Imaging Biomarkers of Hepatic Fibrosis: Reliability and Accuracy of Hepatic Periportal Space Widening and Other Morphologic Features on MRI

Noninvasive diagnosis of liver cirrhosis: qualitative and quantitative imaging biomarkers

A diagnosis of cirrhosis initiates a shift in the management of chronic liver disease and affects the diagnostic workflow and treatment decision of primary liver cancer. Liver biopsy remains the gold standard for cirrhosis diagnosis, but it is invasive and susceptible to sampling bias and observer variability. Various qualitative and quantitative imaging biomarkers based on ultrasound, CT and MRI have been proposed for noninvasive diagnosis of cirrhosis. Qualitative imaging features are easy to apply but have moderate diagnostic sensitivity. Elastography techniques allow quantitative assessment of liver stiffness and are highly accurate for cirrhosis diagnosis. Ultrasound elastography are widely used in clinical practice, while MR elastography has narrower availability. Although not applicable in clinical practice yet, other quantitative imaging features, including liver surface nodularity, linear and volumetric measurement, extracellular volume fraction, liver enhancement on hepatobiliary phase, and parameters derived from diffusion-weighted imaging, can provide additional information of liver morphology, perfusion, and function, thus may increase diagnosis performance. The introduction of radiomics and deep learning has further improved diagnostic accuracy while reducing subjectivity. Several imaging features may also help to assess liver function and outcomes in patients with cirrhosis. In this review, we summarize the qualitative and quantitative imaging biomarkers for noninvasive cirrhosis diagnosis, and the assessment of liver function and outcomes, and discuss the challenges and future directions in this field.

Noninvasive prediction of insufficient biochemical response after ursodeoxycholic acid treatment in patients with primary biliary cholangitis based on pretreatment nonenhanced MRI

OBJECTIVES

To explore the feasibility of pretreatment nonenhanced magnetic resonance imaging (MRI) in predicting insufficient biochemical response to ursodeoxycholic acid (UDCA) in patients with primary biliary cholangitis (PBC).

METHODS

From January 2009 to April 2022, consecutive PBC patients who were treated with UDCA and underwent nonenhanced MRI within 30 days before treatment were retrospectively enrolled. All MR images were independently evaluated by two blinded radiologists. Uni- and multivariable logistic regression analyses were performed to develop a predictive model for 12-month insufficient biochemical response. Model performances were evaluated by computing the area under the receiver operating characteristic curve (AUC), sensitivity, and specificity.

RESULTS

A total of 74 patients (50.6 ± 11.9 years; 62 females) were included. Three pretreatment MRI features, including hepatomegaly (odds ratio [OR]: 4.580; p = 0.011), periportal hyperintensity on T2-weighted imaging (T2WI) (OR: 4.795, p = 0.008), and narrowing of the bile ducts (OR: 3.491; p = 0.027) were associated with 12-month insufficient biochemical response in the multivariable analysis. A predictive model based on the above indicators had an AUC of 0.781, sensitivity of 85.4%, and specificity of 61.5% for predicting insufficient biochemical response.

CONCLUSIONS

A noninvasive model based on three pretreatment MRI features could accurately predict 12-month insufficient biochemical response to UDCA in patients with PBC. Early identification of PBC patients at increased risk for insufficient response can facilitate the timely initiation of additional treatment.

CLINICAL RELEVANCE STATEMENT

A noninvasive predictive model constructed by incorporating three pretreatment MRI features may help identify patients with primary biliary cholangitis at high risk of insufficient biochemical response to ursodeoxycholic acid and facilitate the timely initiation of additional treatment.

KEY POINTS

• Noninvasive imaging features based on nonenhanced pretreatment MRI may predict an insufficient biochemical response to UDCA in PBC patients. • A combined model based on three MRI features (hepatomegaly, periportal hyperintensity on T2-weighted imaging, and narrowing of the bile ducts) further improved the predictive efficacy for an insufficient biochemical response to UDCA in PBC patients, with high sensitivity and specificity. • The nomogram of the combined model showed good calibration and predictive efficacy for an insufficient biochemical response to UDCA in PBC patients. In particular, the calibration curve visualised the clinical applicability of the prediction model.

Imaging Phenotypes and Evolution of Hepatic Langerhans Cell Histiocytosis on CT/MRI: A Retrospective Study of Clinical Cases and Literature Review

(1) Background: pathological changes in hepatic Langerhans cell histiocytosis (LCH) have been observed; however, corresponding imaging findings can appear vague to physicians and radiologists. The present study aimed to comprehensively illustrate the imaging findings of hepatic LCH and to investigate the evolution of LCH-associated lesions. (2) Methods: LCH patients with liver involvement treated at our institution were retrospectively reviewed along with prior studies in PubMed. Initial and follow-up computed tomography (CT) and magnetic resonance imaging (MRI) were systematically reviewed, and three imaging phenotypes were created based on the lesion distribution pattern. Clinical features and prognoses were compared among the three phenotypes. Liver fibrosis was evaluated visually on T2-weighted imaging (T2WI) and diffusion-weighted imaging (DWI), and apparent diffusion coefficient (ADC) values of the fibrotic areas were measured. Descriptive statistics and a comparative analysis were used to analyze the data. (3) Results: based on the lesion distribution pattern on CT/MRI scans, patients with liver involvement were categorized as the disseminated lesion phenotype, scattered lesion phenotype, and central periportal lesion phenotype. Patients with scattered lesion phenotype were typically adults, and only a few of them had hepatomegaly (n_present = 1, 1/6, 16.7%) and liver biochemical abnormalities (n_present = 2, 2/6, 33.3%); patients with central periportal lesion phenotype were typically young children, and hepatomegaly and biochemical abnormalities were more apparent in these patients than those with another phenotype; and those with the disseminated lesion phenotype were found in all age groups, and the lesions evolved rapidly on medical imaging. Follow-up MRI scans show more details and can better document the evolution of lesions than CT. T2-hypointense fibrotic changes, including the periportal halo sign (n_present = 2, 2/9, 22.2%), patchy liver parenchyma changes (n_present = 6, 6/9, 66.7%), and giant hepatic nodules close to the central portal vein (n_present = 1, 1/9, 11.1%), were found, while fibrotic changes were not observed in patients with the scattered lesion phenotype. The mean ADC value for the area of liver fibrosis in each patient was lower than the optimal cutoff for significant fibrosis (METAVIR Fibrosis Stage ≥ 2) in a previous study that assessed liver fibrosis in chronic viral hepatitis. (4) Conclusions: The infiltrative lesions and liver fibrosis of hepatic LCH can be well characterized on MRI scans with DWI. The evolution of these lesions was well demonstrated on follow-up MRI scans.

Computed tomography imaging features to evaluate the severity of portal hypertension and predict the rebleeding risk after endoscopic treatment in cirrhotic patients with variceal hemorrhage

PURPOSE

To investigate the association of computed tomography (CT) imaging features and severity of portal hypertension (PH) and develop a nomogram to predict high-risk PH in cirrhotic patients with gastroesophageal variceal hemorrhage (GVH).

METHODS

The study retrospectively enrolled 158 cirrhotic patients with a history of endoscopic treatment for GVH. Hepatic vein pressure gradient (HVPG) was measured and the patients were classified into high-risk (HVPG > 16 mmHg) or low-risk (HVPG ≤ 16 mmHg) PH group. Pre-treatment CT features, including cavernous transformation of portal vein (CTPV), hilar periportal space (a distance between right portal vein and posterior edge of segment IV of the liver), and depth of right posterior hepatic notch sign (a sharp indentation in the right medial posterior liver surface), were evaluated. Risk factors associated with high-risk PH were analyzed, and a nomogram based on the imaging features was developed.

RESULTS

High-risk PH group showed a higher rebleeding rate after treatment than that of the low-risk (P = 0.029). Multivariate analysis indicated that larger hilar periportal space (P < 0.001), less frequencies of CTPV (P = 0.044) and deeper right posterior hepatic notch (P < 0.001) were independent risk factors associated with high-risk PH. A nomogram based on the three CT imaging features was established to predict high-risk PH with an excellent discrimination (c-statistic 0.854).

CONCLUSION

The nomogram based on CT features of hilar periportal space, depth of right posterior hepatic notch and CTPV can help to distinguish cirrhotic patients with high-risk PH, who are more vulnerable of variceal rebleeding after endoscopic treatment.

Quality Control of Magnetic Resonance Elastography Using Percent Measurable Liver Volume Estimation

BACKGROUND

Although studies have described factors associated with failed magnetic resonance elastography (MRE), little is known about what factors influence usable elastography data.

PURPOSE

To identify factors that have a negative impact on percent measurable liver volume (pMLV), defined as the proportion of usable liver elastography data relative to the volume of imaged liver in patients undergoing MRE.

STUDY TYPE

Retrospective.

SUBJECTS

A total of 264 patients (n = 132 males, n = 132 females; mean age = 57 years) with suspected or known chronic liver disease underwent MRE paired with a liver protocol MRI.

FIELD STRENGTH/SEQUENCE

MRE was performed on a single 1.5 T scanner using a two-dimensional gradient-recalled echo phase-contrast sequence with a passive acoustic driver overlying the right hemiliver.

ASSESSMENT

Stiffness maps (usable data at 95% confidence) and liver contours on magnitude images of the MRE acquisition were manually traced and used to assess mean stiffness and pMLV. Hepatic fat fraction and R² * values were also calculated. The distance from the acoustic wave generator on the skin surface to the liver edge was measured. Two radiologists performed the MR analyses with 50 overlapping cases for inter-reader analysis.

STATISTICAL TESTS

Linear regression was performed to identify factors significantly associated with pMLV. Intraclass correlation was performed for inter-reader reliability.

RESULTS

pMLV was 31% ± 20% (range 0%-86%). Complete MRE failure (i.e. pMLV = 0%) occurred in 10 patients (4%). Multivariate linear regression identified higher hepatic fat fraction, R² *, BMI, and driver-to-liver surface distance; male sex; and lower mean liver stiffness was significantly independently associated with lower pMLV. Intraclass correlation for pMLV was 0.96, suggestive of excellent reliability.

DATA CONCLUSION

Higher fat fraction, R² *, BMI, driver-to-liver surface distance, male sex, and lower mean liver stiffness were associated with lower pMLV. Optimization of image acquisition parameters and driver placement may improve MRE quality, and pMLV likely serves as a diagnostic utility quality control metric.

LEVEL OF EVIDENCE

3 TECHNICAL EFFICACY STAGE: 2.