The CT/MR imaging LI-RADS v2018 core document provides technical recommendations for MR imaging of HCC. These imaging parameters are applicable for the evaluation of known or suspected HCC. Parameters for a full liver MR imaging are described later ( Fig. 1 ).

• •

  1.5 T or 3T scanner

• •

  Torso phased-array coil

• •

  Required sequences: in-and-opposed phase, T2 weighted (fat suppression is optional), and multiphase T1 weighted pre- and post-contrast:

  • ○

    Extracellular contrast agent and gadobenate dimeglumine: arterial phase (late arterial phase strongly preferred), portal venous phase, and delayed phase Optional 1–3-h hepatobiliary phase if gadobenate dimeglumine is used

  • ○

    Gadoxetate disodium contrast agent: arterial phase (late arterial phase strongly preferred), portal venous phase, transitional phase (2–5 minutes after injection), and hepatobiliary phase

  • ○

    The injection rate for both gadobenate dimeglumine and gadoxetate disodium is the same at 1 to 2 mL/sec

• •

  Additional suggested sequences : diffusion-weighted imaging, subtraction imaging (if not able to be easily generated on PACS), and multiplanar acquisition

Imaging protocol of full liver MR imaging with hepatobiliary contrast. The liver lesion ( arrows ) demonstrates non-rim arterial phase hyperenhancement on the arterial phase, and non-peripheral washout and capsule on the portal venous and translational phases. This lesion is a hepatobiliary defect on the hepatobiliary phase. The lesion is mildly T2 dark and does not have fat as seen in the in-and-opposed phase images. On the high b-value DWI image, the lesion has restricted diffusion.

To address the economic and logistical barriers of full liver MR imaging, abbreviated MR imaging (AMRI) protocols with shorter scan times have been developed. A recent meta-analysis showed promising pooled sensitivity of 86% and specificity 95% with AMRI and per-lesion sensitivity of 77% in the detection of HCC with sensitivity of abbreviated MR imaging for detection of HCC less than 2 cm lower than HCC greater than 2 cm (69% and 86%, respectively). ^, AMRI protocols can be completed in ∼10 to 15 minutes, which is considerably shorter than a standard liver MR imaging and has comparable diagnostic accuracy to full multiphase examinations. Three AMRI techniques have been explored, which vary in the included sequences and are designed to evaluate different features of HCC.

These 3 strategies include dynamic contrast-enhanced (DCE-AMRI), hepatobiliary phase (HBP-AMRI), and non-contrast (NC-AMRI) Table 2 . A retrospective study by Violi and colleagues compared these 3 AMRI types in a screening cohort and found that both DCE- and HBP-AMRI performed well, though no independent reference standard was used to validate results. While all 3 AMRI protocols appear promising, especially in terms of sensitivity and inter-reader agreement, further prospective studies in true screening populations are needed to clarify their relative diagnostic performance, reproducibility, clinical utility, and cost effectiveness.

Each AMRI protocol is tailored to the specific imaging feature of HCC being evaluated, prioritizing those with the highest diagnostic yield within the constraints of shortened scan time.

• •

  Pre-contrast and multiphase postcontrast T1 (multiphase only in DCE-AMRI) : Accesses arterial phase hyperenhancement, washout, and enhancing capsule

• •

  Hepatobiliary phase T1 (only HBP-AMRI) : Identifies lesions with reduced hepatobiliary uptake due to decreased organic anion transporting polypeptides (OATP) transporter expression in hepatocytes

• •

  In-phase and opposed-phase T1 : Identifies fat or iron content differing from the background liver. Can be acquired alongside pre-contrast T1 on some MR imaging systems

• •

  T2-weighted imaging : Identifies mild to moderate hyperintensity. Marked hyperintensity (e.g., cysts and hemangiomas) or marked hypointensity suggests benignity

• •

  Diffusion-weighted imaging ( DWI ) / Apparent diffusion coefficient ( ADC ) : Accesses diffusion restriction versus T2 shine-through

DCE-AMRI captures pre-contrast, arterial, portal venous, and delayed phase images with an extracellular contrast agent before and following IV contrast administration while eliminating T2-weighted, diffusion-weighted, and in-and-opposed phase sequences. The post-contrast images of DCE-AMRI allow for detection of hallmark features of HCC based on lesion vascularity, which are arterial phase hyperenhancement, washout, and capsular enhancement ( Fig. 2 ). Inclusion of the full dynamic contrast enhancement sequences also allows for full lesion categorization at the time of screening without the need for follow-up diagnostic imaging.

Imaging protocol of dynamic contrast-enhanced (DCE) abbreviated MRI. The DCE-AMRI is comprised of only the pre-contrast, arterial, portal venous, and delayed phases of multiphasic MRI imaging. The liver lesion ( arrows ) demonstrates non-rim arterial phase hyperenhancement on the arterial phase, and non-peripheral washout and capsule on the portal venous and delayed phases.

Retrospective studies that simulate DCE-AMRI utilizing extracted sequences from full liver MR imaging studies have found a similar diagnostic performance of DCE-AMRI compared to the full liver MR imaging protocol. ^, In a recent prospective dual-center study out of Korea with a majority of the patients having hepatitis B cirrhosis, compared to biannual US, annual DCE-AMRI had a higher diagnostic yield for HCC diagnosis with no difference in false referral rate. In a recent case-control study, DCE-AMRI showed higher sensitivity and specificity compared to US in patients with cirrhosis, with significantly higher sensitivity than US in patients with Child-Pugh B cirrhosis but not Child-Pugh A or C cirrhosis.

A weakness of DCE-AMRI is that this technique has the potential for overcalling benign perfusion alterations and/or vascular shunts due to enhancement characteristics. The absence of additional sequences for the downgrading of suspicious lesions may also result in a greater need for follow-up imaging. Additional limitations of this technique include susceptibility to motion and contrast timing issues.

HBP-AMRI focuses on the hepatobiliary phase appearance after gadoxetate disodium contrast administration, enabling the detection of lesions with impaired hepatocyte function. HCC typically appears hypointense on T1-weighted hepatobiliary phase images due to absent or reduced OATP transporter activity, which impairs contrast uptake. As a result, suspicious lesions appear hypointense relative to the normal surrounding liver parenchyma on the delayed hepatobiliary phase sequence. This approach can be useful for identifying early, small (less than 2 cm) HCC not yet demonstrating classic enhancement detectable on DCE-AMRI. HBP-AMRI has the most robust data of all the AMRI protocols, as it has been clinically implemented at several academic centers over the last 10 years. Benefits of this strategy include high diagnostic accuracy and favorable cost-effectiveness in high-risk populations. ^,,,, Multiple studies have confirmed HBP-AMRI’s favorable cost-effectiveness compared to full MR imaging and US, with one analysis showing a $3000 per quality-adjusted life year gained advantage over US due to its improved diagnostic accuracy.

The patient undergoing HBP-AMRI screening receives a hand-injected intravenous contrast agent in the waiting room and undergoes imaging of all included sequences approximately 20 minutes after injection—including hepatobiliary phase T1, T2, with or without DWI ( Fig. 3 ). DWI is typically included to maximize diagnostic accuracy ^, without substantially increasing acquisition time (approximately 2 minutes).

Imaging protocol of hepatobiliary phase (HBP) abbreviated MR imaging. The HBP-AMRI is composed only of the 20-minute hepatobiliary phase, T2 weighted-imaging, and diffusion weighted imaging. The liver lesion ( arrows ) is a defect on the hepatobiliary phase, is mildly T2 bright, and demonstrates restricted diffusion.

A key advantage of HBP-AMRI is its extended imaging window, as hepatocyte contrast uptake persists beyond 20 minutes. This allows greater flexibility for acquiring diagnostic post-contrast images compared to the narrower timing window required for DCE-AMRI, and the ability to repeat sequences if needed. In 2024, gadoxetate disodium was reclassified by the American College of Radiology as a Group II gadolinium-based contrast agent, indicating a lower risk for nephrogenic systemic fibrosis. This reclassification may slightly expand the pool of patients eligible for HBP-AMRI-based screening, particularly in patients previously excluded due to renal impairment.

Limitations for HBP-AMRI include the need for repeat diagnostic MR imaging after lesion identification in order to fully characterize an observation, reduced utility in patients with elevated bilirubin above 3 mg/dL, diminished background hepatic uptake in advanced cirrhosis, false positive findings due to nonfunctioning hepatocytes, and poor sensitivity for HCC in areas of confluent fibrosis. Moreover, HBP-AMRI is preferred in patients with normal bilirubin, which limits its applicability in patients with poor liver function. Lastly, HBP-AMRI has increased associated costs compared to DCE and NC-AMRI. It is expected, however, that shorter scan time and improved radiologist efficiency in reading the examination may offset this difference.

NC-AMRI protocol composed of in-and-opposed phase T1-weighted images, T2-weighted images, and DWI images ( Fig. 4 ). Similar to HBP-AMRI, key sequences may be repeated if initial images are degraded by artifact, as this technique does not rely on the dynamic post-contrast appearance of lesions. NC-AMRI relies on the DWI sequence to find lesions that have restricted diffusion and includes the T2-weighted sequence to evaluate for lesions that are very bright like cysts and hemangiomas that may cause T2 shine-through on the DWI sequence.

Imaging protocol of non-contrast (NC) abbreviated MR imaging. NC-AMRI is composed of precontrast T1-weighted imaging, T2-weighted imaging, in-and-opposed phase imaging, and diffusion weighted imaging. The liver lesion ( arrows ) is dark on the precontrast T1-weighted image, is mildly T2 bright, does not contain fat on the in-and-opposed phase images, and demonstrates restricted diffusion.

A recent meta-analysis showed sensitivity for NC AMRI was comparable to contrast-enhanced AMRI (86% and 94%) respectively. Other retrospective studies have evaluated NC-AMRI in HCC-enriched populations, nonsurveillance populations, with per-patient sensitivities for HCC ranging from 84% to 92%. ^,, Additionally, 2 prospective studies on NC-AMRI have evaluated the performance in a surveillance population. One prospective study in a surveillance population showed that annual NC-AMRI did not have a statistically significantly higher sensitivity for HCC than biannual US. The second study showed that NC-abbreviated MR imaging had a sensitivity of 62% in a surveillance population, but a high specificity ranging from 96% to 100%.

NC-AMRI provides the most limited ability to characterize liver lesions due to the inability to evaluate for major imaging findings of HCC as with DCE-AMRI or hypointense lesions in HBP-AMRI, but may be favored because of shorter scan times, no need for intravenous access or contrast administration, and it is the least costly among AMRI protocols. Furthermore, increased public awareness of gadolinium deposition has led to apprehension among some patients regarding the need for repeated lifetime MR imaging contrast use. NC-AMRI may be an alternative for patients when specific contrast agents supply is limited, in patients who have documented severe allergies, and in those who strongly oppose the use of gadolinium-based MR imaging contrast.

A limitation of NC-AMRI is its reliance on the DWI sequence, which is often compromised by susceptibility artifact in the upper abdomen near the dome, along the liver margin, and especially the left lobe due to adjacent cardiac motion. ^, This limitation may increase false negatives if lesions are present in artifact-prone areas. In a recent meta-analysis, NC-AMRI technique was found to have lower but acceptable sensitivity for early-stage HCC compared to HBP-AMRI (79% vs 87%, respectively). This study did however show higher specificity of NC-AMRI compared to HBP-AMRI, possibly due to lower false positive rates related to impaired hepatocyte function in some benign lesions and many NC-AMRI studies performed in HCC-enriched populations compared to most of the HBP-AMRI studies performed in surveillance or simulated surveillance populations with lower incidences of HCC.