Objective:

To compare the unnecessary fine-needle aspiration biopsy (FNAB) rates and performance of FNAB of 2017 American Colle

ge of Radiology (ACR) Thyroid Imaging Reporting and Data System (TI-RADS) and 2015 American Thyroid Association (ATA) guideline for thyroid nodules.

Methods:

All consecutive patients with thyroid nodules ＞1 cm who underwent thyroidectomy at Peking Union Medical College Hospital were retrospectively reviewed from December 2014 to January 2016. According to the ACR TI-RADS and ATA guideline

the performance of FNAB and the rate of unnecessary biopsy were calculated. Final diagnosis of malignancy was based on histopathology. The size threshold suggested by the ACR TI-RADS and ATA guideline for FNAB were applied to the ATA guideline

ACR TI-RADS

respectively.

Results:

Of the 1 427 thyroid nodules

797 were malignant

and 630 were benign. ACR TI-RADS showed better accuracy (AUC: 0.67

vs

0.57;

P

＜0.001). ACR TI-RADS had a higher specificity (42.9%

vs

14.3%

P

＜0.05) and lower unnecessary biopsy rate (33.0%

vs

40.6%

P

＜0.001)

whereas the ATA guideline pattern yielded a higher sensitivity (99.2%

vs

91.7%

P

＜0.001). Applying the FNAB size threshold recommended by the ACR TI-RADS to the ATA guideline

the unnecessary biopsy rate was significantly decreased (31.5%)

and the performance of FNAB was significantly improved (AUC: 0.70).

Conclusion:

ACR TI-RADS had a higher specificity and lower unnecessary biopsy rate for nodules ＞1 cm compared to ATA guideline pattern. Applying the FNAB size threshold recommended by the ACR TI-RADS to the ATA guideline

the unnecessary biopsy rate was significantly decreased and the performance of FNAB was significantly improved.