Correlation of surgeon-performed parathyroid ultrasound with the Perrier classification and gland weight

Background Ultrasound localisation of parathyroid glands correlates with gland weight. We hypothesise that gland identification is also dependent on anatomical location. Perrier et al. have described a uniform and reliable nomenclature for parathyroid locations. We aimed to correlate surgeon-performed ultrasound (SUS) with intra-operative Perrier classification and gland weight. Methods Review of a prospectively maintained single operator SUS database of 194 patients referred with non-familial primary hyperparathyroidism (PHPT) at a tertiary centre between 2010 and 2015. Patients underwent MIBI localisation as well as on table SUS. Intra-operative pathological gland locations were classified according to the Perrier nomenclature. Results Mean weight of pathological glands found and missed by SUS was 1.07 ± 0.1 g and 0.48 ± 0.08 g respectively ( p = 0.0001, unpaired t test). The weight of glands identified was greater than that of missed glands for each of the Perrier locations ( p < 0.001, Mann-Whitney). The proportion of pathological glands found at each Perrier location varied significantly ( p < 0.0001, Chi Square); so we find proportionally more B-, D-, E- and F-type glands and miss more A- and C-type glands. The median weight of glands missed on SUS varied significantly across the Perrier groups (Kruskal-Wallis, p = 0.0034) and suggests that SUS can miss quite large glands (> 0.5 g) in locations B, C and F; whereas missed glands in locations A, D and E were all small (< 0.5 g). Conclusion Whilst gland identification correlates well with gland weight, anatomical location has a significant impact on failure of localisation irrespective of gland weight. For the surgeon operating on PHPT patients with negative US localisation, particular attention should be paid to locations C, D and A as these are the sites where pathological glands are most often missed on pre-operative US.