Repeat thyroid FNAC: Inter-observer agreement among high- and low-volume centers in Naples metropolitan area and correlation with the EU-TIRADS

Abstract

Our institution (University Hospital “L. Vanvitelli” - Naples, Italy) is a high-volume (HV) center in Naples metropolitan area and many patients are referred there to repeat thyroid fine-needle aspiration cytology (FNAC) after initial FNAC performed in low-volume institutions (LV). The aims of the study were to 1) examine the inter-observer agreement between HV and LV institutions according to the Italian thyroid cytology system, and 2) explore how the discordant FNAC reports were distributed in the European Thyroid Imaging and Reporting Data System (EU-TIRADS) categories. All consecutive cases of repeat FNAC performed at University Hospital “L. Vanvitelli” from January 2016 to December 2021 were retrospectively reviewed. Fleiss’ kappa (κ) was used to assess the inter-observer agreement, and categorical variables were compared by chi-square testing. P < 0.05 was considered statistically significant. A total of 124 nodules from 124 adults (mean age 49 years; mean maximum diameter 19 mm) were evaluated. Initial FNAC reports at LV were: 4 (3.2%) TIR1c, 64 (51.6%) TIR2, 48 (38.7%) TIR3A, 8 (6.5%) TIR3B, 0 TIR4, 0 TIR5. The overall FNAC reports were significantly different between the LV and HV institutions. At repeated FNAC, cytological diagnosis was unchanged in 64 (51.6%) cases including TIR2 and TIR3A results. A downgraded FNAC diagnosis (i.e., TIR2 vs TIR3A, TIR2 vs TIR3B) was observed in 36 (29%) nodules. An upgraded FNAC diagnosis (i.e., TIR3B vs TIR2, TIR3B vs TIR3A, TIR4 vs TIR3A, TIR5 vs TIR2, TIR5 vs TIR3B) was recorded in 24 (19.4%) nodules. The weighted inter-observer agreement between LV and HV institutions was poor (κ=0.133). Changed FNAC results were significantly (p=0.0023) more frequent in nodules at intermediate/high-risk (i.e., EU-TIRADS 4/5) than in those at no/low risk (EU-TIRADS 2/3) [i.e., 32/48 (66.7%) and 28/76 (36.8%), respectively]. Downgraded FNAC results were significantly more frequent in EU-TIRADS 2/3 (p=0.001) while upgraded FNAC were present only in EU-TIRADS 4/5 (24/24, 100.0%). The inter-observer agreement among LV and HV thyroid services was poor. The EU-TIRADS 4 and 5 categories included all the malignant nodules with FNAC results reclassified as higher risk (i.e., TIR3B-TIR4-TIR5) by the high-volume cytology service.