Remote frozen section examination of parathyroidectomy specimens by telepathology using Mikroscan D2 and Aperio LV1 : a validation study
Edward Chandraratnam, L Santos, Shaun Chou, Jun Dai, L Syeda, Guy D. Eslick, Ronald Yoon-Kong Chin · Internal Medicine Journal · 2017
Patients presenting with primary hyperparathyroidism have one of the three possible histopathological diagnoses: (i) parathyroid hyperplasia, (ii) parathyroid adenoma and (iii) parathyroid carcinoma. Each of these requires different surgical approaches: (i) removal of 3.5 glands, (ii) removal of adenoma and (iii) removal of carcinoma plus ipsilateral neck dissection respectively. Therefore, intraoperative pathology consultation plays a key role in parathyroidectomy. Telepathology, where the slide is interpreted off-site through dedicated telepathology robotic systems, may overcome some of the logistical difficulties traditionally associated with in-house microscopy for frozen section interpretation, for example availability of a local surgical pathologist, travelling time and time spent waiting for surgeons to provide specimens. The aim of this study is to validate two robotic telepathology (RT) systems (Aperio LV1 and Mikroscan D2) for parathyroid frozen section diagnosis. To do this, we compared the diagnostic accuracy and speed taken to make a diagnosis through the two RT systems with in-house frozen section diagnosis through microscopy. Three pathology services in Sydney (Austpath Laboratories, SWAPS and ICPMR) each submitted 20 consecutive parathyroidectomy cases with frozen section specimens to the study. This resulted in a total of 60 cases, with 84 slides. Three pathologists (EC, LS and SC) interpreted these cases: first with the Aperio LV1 system, then with the Mikroscan D2 system and finally with in-house microscopy. Each pathologist had a 2-week washout period between the three systems. We recorded the diagnoses and the time taken to arrive at the diagnosis. Aperio LV1 had a slightly higher rate of concordant diagnoses with in-house microscopy compared to Mikroscan D2 for pathologists EC (92.9% vs 90.5%) and LS (95.2% vs 94.7%). We recorded a lower rate of concordant diagnoses for pathologists SC/MF with Aperio LV1 compared to Mikroscan D2 (79.8% vs 88.1%). This may have been due to inter-observer variability in addition to inter-method variability, as SC interpreted the slide set with in-house microscopy and Mikroscan D2, while MF completed the slide set with Aperio LV1. Average slide interpretation time was 35.9 s for in-house microscopy, 94.4 s for Aperio LV1 and 232.5 s for Mikroscan D2. RT with Aperio LV1 and Mikroscan D2 for parathyroid frozen section interpretation has a high rate of concordant diagnoses compared with in-house microscopy. Slide interpretation times with these two RT systems are longer than in-house interpretation but less for Aperio LV1 compared with Mikroscan D2.