Thyroid imaging reporting and data system (TI-RADS)

Last revised by Dr Yuranga Weerakkody on 06 Nov 2020

Thyroid imaging reporting and data system (TI-RADS) refers to any of several risk stratification systems for thyroid lesions, usually based on ultrasound features, with a structure modelled off BI-RADS. The following article describes the initial iterations proposed by individual research groups, none of which gained widespread use. Refer to separate articles for the latest systems supported by various professional societies:

A TI-RADS was first proposed by Horvath et al. in 2009 1. The system has fair interobserver agreement 4.

  • TI-RADS 1: normal thyroid gland
  • TI-RADS 2: benign conditions (0% risk of malignancy)
  • TI-RADS 3: probably benign nodules (<5% malignancy)
  • TI-RADS 4: suspicious nodules (5-80% malignancy)
    • TI-RADS 4a: undetermined (5-10% malignancy)
    • TI-RADS 4b: suspicious (10-80% malignancy)
  • TI-RADS 5: probably malignant nodules (>80% malignancy)
  • TI-RADS 6: biopsy proven malignancy

Thyroid nodules could be classified into one of 10 ultrasound patterns, which had a corresponding TI-RADS category.

  • colloid type 1: anechoic with hyperechoic spots, nonvascularised
  • colloid type 2: mixed echogenicity with hyperechoic spots, nonexpansile, nonencapsulated, vascularized, spongiform/"grid" aspect
  • colloid type 3: mixed echogenicity or isoechoic with hyperechoic spots and solid portion, expansile, nonencapsulated, vascularized
  • simple neoplastic pattern: solid or mixed hyperechoic, isoechoic, or hypoechoic; encapsulated with a thin capsule
  • de Quervain pattern: hypoechoic, ill-defined borders, without calcifications
  • suspicious neoplastic pattern: hyperechoic, isoechoic, or hypoechoic; encapsulated with a thick capsule; hypervascularised; with calcifications (coarse or microcalcifications)
  • malignant pattern A: hypoechoic, nonencapsulated with irregular margins, penetrating vessels
  • malignant pattern B: isoechoic or hypoechoic, nonencapsulated, hypervascularised, multiple peripheral microcalcifications
  • malignancy pattern C: mixed echogenicity or isoechoic without hyperechoic spots, nonencapsulated, hypervascularised

The authors stated that TI-RADS 4 and 5 nodules must be biopsied.

In 2009, Park et al. proposed a system with five categories, which, like BI-RADS, each carried a management recommendation 2. The risk of malignancy was derived from thyroid ultrasound (TUS) features.

  • TUS 0: no nodule
  • TUS 1: highly suggestive of benign (0-7% risk of malignancy)
  • TUS 2: probably benign (8-23% malignancy)
  • TUS 3: indeterminate (24-50% malignancy)
  • TUS 4: probably malignancy (51-90% malignancy)
  • TUS 5: high suggestive of malignancy (91-100% malignancy)

The probability of malignancy was based on an equation derived from 12 features 2.

The equation was as follows: z = -2.862 + 0.581X1 - 0.481X2 - 1.435X3 + 1.178X4 + 1.405X5 + 0.700X6 + 0.460X7 + 0.648X8 - 1.715X9 + 0.463X10 + 1.964X11 + 1.739X12. Each variable is valued at 1 for the presence of the following and 0 otherwise:

  • X1: shape taller than wide
  • X2: perinodular halo
  • X3: well circumscribed
  • X4: microlobulation
  • X5: infiltrative margin
  • X6: marked hypoechoic
  • X7: hypoechoic
  • X8: homogeneous echotexture
  • X9: mainly cystic
  • X10: solid
  • X11: microcalcification
  • X12: abnormal lymph node
  • TUS 1: no additional ultrasound is recommended if clinically not needed
  • TUS 2: long term ultrasound follow up if clinically needed
  • TUS 3: fine needle aspiration biopsy and short term (6 month) follow up if nondiagnostic cytological result
  • TUS 4: fine needle aspiration biopsy and immediate reaspiration if nondiagnostic cytological result
  • TUS 5: consider surgery regardless of fine needle aspiration biopsy result

The above systems were difficult to apply clinically due to their complexity, leading Kwak et al. to propose a simpler TI-RADS in 2011 2. The system is sometimes referred to as TI-RADS Kwak 6.

The category definitions were similar to BI-RADS, based on the risk of malignancy depending on the presence of suspicious ultrasound features:

  • TI-RADS 1: negative (0% risk of malignancy)
  • TI-RADS 2: benign (0% malignancy)
  • TI-RADS 3: probably benign (1.7% malignancy)
    • no suspicious features on ultrasound
  • TI-RADS 4a: low suspicion for malignancy (3.3% malignancy)
    • one suspicious feature
  • TI-RADS 4b: intermediate suspicion for malignancy (9.2% malignancy)
    • two suspicious features
  • TI-RADS 4c: moderate concern but not classic for malignancy (44.4-72.4% malignancy)
    • three or four suspicious features
  • TI-RADS 5: highly suggestive of malignancy (>87.5% malignancy)
    • five suspicious features

The following features were considered suspicious:

  • solid component
  • hypoechogenicity, especially marked hypoechogenicity
  • microlobulated or irregular margins
  • microcalcifications
  • taller than wide shape

The study included only nodules ≥1 cm in greatest dimension. The authors suggested, as with BI-RADS, that biopsy candidates were those nodules categorized as TI-RADS category 4 or 5, meaning demonstrating at least one suspicious sonographic feature.

In 2013, Russ et al. published a simplified TI-RADS that was prospectively validated 5. The system is sometimes referred to as TI-RADS French 6.

  • TI-RADS 1: normal thyroid gland
  • TI-RADS 2: benign nodule
  • TI-RADS 3: highly probable benign nodule
  • TI-RADS 4a: low suspicion for malignancy
  • TI-RADS 4b: high suspicion for malignancy
  • TI-RADS 5: malignant nodule with more than two criteria of high suspicion

Constantly benign patterns include

  • simple cyst
  • spongiform nodule
  • "white knight" pattern in the setting of thyroiditis (numerous hyperechoic round pseudonodules with no halo or central vascularizaton)
  • isolated macrocalcification
  • nodular hyperplasia (isoechoic confluent micronodules located within the inferior and posterior portion of one or two lobes, usually avascular and seen in simple goiters)

Very probably benign nodules are those that are both

  • isoechoic or hyperechoic, and
  • no sign of high suspicion (regular shape and borders, no microcalcifications)

Mildly suspect nodules are both

  • mildly hypoechoic, and
  • no sign of high suspicion

Highly suspicious features include

  • taller than wide shape
  • irregular borders
  • microcalcifications
  • markedly hypoechoic
  • high stiffness with sonoelastography (if available)

TI-RADS 4b applies to the lesion with one or two of the above signs and no metastatic lymph node is present.

TI-RADS 4c applies to the lesion with three to five of the above signs and/or a metastatic lymph node is present.

The authors proposed the following criteria, based on French Endocrine Society guidelines, for when to proceed with fine needle aspiration biopsy:

  • if >7 mm, biopsy is recommended if TI-RADS 4b and 5 or if patient has risk factors (family history of thyroid cancer or childhood neck irradiation)
  • if >10 mm, biopsy is recommended if TI-RADS 4a or if TI-RADS 3 that has definitely grown (≥2 mm in two dimensions and >20% in volume)

ADVERTISEMENT: Supporters see fewer/no ads

Updating… Please wait.

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

 Thank you for updating your details.