To determine whether image texture parameters analysed on pre-operative contrast-enhanced computed tomography (CT) can predict overall survival and recurrence-free survival in patients with hepatocellular carcinoma (HCC) treated by surgical resection.

We retrospectively included all patients operated for HCC who had liver contrast-enhanced CT within 3 months prior to treatment in our centre between 2010 and 2015. The following texture parameters were evaluated on late-arterial and portal-venous phases: mean grey-level, standard deviation, kurtosis, skewness and entropy. Measurements were made before and after spatial filtration at different anatomical scales (SSF) ranging from 2 (fine texture) to 6 (coarse texture). Lasso penalised Cox regression analyses were performed to identify independent predictors of overall survival and recurrence-free survival.

Forty-seven patients were included. Median follow-up time was 345 days (interquartile range [IQR], 176-569). Nineteen patients had a recurrence at a median time of 190 days (IQR, 141-274) and 13 died at a median time of 274 days (IQR, 96-411). At arterial CT phase, kurtosis at SSF = 4 (hazard ratio [95% confidence interval] = 3.23 [1.35-7.71] p = 0.0084) was independent predictor of overall survival. At portal-venous phase, skewness without filtration (HR [CI 95%] = 353.44 [1.31-95102.23], p = 0.039), at SSF2 scale (HR [CI 95%] = 438.73 [2.44-78968.25], p = 0.022) and SSF3 (HR [CI 95%] = 14.43 [1.38-150.51], p = 0.026) were independently associated with overall survival. No textural feature was identified as predictor of recurrence-free survival.

In patients with resectable HCC, portal venous phase-derived CT skewness is significantly associated with overall survival and may potentially become a useful tool to select the best candidates for resection.

• HCC heterogeneity as evaluated by texture analysis of contrast-enhanced CT images may predict overall survival in patients treated by surgical resection. • Among texture parameters, skewness assessed at different anatomical scales at portal-venous phase CT is an independent predictor of overall survival after resection. • In patients with HCC, CT texture analysis may have the potential to become a useful tool to select the best candidates for resection.

To evaluate whether pathologically early hepatocellular carcinoma (HCC) exhibited local tumor progression after radiofrequency ablation (RFA) less often than typical HCC.

Fifty pathologically early HCCs [tumor diameter (mm): mean, 15.8; range, 10-23; follow-up days after RFA: median, 1213; range, 216-2137] and 187 typical HCCs [tumor diameter (mm): mean, 15.6; range, 6-30; follow-up days after RFA: median, 1116; range, 190-2328] were enrolled in this retrospective study. The presence of stromal invasion (namely, tumor cell invasion into the intratumoral portal tracts) was considered to be the most important pathologic finding for the diagnosis of early HCCs. Typical HCC was defined as the presence of a hyper-vascular lesion accompanied by delayed washout using contrast-enhanced computed tomography or contrast-enhanced magnetic resonance imaging. Follow-up examinations were performed at 3-mo intervals to monitor for signs of local tumor progression. The local tumor progression rates of pathologically early HCCs and typical HCCs were then determined using the Kaplan-Meier method.

During the follow-up period for the 50 pathologically early HCCs, 49 (98%) of the nodules did not exhibit local tumor progression. However, 1 nodule (2%) was associated with a local tumor progression found 636 d after RFA. For the 187 typical HCCs, 46 (24.6%) of the nodules exhibited local recurrence after RFA. The follow-up period until the local tumor progression of typical HCC was a median of 605 d, ranging from 181 to 1741 d. Among the cases with typical HCCs, local tumor progression had occurred in 7.0% (7/187), 16.0% (30/187), 21.9% (41/187) and 24.6% (46/187) of the cases at 1, 2, 3 and 4 years, respectively. Pathologically early HCC was statistically associated with a lower rate of local tumor progression, compared with typical HCC, when evaluated using a log-rank test (P = 0.002).

The rate of local tumor progression for pathologically early HCCs after RFA was significantly lower than that for typical HCCs.

The purpose of this study was to evaluate whether the hypervascularity of hepatocellular carcinomas (HCCs) on contrast-enhanced ultrasonography (CEUS) prior to radiofrequency ablation (RFA) is a significant risk factor for local recurrence after RFA.

Institutional review board approval and informed consent were obtained. Overall, 208 patients (mean age, 71.7 years; range, 50-87 years; 137 men, 71 women) with 282 HCCs treated with RFA were analyzed retrospectively. The mean maximum tumor diameter was 15.7mm. We compared the abilities of CEUS and contrast-enhanced computed tomography (CECT) to detect hypervascularity in HCCs. We then classified the HCCs into two groups according to the arterial-phase CEUS findings: a "hypervascular group" with whole or partial hypervascular areas within the lesions compared with the surrounding liver parenchyma, and a "non-hypervascular group" with isovascular or hypovascular areas within the lesions. We assessed the cumulative rate of local recurrence after RFA, and we also evaluated the risk factors for local recurrence using a univariate analysis.

The detection rate for hypervascular HCCs was significantly higher using CEUS (78%, 221/282) than that using CECT (66%, 186/282) (P<0.001). Using the CEUS findings, the cumulative rate of local recurrence was significantly higher in the hypervascular group (41.2%, 56/221) than in the non-hypervascular group (18.4%, 6/61) (P=0.007). A univariate analysis revealed that hypervascularity on CEUS was an independent risk factor for local recurrence (P=0.010).

Hypervascularity in HCCs as observed using CEUS is a significant risk factor for local recurrence after RFA.

Purpose To characterize vessel occlusion rates and their role in local tumor progression in patients with hepatocellular carcinoma (HCC) who underwent microwave tumor ablation. Materials and Methods This institutional review board approved, HIPAA-compliant retrospective review included 95 patients (75 men and 20 women) with 124 primary HCCs who were treated at a single center between January 2011 and March 2014. Complete occlusion of the portal veins, hepatic veins, and hepatic arteries within and directly abutting the ablation zone was identified with postprocedure contrast material-enhanced computed tomography. For each vessel identified in the ablation zone, its size and antenna spacing were recorded and correlated with vascular occlusion with logistic regression analysis. Local tumor progression rates were then compared between patent and occluded vessels for each vessel type with Fisher exact test. Results Occlusion was identified in 39.7% of portal veins (29 of 73), 15.0% of hepatic veins (six of 40), and 14.2% of hepatic arteries (10 of 70) encompassed within the ablation zone. Hepatic vein occlusion was significantly correlated with a smaller vessel size (P = .036) and vessel-antenna spacing (P = .006). Portal vein occlusion was only significantly correlated with a smaller vessel size (P = .001), particularly in vessels that were less than 3 mm in diameter. Local tumor progression rates were significantly correlated with patent hepatic arteries within the ablation zone (P = .02) but not with patent hepatic (P = .57) or portal (P = .14) veins. Conclusion During microwave ablation of HCC, hepatic veins and arteries were resistant to vessel occlusion compared with portal veins, and only arterial patency within an ablation zone was related to local tumor progression. ^© RSNA, 2016.

Radiofrequency ablation (RFA) has gained a wide acceptance as a first-line therapeutic option for small hepatocellular carcinoma (HCC). For very early-stage HCC, despite a higher rate of local tumour progression, RFA is considered as a viable alternative to surgical resection owing to its comparable long-term survival, reduced morbidity, and greater preservation of hepatic parenchyma. For HCCs larger than 2 cm, RFA can contribute to near-curative therapy when combined with chemoembolization. RFA can be used as part of a multimodal treatment strategy for more advanced or recurrent cases, and could be a useful bridging therapy for patients who are waiting for liver transplantation. However, the use of RFA is still limited in treating large tumours and some tumours in high-risk locations. To overcome its current limitations, other ablation techniques are being developed and it is important to validate the role of other techniques for enhancing performance of ablation therapy for HCC.

Our aim was to determine how well ablative margin (AM) grading assessed by magnetic resonance imaging (MRI) with ferucarbotran administered prior to radiofrequency ablation (RFA) predicts local tumor progression in comparison with enhanced computed tomography (CT).

101 hepatocellular carcinomas were treated by RFA after ferucarbotran administration. We performed T2*-weighted MRI after 1 week and enhanced CT after 1 month. The assessment was categorized in three grades: AM(+): high-intensity area with continuous low-intensity rim; AM zero: high-intensity area with discontinuous low-intensity rim; and AM(-): high-intensity area extending beyond the low-intensity rim.

AM(+), AM zero, AM(-) and indeterminable were found in 47, 36, 8 and 10 nodules, respectively. The overall agreement rate between MRI and enhanced CT for the diagnosis of AM was 71.3%. The κ coefficient was 0.523 (p < 0.001), indicating moderate agreement. Multivariate logistic regression showed that a significant factor for the achievement of AM(+) on MRI was only segment location (odds ratio 5.9, non-segment 4 + 8 vs. segment 4 + 8). The cumulative local tumor progression rates (4.4, 7.6, and 7.6% in 1, 2, and 3 years) in 47 AM(+) nodules were significantly lower than those (13.9, 33.4, and 41.8% in 1, 2, and 3 years) in 36 AM zero nodules. A multivariate Cox proportional hazards model identified contiguous vessels (odds ratio 12.0) and AM(+) on MRI (odds ratio 0.19) as independent factors for local tumor progression.

AM assessment by MRI using ferucarbotran can predict local tumor progression after RFA and enable early and less invasive diagnosis.

To evaluate the differences in enhancement pattern of hepatocellular carcinoma (HCC) 20 mm or smaller and enhancement effects of hepatic vessels on early dynamic contrast-enhanced magnetic resonance imaging (MRI) obtained with gadoxetic acid and gadopentetate dimeglumine in the same patients with cirrhosis.

We reviewed MR images using gadoxetic acid and gadopentetate dimeglumine in the same 34 patients with 42 histologically confirmed HCCs (median diameter, 14.5 mm). The percentage enhancements (PEs) of HCC, the hepatic artery and portal vein and relative contrasts (RCs) between HCC and the liver were calculated and analyzed statistically.

The PEs of HCC, the hepatic artery and portal vein were significantly lower for gadoxetic acid in comparison with gadopentetate dimeglumine in the arterial phase (p = 0.0256 for HCC, p < 0.0001 for hepatic artery) and portal phase (p < 0.0001 for HCC, portal vein). The RC between HCC and the liver was significantly lower for gadoxetic acid in comparison with gadopentetate dimeglumine in the arterial phase (p = 0.0422), but was not significantly different in the portal phase (p = 0.1133). Forty-one of the 42 (97.62 %) nodules showed arterial hypervascularization. Of these, 31 (75.61 %) nodules were hypointense in the portal phase for gadoxetic acid, and 22 (53.66 %) were hypointense for gadopentetate dimeglumine (p = 0.038).

Compared with gadopentetate dimeglumine, gadoxetic acid-enhanced MRI demonstrated a different enhancement pattern of inferior arterial enhancement and was more rapidly hypointense in the portal phase for HCC. It showed markedly lower enhancement for hepatic artery and portal vein in the patients with cirrhosis.

The purpose of our study was to determine whether gadoxetate disodium-enhanced MRI improves the detection of locally recurrent hepatocellular carcinoma (HCC) after radiofrequency ablation therapy.

Forty-two consecutive patients (26 men and 16 women; mean age, 70.9 years) who had undergone radiofrequency ablation therapy for HCC were examined by gadoxetate disodium-enhanced MRI. Three independent radiologists reviewed two sets of MR images separately. Set 1 contained unenhanced T1-weighted, T2-weighted, and contrast-enhanced extracellular phase images, and set 2 contained unenhanced T1-weighted, T2-weighted, and contrast-enhanced extracellular and hepatocyte phase images. Diagnostic accuracies were determined by receiver operating characteristic (ROC) curve analysis.

Of the 61 ablated lesions included, 10 developed local recurrence and 51 did not. Sensitivity (60-80% and 70-80% for sets 1 and 2, respectively), specificity (90-98% and 92-100%), and accuracy (87-95% and 89-97%) were not improved by adding hepatocyte phase images. Rather, mean area under the ROC curve (AUC) significantly decreased from 0.94 for set 1 to 0.82 for set 2 (p = 0.046), and the AUC of the radiologist with the least experience significantly decreased from 0.91 to 0.75 (p = 0.037).

For gadoxetate disodium-enhanced MRI, the detection of local recurrence of HCC after radiofrequency ablation therapy was not found to be improved by including hepatocyte phase images. Furthermore, detection may be further impaired when a less-experienced radiologist interprets the images.

Despite the high complete necrosis rate of radio-frequency ablation (RFA) or the complete removal following curative hepatic resection (HR), recurrent hepatocellular carcinoma (HCC) remains a significant problem. The aim of the study is to identify some intraoperative ultrasound (IOUS) patterns, predicting intrahepatic recurrences. From January 1997 to July 2009, 410 patients with HCC were treated (162 HR and 248 RFA through a surgical access). All patients were submitted to IOUS examination: 148 IOUS were performed during the laparotomic access while 262 IOUS were performed during the laparoscopic access. Primary HCC was classified according to diameter, HCC pattern (nodular or infiltrative), echogenicity (hyper- or hypo-echoic), echotexture (homogeneous or inhomogeneous), capsular invasion, mosaic pattern, nodule in nodule aspect and infiltration of portal vessels. Number of HCC nodules was also considered. Multivariate analysis (Cox model) was performed to determine features associated with recurrent HCC using IOUS patterns that independently predicted recurrent HCC, a IOUS score was developed. The patients were followed for 3-127 months, (median follow-up: 21.5 months). In 220 patients (54%), intrahepatic recurrences occurred. In 155 patients (38%), distant intrahepatic recurrences arose in different segments at the primary tumor site. In 65 HCC cases (16%), local recurrences were found. At multivariate analysis, multiple nodules, HCC diameter (>20 mm), HCC pattern (infiltrative), hyperechoic nodule and portal infiltration were statistically significant for risk factor of intrahepatic recurrences. Therefore, a IOUS scoring system was calculated on the basis of multivariate analysis and identified three risk categories of patients: in group 1 recurrences occurred in 37%, group 2 in 46% and group 3 in 66% (p = 0.0001). IOUS is an accurate staging tool during "surgical" procedures. This study showed an added value of IOUS: it permitted to identify ultrasound patterns, which can predict the risk of HCC recurrences. The calculated IOUS score permits to intraoperatively evaluate the actual surgical choice and to program the best treatment strategies during the follow-up period.