Multicenter evaluation of abbreviated MRI and ultrasound for detecting early-stage hepatocellular carcinoma

Background & aims

Abbreviated MRI (AMRI) has been proposed as an alternative to ultrasound for hepatocellular carcinoma (HCC) surveillance; however, comparative data for AMRI and ultrasound are needed. Thus, we evaluated the sensitivity and specificity of dynamic contrast-enhanced (DCE)-AMRI and ultrasound for early-stage HCC detection in patients with cirrhosis.

Methods

We conducted a multicenter retrospective case–control study among patients with cirrhosis (cases with early-stage HCC as per Milan Criteria; controls without HCC) who underwent an ultrasound and a DCE-MRI within a 6-month period between 2012 and 2019. HCC diagnosis was confirmed by imaging alone in 85% and by histopathology in 15% of patients. Dynamic AMRI examinations were simulated from the full MRI by selecting relevant sequences. Independent, blinded interpretations of ultrasounds and AMRI results were performed using Liver Imaging Reporting and Data System algorithms. Ultrasounds were considered positive if US-3 observations were detected. AMRI was considered positive if LR-4, LR-5, or LR-M were detected. Per-patient sensitivity and specificity for early-stage HCC detection were estimated, and cross-modality differences were tested.

Results

We included 216 cases and 432 controls. Patient-level sensitivity and specificity of AMRI were significantly higher compared with ultrasound: 80.1% (95% CI 76.1–83.6) vs. 71.1% (95% CI 66.6–75.2), p <0.001, and 91.9% (95% CI 89.9–93.5) vs. 72.3% (95% CI 69.3–75.2), p <0.001, respectively. AMRI sensitivity was significantly higher compared with ultrasound among patients with Child-Pugh B cirrhosis (80.8% vs. 57.4%, p <0.001) but not among those with Child-Pugh A (84.7% vs. 78.6%, p = 0.07) or Child-Pugh C cirrhosis (52.6% vs. 68.4%, p = 0.18).

Conclusions

Dynamic AMRI may be more sensitive and specific for early-stage HCC detection in patients with cirrhosis compared with ultrasound, although its relative benefit might be smaller in patients with Child-Pugh A cirrhosis. Larger direct comparative data sets are needed, particularly among patients with Child-Pugh C cirrhosis who may benefit from alternative surveillance strategies.

Impact and implications

Abbreviated MRI (AMRI) is increasingly recognized as an alternative to ultrasound for hepatocellular carcinoma (HCC) surveillance. However, existing data are limited by single-center samples, spectrum bias, and lack of comparative data for AMRI vs. ultrasound. We found that AMRI had significantly higher per-patient sensitivity and specificity compared with ultrasound for the detection of early-stage HCC, although its relative benefit might be smaller in patients with Child-Pugh A cirrhosis, and both modalities underperformed in patients with Child-Pugh C cirrhosis. If sufficiently validated, AMRI could be adopted into practice guidelines for HCC surveillance and serve as a preferred alternative in select subgroups of patients.