All Pathways
RadiologyDiagnostic

Liver Lesion Characterization (LI-RADS v2018)

Liver Lesion Characterization (LI-RADS v2018): Untreated liver observation on multiphase CT/MRI → Do not use LI-RADS: under 18, vascular or congenital c...

Pathway Overview

21 steps

Algorithm Steps

21 total

  1. 01Start

    Untreated liver observation on multiphase CT/MRI

    Adult at high risk for HCC. No pathology proof. CT or MRI with extracellular contrast, or MRI with hepatobiliary contrast.

  2. 02Warning

    Do not use LI-RADS: under 18, vascular or congenital cause of cirrhosis

    In these patients LI-RADS is not validated. Benign nodules can look like HCC, so an LR-5 can be a false HCC diagnosis.

    • Age under 18 years
    • Cirrhosis from congenital hepatic fibrosis, or from a vascular disorder: Budd-Chiari syndrome, HHT, chronic portal vein occlusion, cardiac congestion, diffuse nodular regenerative hyperplasia
    • Path-proven malignancy or path-proven non-hepatocellular benign lesion (for example haemangioma): report the pathology, not a LI-RADS category
  3. 03Decision

    LI-RADS applies? Adult at high risk for HCC

    Cirrhosis, chronic hepatitis B, or current or prior HCC (includes adult liver transplant candidates and recipients). None of the exclusions above.

  4. If Yes
    1. 04Decision

      Eligible: can the observation be categorised?

      Not categorisable if image degradation or a missing phase prevents it.

    2. If Yes
      1. 05Decision

        Categorisable: definite tumour in vein?

        Unequivocal enhancing soft tissue in a vein, with or without a visible mass. If unsure, it is not tumour in vein.

      2. If Yes
        1. 06End

          LR-TIV: tumour in vein

          Multidisciplinary discussion for tailored workup. May include biopsy. Report: next to a targetoid mass, 'may be due to non-HCC malignancy'; next to an LR-5 mass, 'definitely due to HCC'; otherwise 'probably due to HCC'.

        If No
        1. 07Decision

          No tumour in vein: definitely benign?

          Examples: cyst, haemangioma, perfusion alteration (for example arterioportal shunt), focal fat deposition or sparing, hypertrophic pseudomass, confluent fibrosis or focal scar, spontaneous disappearance. A nodule that looks like FNH or hepatocellular adenoma is never LR-1: usually LR-3; LR-2 only with caution.

        2. If Yes
          1. 08End

            LR-1: definitely benign

            Return to surveillance in 6 months.

          If No
          1. 09Decision

            Not definitely benign: probably benign?

            Probable forms of the benign entities above, or a distinctive solid nodule under 20 mm with no major feature and no LR-M feature, and either no ancillary feature of malignancy or ancillary features of both malignancy and benignity.

          2. If Yes
            1. 10End

              LR-2: probably benign

              Return to surveillance in 6 months. Consider repeat diagnostic imaging in 6 months or less.

            If No
            1. 11Decision

              Not benign: LR-M criteria met?

              Targetoid mass (rim APHE, peripheral washout, delayed central enhancement, or targetoid diffusion restriction, transitional or hepatobiliary phase appearance). Or nontargetoid mass (no tumour in vein, not meeting LR-5) with infiltrative appearance, marked diffusion restriction, necrosis or severe ischaemia, or another feature of non-HCC malignancy. If unsure between LR-M and LR-4 or LR-5, choose LR-M.

            2. If Yes
              1. 12End

                LR-M: probably or definitely malignant, not necessarily HCC

                Multidisciplinary discussion for tailored workup. Often includes biopsy. May be cholangiocarcinoma, combined HCC-cholangiocarcinoma, metastasis or atypical HCC.

              If No
              1. 13Action

                Not LR-M: assess major features

                Count a feature only when it is unequivocal. If unsure, count it as absent. Gadoxetate MRI: count washout only in the portal venous phase. Transitional or hepatobiliary phase hypointensity is not washout (ancillary feature only).

                • Size: outer edge to outer edge, in the phase where the edge is clearest; avoid the arterial phase
                • Nonrim arterial phase hyperenhancement (APHE)
                • Additional major features: nonperipheral washout, enhancing capsule, threshold growth
                • Threshold growth: size increase of 50% or more in 6 months or less
              2. 14Action

                Apply the CT/MRI v2018 diagnostic table

                Category depends on APHE, size and the number of additional major features. LR-5 needs nonrim APHE and size 10 mm or more.

                • No APHE: 0 features LR-3; 1 feature LR-3 if under 20 mm, LR-4 if 20 mm or more; 2 or more LR-4
                • Nonrim APHE, under 10 mm: 0 features LR-3; 1 or more LR-4
                • Nonrim APHE, 10-19 mm: 0 features LR-3; capsule only LR-4; washout only or threshold growth only LR-5; 2 or more LR-5
                • Nonrim APHE, 20 mm or more: 0 features LR-4; 1 or more LR-5
              3. 15Action

                Optional: adjust with ancillary features, then apply tiebreak

                Ancillary features can never upgrade to LR-5. Distinctive nodule under 20 mm with no major feature: the ancillary feature of malignancy (with none of benignity) that excluded LR-2 gives LR-3. Do not upgrade it again.

                • 1 or more features favouring malignancy: upgrade by 1 category, up to LR-4
                • 1 or more features favouring benignity: downgrade by 1 category
                • Features for both malignancy and benignity: do not adjust
                • If unsure between two categories, choose the one with lower certainty
              4. 16Decision

                Final category from table and adjustment

                Go to the one step below that matches the final category (LR-3, LR-4 or LR-5). Use the table category after any ancillary adjustment. If ancillary features downgrade LR-3 to LR-2: return to surveillance in 6 months; consider repeat diagnostic imaging in 6 months or less.

              5. LR-3
              6. 17End

                LR-3: intermediate probability of malignancy

                Repeat or alternative diagnostic imaging in 3 to 6 months.

              7. LR-4
              8. 18End

                LR-4: probably HCC

                Multidisciplinary discussion for tailored workup. May include biopsy.

              9. LR-5
              10. 19End

                LR-5: definitely HCC

                Nonrim APHE, size 10 mm or more and the required major features. HCC is confirmed on imaging. Multidisciplinary discussion for consensus management.

      If No
      1. 20End

        LR-NC: not categorisable (degraded or missing images)

        Repeat or alternative diagnostic imaging in 3 months or less.

    If No
    1. 21End

      Not eligible: do not assign a LI-RADS category

      No HCC risk factor, under 18, cirrhosis from congenital hepatic fibrosis or a vascular disorder, or a path-proven lesion (report the pathology). Describe the observation and discuss with the hepatology or HCC team.

Guideline Source

ACR CT/MRI LI-RADS v2018 Core (described in Chernyak et al., Radiology 2018)

Clinical Safety Information

Clinical Decision Support — Not a Substitute for Clinical Judgment

Individual patient factors may require deviation from these recommendations.

Known Limitations

  • CT/MRI LI-RADS only for adults at high risk for HCC (cirrhosis, chronic hepatitis B, current or prior HCC). Not for under 18, or cirrhosis from congenital hepatic fibrosis or a vascular disorder.
  • Does not cover treated observations (use LI-RADS CT/MRI treatment response v2024), CEUS LI-RADS or ultrasound surveillance.
  • Categories state imaging probability only. Final management needs multidisciplinary review of the whole patient.

Contraindicated Populations

Patients under 18 yearsPatients without cirrhosis, chronic hepatitis B, or current or prior HCCCirrhosis due to congenital hepatic fibrosisCirrhosis due to a vascular disorder (Budd-Chiari syndrome, hereditary haemorrhagic telangiectasia, chronic portal vein occlusion, cardiac congestion, diffuse nodular regenerative hyperplasia)Path-proven malignancy or path-proven non-hepatocellular benign lesion

Applicable Regions

USAUEUAPAC

AU: No Australian variant of LI-RADS. The ACR CT/MRI v2018 criteria apply. Follow the local HCC multidisciplinary team.

EU: EASL guidance uses its own non-invasive HCC criteria. Follow the local multidisciplinary team.

US: ACR LI-RADS CT/MRI v2018 is integrated into the AASLD 2018 HCC guidance. OPTN class 5A has adopted the LR-5 criteria.

APAC: Regional societies (for example APASL) use their own HCC imaging criteria. Follow the local multidisciplinary team.

Version 2Next review: 2027-09-30

Frequently Asked Questions

What is the Liver Lesion Characterization (LI-RADS v2018)?

The Liver Lesion Characterization (LI-RADS v2018) is a diagnostic clinical algorithm for Radiology. It provides a structured decision tree to guide clinical decision-making, based on ACR CT/MRI LI-RADS v2018 Core (described in Chernyak et al., Radiology 2018).

What guideline is the Liver Lesion Characterization (LI-RADS v2018) based on?

This algorithm is based on ACR CT/MRI LI-RADS v2018 Core (described in Chernyak et al., Radiology 2018) (DOI: 10.1148/radiol.2018181494).

What are the limitations of the Liver Lesion Characterization (LI-RADS v2018)?

Known limitations include: CT/MRI LI-RADS only for adults at high risk for HCC (cirrhosis, chronic hepatitis B, current or prior HCC). Not for under 18, or cirrhosis from congenital hepatic fibrosis or a vascular disorder.; Does not cover treated observations (use LI-RADS CT/MRI treatment response v2024), CEUS LI-RADS or ultrasound surveillance.; Categories state imaging probability only. Final management needs multidisciplinary review of the whole patient.. Individual patient factors may require deviation from these recommendations.

Get AI-Powered Analysis Alongside This Algorithm

In AttendMe.ai, the Liver Lesion Characterization (LI-RADS v2018) appears automatically when your clinical question matches — alongside evidence from 3M+ peer-reviewed articles.

Try AttendMe Free