The association of the Bethesda reporting system and the AACE/ACE/AME US classification system for improving the diagnostic accuracy of thyroid nodules. An Institutional prospective cohort study

Giuseppina Orlando¹, Alessandro Corigliano¹, Sergio Mazzola², Irene Vitale¹, Riccardo Guercio¹, Marco Dominici¹, Claudia Carolla¹, Giulia Rotolo¹, Roberto Gullo¹, Giuseppe Carollo¹

¹Unit of General and Emergency Surgery, Department of General, Emergency and Transplant Surgery, Policlinico “P. Giaccone”, University of Palermo, Via L. Giuffré 5, 90100 Palermo, Italy - ²Unit of Clinical Epidemiology and Tumor Registry, Department of Laboratory Diagnostics, Policlinico “P. Giaccone”, University of Palermo, Via L. Giuffré 5, 90100 Palermo, Italy

Introduction: The Bethesda system is widely accepted for reporting thyroid cytopathology. Its association with ultrasounds (US) leads thyroid nodules’ diagnosis, expecially by using an efficacious US reporting system. Aim of our study is to evaluate if the association between the Bethesda system and a 3-categories US risk stratification system, proposed by American Association of Clinical Endocrinologists (AACE) / American College of Endocrinology (ACE) / Associazione Medici Endocrinologi (AME), improves the diagnostic performance compared to the cytology alone, decreasing the risk of false negatives. 

Materials and methods: 574 patients suffering from thyroid nodules that underwent US, cytology and thyroidectomy were prospectively recruited. We included: benign lesions, papillary thyroid carcinoma (PTC), follicular or Hürtle cells carcinoma (both included in FTC), aggressive tumors (AT). All patients were stratified and grouped according to definitive histology as a reference. In each group the cytology and the US risk were intersected each other.  The chi-square test was applied to verify the concordance US/cytology. Negative and positive predictive values (NPV and PPV) were also calculated, as well as the false negatives (FN) with 95% confidence interval (CI). Rstudio IDE (33) software version 3.4.1 (2017-06-30) was used.

Results: Among 231 benign, chi-square test showed a p-value = 0,2 (non-significant), that means an overlapping US/cytology non-useful in improving diagnosis; NPV=58%. In 293 PTC, chi-square test resulted in p<0,05; PPV = 57%, FN = 14% (95% CI = 7%-15%). In 31 FTC PPV was 68%,  FN 12,9% (95% CI = 1%-25%), in 19 AT PPV = 89%, FN = 5,3% (95% CI = 0-10%).  In Bethesda 3 examined isolatedly, we found predictive values indicating risk of malignancy in intermediate and even more in high US risk. In Bethesda 4, predictive values indicated high risk of malignancy in all US risk classes.

Conclusion: The US risk stratification system proposed is easy to apply. Its association with cytology is useful expecially for reducing underestimated malignancies: a reduced rate of FN and a satisfying accuracy were observed in PTC and even more in aggressive tumors. In Bethesda 3 and 4 categories, the risk of malignancy could be stratified in different levels.