SS 201a - Hepatocellular carcinoma (HCC): detection, characterisation and therapeutic response
SS 201a - Hepatocellular carcinoma (HCC): detection, characterisation and therapeutic responseWednesday, March 1, 10:30 - 12:00 Room: B Session Type: Scientific Session Topics: Oncologic Imaging, Abdominal Viscera Moderators: D. Ippolito (Monza/IT), S. A. Taylor (London/UK) Add session to my schedule In your schedule (remove)
Feasibility of 10-min delayed hepatocyte phase imaging using a 30° flip angle in Gd-EOB-DTPA-enhanced liver MRI for the detection of hepatocellular carcinoma in patients with cirrhosis
Purpose: To compare 10-min delayed hepatocyte phase imaging (HPI) using a 30° flip angle (FA) (10m-FA30) and 20-min delayed HPI using a 10° FA (20m-FA10) or 30° FA (20m-FA30) in Gd-EOB-DTPA-enhanced MRI in patients with chronic hepatitis or cirrhosis, in terms of contrast-to-noise ratio (CNR) for hepatocellular carcinoma (HCC) and detection sensitivity for focal hepatic lesions (FHLs).
Methods and Materials: 104 patients with 168 HCCs and 55 benign FHLs who underwent Gd-EOB-DTPA-enhanced MRI with 10m-FA30, 20m-FA10, and 20m-FA30 were enrolled. Patients were divided into two groups according to the Child-Pugh classification: group A with Child-Pugh A cirrhosis and group B with Child-Pugh B or C cirrhosis. CNR for HCCs and the detection sensitivity for FHLs were compared between 10m-FA30 and 20m-FA10 or 20m-FA30.
Results: In group A, the CNR for HCCs on 10m-FA30 (165.8±99.7) was significantly higher than that on 20m-FA10 (113.4±71.4) and lower than that of 20m-FA30 (210.2±129.3). However, there was no significant difference in the detection sensitivity between 10m-FA30 (95.0%) and 20m-FA10 (94.7%) or 20m-FA30 (94.7%). In group B, the CNR (54.0±36.4) for HCCs and the detection sensitivity (94.2%) for 10m-FA30 were significantly higher than those for 20m-FA10 (41.8±36.4 and 80.8%, respectively) and were not different from those for 20m-FA30 (62.7±44.4 and 93.3%, respectively).
Conclusion: The diagnostic performance of 10m-FA30 was similar to or higher than 20m-FA10 or 20m-FA30 in both groups A and B. This finding indicates that 10m-FA30 could replace 20-min delayed HPIs regardless of patient’s liver function and reduce the delay time by 10 minutes.
Assessment of gadoxetic acid-enhanced MRI phases for LI-RADS categorisation and non-invasive grading of hepatocellular carcinoma
Purpose: This retrospective study evaluated the impact of transitional and hepatobiliary phase on LI-RADS classification and non-invasive grading in patients with hepatocellular carcinoma (HCC).
Methods and Materials: 61 patients with 61 histologically confirmed HCCs underwent MRI with gadoxetic acid. In all phases (arterial [AP], portal venous [PVP], transitional [TP], hepatobiliary [HP]) signal intensities (SI) of HCC and adjacent liver parenchyma were measured. For quantitative assessment ratio between SI of HCC and SI of normal liver were calculated. LI-RADS classification (LR-1 to LR-5) were done in A) LI-RADS major criterions with assessment of washout criterion only in PVP, B) same as A + ancillary feature “transitional phase hypointensity” and C) same as B + ancillary feature “hepatobiliary phase hypointensity”. Histopathological HCC grading results were correlated with the SI HCC ratio (= SI mass HP / SI mass AP).
Results: None HCC was graded as LR-1 or LR-2. The inclusion of the feature “transitional phase hypointensity” led to a non-significant upgrading of 5/9 HCC from LR-3 to LR-4 (p = 0.062). The addition of the feature “hepatobiliary phase hypointensity” led to a statistical significant upgrading of 8/9 HCC from LR-3 to LR-4 (p = 0.008). None HCC with initial LR-4 was upgraded. None lesions were downgraded. The HCC SI ratios between G1 vs. G2/G3 tumours were statistically significant different (p = 0.008).
Conclusion: The addition of transitional and hepatobiliary gadoxetic acid-enhanced MRI phases improves LI-RADS categorisation. Quantitative assessment of HCC SI ratio enables non-invasive grading.
Comparison between acoustic radiation force impulse quantification data and perfusion-CT parameters in hepatocellular carcinoma
Purpose: To find out, if ultrasound elastography of hepatocellular carcinoma (HCC) can predict patterns of tumour perfusion in volume perfusion computed tomography (VPCT).
Methods and Materials: 25 consecutive patients (mean age, 68.9; range, 51 - 85 years) with liver cirrhosis suspected of HCC underwent VPCT and acoustic radiation force impulse (ARFI) elastography the same day. Quantitative elasticity values were registered, while blood flow (BF), blood volume (BV) and hepatic perfusion index (HPI) of the HCC lesions were calculate. Additionally, we identified histologic WHO grading, lesion size and localisation. The Siemens Acuson S 3000 HELX-System with Virtual Touch™-Software and Siemens Somatom Definition Flash with Syngo® software were used.
Results: A total of 43 HCC lesions were assessed. Mean shear wave velocity was 2.6 m/s (range, 1.1 - 4.3 m/s). There was no significant linear correlation between the elasticity values and BF (p=0.751), BV (p=0.426) and HPI (p=0.437). However, elasticity values were higher, the larger the tumour was (p=0.008). Shear wave velocity declined with increasing distance of the HCC to the skin surface (p=0.028) and depending on liver segment. In addition, elasticity values were higher in less differentiated HCCs. This trend was not statistically significant (p=0.842).
Conclusion: Tissue elasticity in HCC does not correlate with the degree of tumour vascularisation, but calculated values are influenced both by the tumour size and localization inside the liver.
Evolution of indeterminate hepatocellular nodules at initial Gd-EOB-DPTA-enhanced MRI in cirrhotic patients
Purpose: To retrospectively evaluate the evolution of indeterminate hepatic nodules at initial Gd-EOB-DPTA-enhanced MRI in cirrhotic patients.
Methods and Materials: 33 cirrhotic patients (24 males; mean age 64.9 years) with 69 indeterminate nodules (mean diameter 1.1 cm; range 0.5-2) at initial Gd-EOB-DPTA-enhanced MRI and a Gd-EOB-DPTA-enhanced MRI follow-up of at least 2 years (mean 805 days; range 745-2440) were evaluated. Indeterminate hepatocellular nodules were defined as nodules that cannot be diagnosed as HCC according to ASSLD 2010 criteria. Based on signal intensity at initial MRI, each nodule was classified into 6 groups: A) hyperintense on precontrast T1 phase (12/69); B) hyperintense on precontrast T1 and hepatobiliary phase (4/69); C) hyperintense on hepatobiliary phase (23/69); D) hypointense on hepatobiliary phase (9/69); E) hypointense on transitional and hepatobiliary phase (20/69); F) hypointense on portal-venous, transitional and hepatobiliary phase (1/69). Changes in size (disappearance, regression, no change, growth) and rate of progression to HCC were compared among groups using Chi-square test. Diagnosis of HCC was done by pathology and ASSLD 2010 criteria.
Results: 5/69 (7%) nodules (3 Group E; 2 group D) become HCC. There was no statistically significant difference in progression to HCC among groups. There was no statistically significant difference in rate of nodule disappearance (6%; 8/69); regression (4%; 7/69), no change (55%; 38/69) and growth (23%; 16/69) among groups.
Conclusion: Indeterminate hepatic nodules at Gd-EOB-DPTA-enhanced MRI in cirrhotic patients rarely progress to HCC.
Fate of subcentimeter arterially enhancing and hepatobiliary hypointense lesions seen on gadoxetate-enhanced MRI in patients at risk of HCC
Purpose: To investigate the significance of subcentimeter (≤1 cm) arterially enhancing and hepatobiliary hypointense lesions (SAELs) found on gadoxetate-enhanced MRI in patients at risk of hepatocellular carcinoma (HCC).
Methods and Materials: Two radiologists jointly reviewed gadoxetate-enhanced MRI obtained from September 2008 to March 2013 in 2,311 patients at high risk of HCC, and found 52 SAELs that were confirmed histologically or followed by imaging over 12 months: 17 isolated SAELS in patients without accompanying HCC and 35 in patients with early-stage HCC (solitary or up to 3 nodules ≤ 5cm in size without vascular invasion). SAELs found in patients at beyond early stages were excluded. The imaging findings were compared between the malignant and benign SAELs.
Results: Of total 52 SAELs (in 46 patients), 30 (57.7%) were finally diagnosed as HCCs: 10 (58.8%) among 17 isolated SAELs and 20 (57.1%) among 35 SAELs accompanying early-stage HCCs. All HCCs were resectable with curative intention at the time of diagnosis. Diagnostic accuracy of HCC based on imaging findings of arterial enhancement and venous-phase washout yielded the sensitivity, specificity, and positive predictive values of 83.3%, 50%, and 69.4%, respectively. Venous-phase washout was more frequently observed in the malignant SAELs than in the benign SAELs (57.7% vs. 30.6%; P = 0.01).
Conclusion: SAELs found on gadoxetate-enhanced MRI in patients at risk of HCC have a high malignant potential. Close observation can be an appropriate strategy for isolated SAELs. Venous-phase washout may be helpful to predict malignant SAELs.
Hypovascular hypointense nodules on hepatobiliary phase without T2 hyperintensity on gadoxetic acid-enhanced MRI: long-term outcomes and risk factors of hypervascularisation
Purpose: To evaluate the long-term outcome and risk factors associated with hypervascularisation in hypovascular hypointense nodules on hepatobiliary phase (HBP) without T2 hyperintensity on gadoxetic acid-enhanced magnetic resonance (MR) images in patients with chronic liver disease.
Methods and Materials: The institutional review board-approved retrospective analysis included 97 chronic liver disease patients with 222 hypovascular hypointense nodules on HBP without T2 hyperintensity on gadoxetic acid-enhanced MRI. The following MR features were analysed: nodule size, presence of fat, degree of hypointensity on HBP, degree of signal intensity on diffusion-weighted images (DWI) and T1-weighted images (T1WI). Baseline clinical characteristics were also obtained. Univariate and multivariate analysis with a Cox proportional hazard regression model was used for statistical analysis.
Results: The mean follow-up interval was 997 days (range, 137-1804 days). Of them, 41 nodules (18.5 %) became hypervascular hepatocellular carcinoma (HCC). Both univariate and multivariate Cox analysis identified that previous history of HCC, follow-up nodule size, degree of hyperintensity on T1WI and DWI were significant risk factors for hypervascularisation (P<0.05). The mean growth rate of hypervascularised nodule was 2.89 x 10-3/days and 0.68 x 10-3/days for non-hypervascularised nodule. The cumulative incidence of hypervascularisation was 3.7 % at 1 year, 12.9 % at 2 years, 21.3 % at 3 years.
Conclusion: The previous history of HCC, follow-up nodule size, degree of hyperintensity at T1WI and DWI were associated with progression to hypervascular HCC.
Purpose: To determine the value of CT perfusion (CTP) to predict complete response early after transarterial chemoembolisation (TACE).
Methods and Materials: This prospective study was approved by the institutional review board, written informed consent was obtained from all patients. CTP was performed before and one day (interquartile range[IQR]: 1-2 days) after TACE in 18 patients. 46 lesions (41 hepatocellular carcinoma [HCC] and 5 hypervascular cholangiocarcinoma [CCC]) were amenable for analysis. Blood flow (BF), Blood volume (BV), time to start (TTS), arterial liver perfusion (ALP), portal liver perfusion (PVP) and hepatic perfusion index (HPI) were measured in all lesions, and values before and after TACE, as well as relative change was compared to clinical response on biphasic CT using mRECIST criteria 6 weeks after the procedure. Optimal cut-off values for detecting response were calculated using area under the ROC curves (AUROC). Long-term outcome was assessed using Kaplan-Meier estimates.
Results: CTP parameters were all significantly reduced after TACE in responding patients (PR, CR) while no difference was observed in non-responders. ALPpost was superior in prediction of CR compared to BFpost and BVpost (AUROCALP 0.953 vs. AUROCBF 0.859 and AUROCBV 0.831, p<0.001) with a sensitivity, specificity, PPV, NPV and accuracy of 91%, 92%, 91%, 92% and 91%. Only 4/22 lesions with CR recurred with a median progression local-recurrence free survival of 22.7 months.
Conclusion: CT Perfusion is a feasible tool for early response assessment after transarterial chemoembolisation and can reliably detect lesions with complete response.
HCC showing complete response according to mRECIST on CT after a first session of TACE: is lipiodol deposition a good predictor of local recurrence?
Purpose: To evaluate if the lipiodol deposition pattern can predict local recurrence in hepatocellular carcinoma (HCC) nodules with complete response (CR) according to mRECIST on CT after a first session of conventional chemoembolisation (cTACE).
Methods and Materials: From January 2012 to May 2014 all consecutive patients undergoing a first cTACE for HCC were identified. Inclusion criteria were presence of ≤3 HCCs and available pre- and post-cTACE CECT. Each treated tumour response was classified according to mRECIST. The analysis focused on tumours showing CR. For them, the lipiodol deposition pattern was classified as complete (C-Lip, covering the entire tumour volume), or incomplete (I-Lip). Local recurrence was defined as the reappearance of enhancing areas on arterial phase showing washout on portal/delayed phase within 2 cm from treated tumours on follow-up CT.
Results: Final population included 50 patients (mean age 62+/-12 y; 45 male (90%)) with 82 HCCs (mean 26.8+/-14.2 mm). A total of 46 (52%) HCCs were classified as CR, including 16 (35% - mean 22.9+/-8 mm) with incomplete, and 30 (65% - mean 22.8+/-10 mm) with complete lipiodol deposition. After a median follow-up of 14 months (range 3.2-35.9 months), 15/16 (94%) and 10/30 (30%) of I-Lip and C-Lip HCCs showed local recurrence on CT (p<0.001). No statistical difference regarding delay of recurrence was noted between I-Lip and C-Lip HCCs (mean 334 vs 401 days p=0.519).
Conclusion: Despite showing CR according to mRECIST, HCCs with incomplete lipiodol deposition have a high risk of recurrence and should be considered as incompletely treated.
Purpose: Aim of study is to evaluate accuracy of DW-MRI in evaluating HCC response post-DEB TACE and compare results with DCE MRI.
Methods and Materials: 42 patients with 59 lesions underwent conventional precontrast abdominal MR, DWI, ADC map with ADC value measurement and DCE MR. According to DCE MR, the lesions have been classified into 4 groups (complete response, heterogeneous enhancement, partial nodular enhancement, diffuse enhancement groups). Qualitative DWIs and ADC values were correlated to the DCE MR findings.
Results: Upon comparing qualitative DWI findings to DCE MRI in the evaluation of HCC response to DEB TACE, it showed sensitivity of 83.9%, a specificity of 64.3%, a positive predictive value of 72.2%, and a negative predictive value of 78.3% and overall accuracy of 74.5%. The measured ADC values showed significant difference (P value <0.05) between ADC values measured in active tumoural areas and those measured in necrotic areas with no significant difference between areas of active tumoural enhancement in the different groups. ROC analysis was performed for ADC values showing area under curve 0.7 and maximum combined sensitivity and specificity of 79% and 69.6%, respectively, at cutoff ADC value of 1.395mm²/sec.
Conclusion: DW-MRI is a useful highly sensitive technique in the evaluation of HCC response to DEB TACE, yet it has low specificity related to high number of false-positive results preventing using it solely. In addition, DWI is a reliable method in differentiation between active tumour residue/recurrence and benign perilesional enhancement.
Liver stiffness measured by 2D shear-wave elastography: prognostic values after radiofrequency ablation for hepatocellular carcinoma
Purpose: Liver stiffness (LS) measured by supersonic shear-wave elastography (SWE) can estimate the degree of liver fibrosis. We prospectively evaluated the prognostic value of LS measured using SWE in patients with hepatocellular carcinoma (HCC) treated by radiofrequency ablation (RFA).
Methods and Materials: We prospectively enrolled a total of 145 patients with up to three HCCs≤5cm treated by RFA and who underwent pre-procedural 2D-SWE from January 2012 to December 2013. LS values were measured using real-time SWE (Aixplorer; Supersonic Imagine, France). After a mean follow-up of 32.2±11.5 months, we analysed overall survival (OS) after RFA using the Kaplan-Meier method and Cox proportional hazard regression model. The optimal cut-off LS value to predict OS was determined by the minimal P-value approach.
Results: During the follow-up period, 22 patients died and 11 patients underwent liver transplantation. The estimated 1- and 3-year OS was 96.4% and 85.8%, respectively. The LS value measured by 2D-SWE was a significant predictive factor for OS after RFA for HCC, as was the presence of extrahepatic metastases. The optimal cut-off LS value to predict OS was set at 13.3 kilopascal (kPa). Seventy-nine patients had an LS value ≥13.3kPa, and the estimated 3-year OS was 77.5%, compared to 96.4% in 66 patients with an LS value <13.3kPa. This difference was statistically significant (hazard ratio=5.27 [1.35-20.5]; P=0.017).
Conclusion: The LS value ≥3.3kPa measured by 2D-SWE is a significant predictive factor for OS after RFA for HCC.