Abstract
201
Objectives Recently a nomenclature system for describing parathyroid lesion locations has been proposed (Perrier ND, et al. World J Surg. 2009;33:412-6). We investigated the effectiveness of different imaging options in localizing parathyroid lesions using this new nomenclature.
Methods Data were reviewed retrospectively for 76 pts (61 female, 15 male) with primary hyperparathyroidism. All pts had surgically confirmed lesion locations & weights. 2 experts reviewed 5 image sets: Planar (P) (early & late MIBI pinhole neck imaging along with pinhole 99mTcO4- thyroid imaging and subtraction), MIBI SPECT (S), MIBI SPECT/CT (SCT), Planar+SPECT (PS), and Planar+SPECT/CT (PSCT). Readers evaluated all images in separate reading sessions, independently of one another, without knowledge of other test results or final diagnoses. Studies were graded from 0 (definitely normal) to 4 (definitely abnormal) using the system of Perrier et al.
Results There were 104 parathyroid lesions among the 76 pts. 58 (76%) had single gland disease (SGD) & 18 (24%) had multi-gland disease (MGD) distributed as: 10 pts had 2 gland disease, 6 pts had 3 gland disease & 2 pts had 4 gland disease. Mean lesion wt. was 628±845 mg. Inter-rater agreement assessed by contingency table analysis was significantly better for SCT than for S (contingency coefficient = 0.69 versus 0.59, p=0.03), but coefficients were similar (p>0.05) between SCT and P, PS and PSCT (0.64, 0.64 and 0.63, respectively). ROC analysis indicated that for SGD, PSCT had the highest accuracy of 88±3%, significantly higher than S or SCT (74±4% and 79±4%) though not significantly higher than P (81±4%) or PS (83±4%). PSCT accuracy was higher than the other 4 methods for all lesions, SGD and MGD lesions, and was significantly higher for SGD than for MGD (88±3% and 65±6%, p<0.0001).
Conclusions Using new surgical nomenclature combined planar & SPECT/CT images localize parathyroid lesions more accurately than planar, SPECT and SPECT/CT alone and planar and SPECT in combination