Digital pathology is a practical alternative to on‐site intraoperative frozen section diagnosis in thoracic surgery

Aims Telepathology uses digitised image transfer to allow off‐site reporting of histopathology slides. This technology could facilitate the centralisation of pathology services, which may improve their quality and cost‐effectiveness. The benefits may be most apparent in frozen section reporting, in which turnaround times (TATs) are vital. We moved from on‐site to off‐site telepathology reporting of thoracic surgery frozen section specimens in 2016. The aim of this study was to compare TATs before and after this service change. Methods and results All thoracic frozen section specimens analysed 4 months prior and 4 months following the service change were included. Demographics, operation, sample type, time taken from theatre, time received by laboratory, time reported by laboratory, TAT, frozen section diagnosis, final histopathological diagnosis and final TNM staging were recorded. The results were analysed with spss statistical software version 24. In total, there were 65 samples from 59 patients; 34 before the change and 31 after the change. Specimens included 51 lung, six lymph node, three bronchial, three chest wall and two pleural biopsies. Before the change, the median TAT was 25 min [interquartile range (IQR) 20–33 min]. No diagnoses were deferred. No diagnoses were changed on subsequent paraffin analysis. After the change, with the use of digital pathology, the median TAT was 27.5 min (IQR 21.75–38.5 min). This difference was not significant (P = 0.581). Diagnosis was deferred in one case (3.23%). There was one (3.23%) mid‐case technical failure resulting in the sample having to be transported by courier, resulting in a TAT of 106 min. No diagnoses were changed on subsequent paraffin analysis. Conclusions There was no significant difference in reporting times between digital technology and an on‐site service, although one sample was affected by a technical failure requiring physical transportation of the specimen for analysis. Our study was underpowered to detect differences in accuracy.