RPS 301b - The role of imaging in hepatocellular carcinoma (HCC) management

Purpose:

The purpose of this study was to explore if a short MRI surveillance protocol, the SMS protocol, can be used as surveillance examination for detection of hepatocellular carcinoma.

Methods and materials:

Between 2010 and 2018, all patients who underwent yearly CE-MRI because of US screening failure as surveillance for HCC, with at least 2 examinations, were included. Second last negative MRI examination or MRI examination with HCC detected on full CE-MRI protocol, were selected. Axial T2 weighted images, axial DWI (b=600), and axial T1-in and opposed phase images were sent to a separated server. Three independent reviewers with 18, 13 and 5 years' experience in abdominal imaging evaluated these series. The presence of HCC was scored using a three-point scale as definitely HCC, uncertain, and no HCC. Reference standard was the full CE-MRI according to the international guidelines of the AASLD and EASL.

Results:

215 patients were included in our cohort follow-up study. 39 patients developed HCC (18,1%), of which 37 were detected (94,9%) by the SMS protocol. 2 HCCs were missed (0,9%). 31 patients (14,4%) had suspect lesions, which proved to be false positive. Interobserver agreement was high, the area under the curve was 0.909.

Conclusion:

Surveillance of HCC using the SMS protocol has an excellent accuracy with high interobserver agreement. The SMS protocol might be a promising technique to replace US as screening tool for surveillance of HCC.

Limitations:

Retrospective analysis. No direct comparison with US.

Ethics committee approval

Our local medical ethical committee granted permission for the study and informed consent was waived.

Funding:

No funding was received for this work.

Purpose:

To evaluate per-patient diagnostic performance of a minimised non-contrast MRI protocol for hepatocellular carcinoma (HCC) surveillance in cirrhotic liver, as well as factors affecting diagnostic sensitivity.

Methods and materials:

A total of 226 patients who underwent MRI for HCC surveillance over an eight year period were included in this retrospective study. Set1 consisted of diffusion-weighted imaging and respiratory-triggered, fast-spin echo T2-weighted imaging with fat suppression. Set2 included T1-weighted in/opposed-phase images added to images from Set1. Image sets were scored as positive or negative for HCC according to predetermined criteria by two readers independently. The diagnostic performance of the two sets in conjunction with α-fetoprotein (AFP) was assessed and compared using the McNemar test.

Results:

The sensitivity, specificity, and accuracy of Set1 of reader1/reader2 were 84.4%/87.3%, 86.8%/86.8%, and 85.0%/87.2%, respectively. Those for Set2 were 87.3%/89.6%, 81.1%/79.2%, and 85.8/87.2%, respectively. The sensitivities of the sets were not significantly different (p=0.063). Sensitivities of both sets in conjunction with AFP were higher than those of MRI alone without statistical significance (87.3%/89.6% and, p=0.063/>0.99; 89.6%/89.6%, p=0.125/>0.99). In very early stage HCC, the sensitivities of Set1 and Set2 were 73.1%/76.9% and 76.9%/82.7%, respectively.

Conclusion:

An abbreviated non-contrast MRI protocol consisting of Fat-sat T2WI and DWI is highly sensitive and may be a viable method for HCC surveillance of the cirrhotic liver. The inclusion of T1-weighted in/opposed-phase and AFP may increase this sensitivity.

Limitations:

This was a retrospective study.

Ethics committee approval

Our institutional review board approved this retrospective study; informed consent was waived.

Funding:

No funding was received for this work.

Purpose:

Hepatocellular carcinoma (HCC) is the most common primary liver cancer. Many guidelines recommend surveillance with ultrasound twice per year for patients at risk (e.g. with hepatitis or cirrhosis). However, ultrasound has limited sensitivity for detecting early HCC. We present radiomics based on MRI of the liver parenchyma as an alternative.

Methods and materials:

In our center, high risk patients deemed unsuitable for ultrasound surveillance for HCC had undergone yearly liver MRI. As a proof-of-concept, we aimed to distinguish two extrema: livers in which no HCC developed (i.e. HCC-naive livers) and livers with HCC at first detection. To this end, T2W and diffusion-weighted MRI (B-values: 0/50 and 600+) were collected from 154 patients (74 HCC, 80 HCC-naive). A clinician delineated the liver parenchyma, excluding HCC if present. Within the liver, 410 radiomics features, excluding shape and volume measures, were extracted per sequence. Three decision models, using 1) T2W; 2) DWI; and 3) T2W+DWI, were created through an automated search amongst a variety of machine learning algorithms to find the combination that maximises performance. Evaluation was implemented through a 100x random-split cross-validation, with 80% of the data used for training and model optimisation, and 20% for testing.

Results:

The T2W+DWI radiomics model had a mean area under the curve (AUC) of 0.77. The models using only the T2W or DWI faired similar (AUC of 0.74 and 0.74, respectively).

Conclusion:

Our radiomics approach showed that livers in which a HCC has developed differ in appearance from HCC-naive livers. Our next step includes the extension to longitudinal analysis of the liver to predict HCC development at an earlier stage.

Limitations:

First step towards ultimate aim: prediction in longitudinal setting.

Ethics committee approval

Erasmus MC IRB (MEC-2018-1621)

Funding:

NWO #14929-14930

Purpose:

We sought to compare the diagnostic accuracy of CT-liver perfusion (CTLP) plus MR imaging versus MR imaging alone for detection and characterisation of suspected HCC lesions.

Methods and materials:

33 patients (31 male, 25 cirrhotic) under HCC surveillance underwent Gadoxetic-Acid-Enhanced (GaE) MRI and CTLP in a 6-week interval. In total, 88 pairs of CTLP and GaE-MRI examinations were studied (17 pre-treatment, 67 post-TACE, 4 post-ablation). GaE-MRI was performed on a 1.5T-system (Siemens Vision-Hybrid). Lesions were characterised on MRI using established criteria. CTLP studies were performed on a 128-CT-system (GE Revolution HD) with 140mm z-axis coverage. Lesions with mean value greater than 1.4 HU/sec and 35 ml/100g/min on Mean-Slope-of-Increase and Hepatic-Arterial-Blood-Flow parametric maps respectively were considered as HCC on CTLP. Diagnoses based on GaE-MRI alone and GaE-MRI plus CTLP, were compared with DSA-angiography and 6-month follow-up as gold standard.

Results:

Of the total identified 282 lesions (median diameter 15mm, range 5-110mm), 111 were considered as viable HCCs. GaE-MRI identified 98 true-positive, 157 true-negative lesions, and misdiagnosed 14 false-positive and 13 false-negative lesions, providing 88.3% sensitivity (95%CI; 80.8%-93.6%) and 91.8% specificity (95%CI; 86.6%-95.4%). The combination of GaE-MRI and CTLP identified 102 true-positive, 166 true-negative, 5 false-positive, and 9 false-negative lesions, increasing sensitivity and specificity to 91.9% (95%CI 85.1%-96.2%) and 97% (95%CI 93.3%-99%), respectively.

Conclusion:

The combination of MRI and CTLP may increase accuracy of hepatic nodule characterisation, enabling more efficient patient selection for early and individualised loco-regional treatment.

Limitations:

Use of single manufacturer’s software platform and absence of histological verification can be considered the main limitations of this study.

Ethics committee approval

Institutional Ethics Committee approved the study. All patients provided informed consent.

Funding:

No funding was received for this work.

Purpose:

The American Association for Study of Liver Diseases diagnostic criteria allows non-invasive diagnosis of hepatocellular cancer (HCC) by imaging on basis of arterial phase enhancement followed by washout in portal venous phase. Portal vein thrombosis (PVT) is a common finding in the setting of cirrhosis where venous flow within the liver is altered. Also in our experience, presence of PVT leads to high incidence of atypical arterial enhancement, which may be secondary to compensatory increased arterial supply to the background liver. Our aim is to determine incidence of atypical enhancement pattern in HCCs with PVT as it remains a potential cause for delayed imaging diagnosis.

Methods and materials:

300 patients with HCC and portal vein thrombosis who underwent pretreatment multiphasic CT imaging were drawn retrospectively from our database. Arterial, portal venous, and delayed phase images were assessed qualitatively for lesion hypervascularity and washout by two independent consultant radiologists. Arterioportal shunting was also documented if present. Results were compiled using spss v21, results and p value determined.

Results:

201 lesions (66.8%) lacked typical arterial phase enhancement, however lesion washout was present, while 99 lesions (32.9%) showed typical arterial enhancement with washout (P< 0.001). Tumour thrombus was seen in 263 lesions (87.4%) while bland thrombus documented in 37 lesions (12.3%). Lesions with tumour thrombus showed high incidence of atypical lesion enhancement as compared to bland thrombus 76.4% (P< 0.001).

Arterioportal shunting was seen in 188 lesions (62.5%) (P< 0.001).

Conclusion:

A large proportion of HCC with PVT lack characteristic arterial phase enhancement. This should be given a consideration when making imaging diagnosis of HCC.

Limitations:

Relatively small sample size.

Ethics committee approval

Approved by institutional review board.

Funding:

No funding was recieved for this work.

Purpose:

The aims of this study were to investigate the cumulative incidence of cirrhosis and HCC and to identify specific features distinguishing HCC from benign arterial phase hyper-enhancing (APHE) nodules that developed after the Fontan operation.

Methods and materials:

We retrospectively enrolled 313 post-Fontan patients who had been followed for more than 5 years and had undergone ultrasound or computed tomography (CT) of the liver between January 2000 and August 2018. Cirrhosis was diagnosed clinically.

Results:

During the 5,882.7 person-years follow-up, the estimated cumulative incidence rates of cirrhosis at 5-, 10-, 20-, and 30-years after the Fontan operation were 1.3%, 9.2%, 56.6%, and 97.9%, respectively. On multiphasic CT, 18 patients had APHE nodules ≥1 cm showing washout in the portal venous phase (PVP)/delayed phase, which met current non-invasive HCC diagnosis criteria. Among them only 7 patients (38.9%, 7/18) were diagnosed with HCC. Thus, the annual incidence of HCC was 0.12% during 5,875.9 person-years of follow-up after the Fontan operation. After cirrhosis developed, the annual incidence of HCC was greater: 1.04% (7 HCCs per 672.7 person-years). The appearance of washout in the PVP (P=0.006), the long time elapsed since the initial Fontan operation (P=0.04), the large diameter of nodule (P=0.03), and elevated AFP level (P<0.001) were significantly more common in patients with HCC than those with a benign APHE nodule.

Conclusion:

In post-Fontan patients, cirrhosis is a frequent late complication, especially after 10 years, and HCC is not a rare complication after cirrhosis development. Diagnosis of HCC should not be based solely on the current imaging criteria.

Limitations:

Retrospective design with potential selection bias.

Ethics committee approval

This study was approved by the Institutional Review Board of Seoul National University Hospital (IRB No. 1812-158-998).

Funding:

No funding was received for this work.

Purpose:

This study aimed to evaluate the impact of sarcopenia and skeletal muscle mass loss in the first year after TACE on complete response (CR) and overall survival (OS) in cirrhotic patients with HCC.

Methods and materials:

In a retrospective trial conforming to the Helsinki Declaration, we selected 44 cirrhotic patients treated with TACE between 2014 to 2018 as first therapy for early or intermediate HCC. We divided patients in two groups on the basis of complete or no complete response to treatment (CR/NCR, mRECIST Criteria). We calculated L3-Skeletal Muscle Index (L3SMI) on two CT examinations (before treatment (pre-L3SMI) and after 12 months) and its variation in this period (ΔL3SMI). We also compared OS in patient with ΔL3SMI<0 and ΔL3SMI≥0 and the CR rate between sarcopenic (L3SMI<39 cm²/m² for female, <50 cm²/m² for male) and no sarcopenic patients.

Results:

The two groups were homogeneous in terms of sex, age, BMI, smoke and alcohol habits, Child-Pugh Score, and radiologic appearance lesion (nodular, non-nodular, infiltrative). Univariate analysis demonstrated no significant differences in pre-L3SMI in the two groups (52.8 cm²/m² vs 52.2 cm²/m²; p=0,82) and in ΔL3SMI (-3,57% NCR vs -1,33% CR; p=0,62). Patients with ΔL3SMI<0 and ≥0 had a comparable OS (OS 3-year 13% vs 90%; p=0,17). No differences were highlighted between sarcopenic patients (n=13) and no sarcopenic patients in terms of CR rate (30% vs 58%; p=0,18).

Conclusion:

We observed a lower reduction of muscle mass in patients with CR, a better survival of patients with increased or stable muscle mass, and a higher response rate in no-sarcopenic patients, but without reaching statistical significance. For these reasons, we need further studies.

Limitations:

The cohort was small.

Ethics committee approval

n/a

Funding:

No funding was received for this work.