Hepatobiliary-Phase Hypointense Nodules Without Arterial-Phase Hyperenhancement: Developing a Risk Stratification for Hypervascular Transformation Based on a Real-World Observational Cohort Study

Abstract

Purpose: To develop a risk stratification based on MRI features to predict hypervascular transformation for hepatobiliary-phase (HBP) hypointense nodules without arterial-phase hyperenhancement (APHE).

Materials and methods: This retrospective observational cohort study included 55 HBP hypointense nodules without APHE in 35 patients with chronic liver disease, cirrhosis, or current hepatocellular carcinoma (HCC) who underwent gadoxetic acid-enhanced MRI. The hypervascular transformation during the follow-up MRI(s) was the primary endpoint analyzed for the nodules. Univariable and multivariable Cox proportional hazard regression analyses were performed to identify risk features predicting transformation and assess their predictive value.

Results: Among the 55 nodules, 27 developed hypervascular transformation, while 28 did not. Diffusion-weighted imaging (DWI) hyperintensity (hazard ratio [HR], 4.98; 95% confidence interval [CI]: 1.60, 15.54; p = 0.006) and portal venous phase (PVP) hypointensity (HR, 4.08; 95% CI: 1.43, 11.64; p = 0.009) were associated with hypervascular transformation. DWI hyperintensity and PVP hypointensity had 44.4% (95% CI: 26.0%, 64.4%) and 81.9% (95% CI: 61.3%, 93.0%) sensitivity, while their specificity was 78.2% (95% CI: 64.6%, 87.8%) and 67.9 (95% CI: 47.6%, 83.4%), respectively. The specificity of the combination of two features was 100% (95% CI: 85.0%, 100%). The hypervascular transformation rates for nodules with both, either and neither of the risk MRI findings were 100% (10/10), 60.9% (14/23), and 13.6% (3/22), respectively; the median intervals for transformation were 312 (range: 73-838), 409 (range: 50-1643) and 555 (range: 423-968) days, respectively.

Conclusion: The combination of DWI hyperintensity and PVP hypointensity may be used as a high-risk indicator for the hypervascular transformation of HBP hypointense nodules without APHE; nodules without either feature may be treated as low-risk nodules and could adopt an extended interval follow-up schedule.

Keywords: cancer risk; cirrhosis; gadolinium ethoxybenzyl DTPA; hepatocellular carcinoma; magnetic resonance imaging.

Figure 1.

Flow chart of the patient's inclusion. Abbreviations: HCC: hepatocellular carcinoma; HBP: hepatobiliary phase; APHE: arterial-phase hyperenhancement; TACE: transarterial chemoembolization; RFA: radiofrequency ablation; ECA: extracellular agent; CEUS: contrast-enhanced ultrasonography.

Figure 2.

Hypervascular transformation of a hepatobiliary phase hypointense nodule without arterial phase hyperenhancement manifested with the combination of DWI hyperintensity and PVP hypointensity in a 64-year-old man with hepatitis B–induced liver cirrhosis. On initial gadoxetic acid-enhance MRI (A-D), a 5-mm nodule shows hypointensity on 20-min hepatobiliary phase imaging (A, arrow), isointensity on arterial phase T1-weighted imaging (B), hypointensity on PVP imaging (C, arrow), and hyperintensity on DWI (b = 500 s/mm²) (D, arrow). Follow-up gadoxetic acid-enhanced MRI obtained 445 days later, the nodule shows marked enlargement as measuring 18-mm, unequivocally hyperenhancement than surrounding liver on arterial phase imaging (E, arrow), and hypointensity on 20-min hepatobiliary phase imaging (F, arrow). Nodule specimens after surgical resection confirmed the diagnosis of HCC. Abbreviations: DWI: diffusion-weighted imaging; PVP: portal venous phase; HCC: hepatocellular carcinoma.

Figure 3.

Tree diagram for the nodules in the longitudinal observation. Abbreviations: HBP: hepatobiliary phase; APHE: arterial phase hyperenhancement; PVP: portal venous phase; DWI: diffusion-weighted imaging.

Figure 4.

Kaplan-Meier curves of hypervascular transformation rates according to the risk MRI findings. Hypervascular transformation rates of HBP hypointense nodules without APHE according to both (purple), either (blue), and neither (yellow) of the risk MRI features were evaluated using the Kaplan-Meier method. Nodules with both risk MRI features versus either feature versus neither of features, p < 0.001; Statistical significance was assessed with the log-rank test.