Korean Society of Thyroid Radiology Thyroid Imaging, Reporting and Data System (K-TIRADS)

Dr Henry Knipe and Dr Derek Smith et al.

K-TIRADS is a reporting system designed by the Korean Society of Thyroid Radiology for ultrasound assessment of thyroid nodules and stratification of the requirement for FNA and malignancy. There is also stratification of indications for lymph node sampling.

This is a five-stage system using descriptive, pattern recognition findings on ultrasound, and size measurement.

A number of established benign nodules features (purely cystic/anechoic, comet tail artefact, spongiform) are included in the benign K-TIRADS 2 group.

Assessment of echogenicity, shape/contour and three specific suspicious ultrasound features (microcalcifications, nonparallel orientation [taller than wide], spiculated / microlobulated margin) determine categorization.

Partially cystic or isohyperechoic nodules are considered lower risk than solid hypoechoic nodules.

  • K-TIRADS 1: no nodule
  • K-TIRADS 2: benign
  • K-TIRADS 3 (low): partially cystic / isohyperechoic with no suspicious features
  • K-TIRADS 4 (intermediate): as for K-TIRADS 3 but with any suspicious features or as for K-TIRADS 5 without suspicious features
  • K-TIRADS 5 (high): solid hypoechoic nodule with any suspicious features

Use of FNA / sampling is a combination of the suspicion above and size, unless there are obvious poor prognostic factors (extrathryoid extension or known distant metastases).

  • K-TIRADS 2: if a spongiform nodule is ≥2 cm FNA is recommended
  • K-TIRADS 3: ≥1.5 cm
  • K-TIRADS 4: ≥1 cm
  • K-TIRADS 5: ≥1 cm

The overall risk of malignancy in the original 2016 paper was 0% for K-TIRADS 2, 7.8% for K-TIRADS 3, 25.4% for K-TIRADS 4 and 79.3% for K-TIRADS 5.

Sensitivity for malignancy for each stage is:

  • K-TIRADS 2: 0%
  • K-TIRADS 3: 19.2%
  • K-TIRADS 4: 29.5%
  • K-TIRADS 5: 51.3%

The Korean system was published in 2016, shortly followed by ACR TI-RADS (2017) and EU-TIRADS (2017). Other systems including the ATA and SRU are also in regular use.

A number of comparative studies of the systems 3-5 show the KSThR system has high sensitivity (71-95%) but variable-to-poor specificity (53-67%). The threshold for FNA / sampling is lower than most other systems, which results in high pick up rates but also consistently high unnecessary FNA rates, with only a 41% accuracy 4.

Excellent interobserver agreement regarding decision to biopsy has been shown (Cohen kappa 0.82) 6.

K-TIRADS also describes the assessment of cervical lymph nodes which can be interrogated at the same time as thyroid lesions, both on ultrasound and CT.

Benign features include a normal hilar fatty echo and vascular pattern. Intermediate features include the loss of normal hilar structures and suspicious nodes show cystic/necrotic change, micro or macrocalcification, hyperechogenicity and abnormal peripheral or diffuse vasculature.

FNA is recommended for suspicious nodes >3 mm and intermediate nodes >5 mm (short diameter).

Please note: This article (created May 2019) is based on a proposed classification system, supported by the KSThR, which may help characterization of thyroid lesions. It may be used for risk stratification on the sole discretion of personnel or institution.

Thyroid pathology
Share article

Article information

rID: 68338
System: Head & Neck
Synonyms or Alternate Spellings:
  • K-TI-RADS
  • Korean TI-RADS
  • Korean TIRADS
  • KSThR-TI-RADS
  • KSThR-TIRADS

Support Radiopaedia and see fewer ads

Updating… Please wait.

 Unable to process the form. Check for errors and try again.

 Thank you for updating your details.