on dual tracer MIBI/TcO4 SPECT and those detected on MIBI planar imaging (P = 0.02). In terms of localization as superior versus inferior PA, 23 out of the 90 true positive lesions on dual tracer MIBI/TcO4 SPECT (26%) were mis-localized; all these lesions are located posterior to the inferior thyroid lobe on imaging, 16 of them were superior glands (70%) and 7 were inferior glands (30%). Conclusion: Dual tracer MIBI/TcO4 SPECT is the optimal methodology for detection of PA; sensitivity for lesion detection is significantly better than dual phase MIBI SPECT than with planar imaging. Dual tracer MIBI/TcO4 is significantly better than planar imaging for detection of smaller lesions. Differentiation of superior versus inferior PAs is challenging, since superior PAs are often posterior to the lower thyroid pole.

Optimization of Parathyroid Imaging for Small Adenomas

View Online

Abstract

Details

Metrics