RPS 2616 - Primary and secondary liver tumours

Author Block: C. Wei, Y. ZHU, X. Ma, X. Zhao; Beijing/CN
Purpose or Learning Objective: Microvascular invasion (MVI) serves as an important prognostic factor for hepatocellular carcinoma (HCC) after an operation. However, predicting MVI in patients with HCC is a clinical challenge as MVI is a histopathological diagnosis. The aim of this study is to investigate the feasibility of quantitative dynamic contrast-enhanced MRI (DCE-MRI) for predicting MVI in small solitary HCC (ssHCC).
Methods or Background: A total of 63 patients with pathologically confirmed ssHCC (≤ 3cm) underwent quantitative DCE-MRI studies and received hepatic resection. A dual-input two-compartment exchange model (2CXM) was used to calculate the values of quantitative permeability and perfusion parameters. The differences in parameters between different MVI status groups were analysed. Multivariate logistic regression was used to build the combined prediction model for MVI prediction with the statistically significant parameters. The predictive performance was evaluated using ROC analysis.
Results or Findings: Among the 63 patients with ssHCC, 22 (34.9%) exhibited MVI positive. The MVI positive group had higher volume transfer constant (Ktrans), reverse reflux rate constant (kep), portal vein blood flow (BFpv), and lower extracellular extravascular volume fraction (ve), hepatic arterial perfusion index (HPI) values than negative group (1.532 min-1 vs. 0.853 min-1, 0.547 min-1 vs. 0.362 min-1, 84.63 mL/min/100g vs. 34.95 mL/min/100g, 0.316 vs. 0.582, 65.32 vs. 84.59, respectively) (P<0.05). Quantitative parameters Ktrans, kep and BFpv values independently associated with MVI with OR values of 4.36, 2.53 and 3.74 (P<0.05) through multivariate logistic regression. ROC analysis showed that the AUC, sensitivity, specificity in predicting MVI by combined Ktrans, kep and BFv values were 0.903, 87.6%, 95.4%, respectively.
Conclusion: Quantitative DCE-MRI derived parameters showed potential value in the prediction MVI in ssHCC.
Limitations: The sample size was relatively small.
Ethics committee approval: Approved by the Independent Ethics Committee of the Cancer Hospital, CAMS (no. 20/412-2608).
Funding for this study: No funding was received for this study.

Author Block: M. Shantarevich, E. V. Kondratyev, G. G. Karmazanovsky; Moscow/RU
Purpose or Learning Objective: Poor tumour differentiation of hepatocellular carcinoma (HCC) correlates with lower overall and disease-free survival. Therefore, accurate non-invasive preoperative prediction of the tumour histologic grade is crucial for patient prognosis. The purpose of this study was to investigate the value of radiomics analysis of contrast-enhanced computed tomography images (CECT) in estimating the histologic tumour grade before surgery in patients with HCC.
Methods or Background: The 36 patients with HCC and preoperative liver CECT who had undergone surgical resection were retrospectively enrolled in the study (25 patients with the tumour Grade 1+Grade2, and 11 patients - with Grade 3). The LIFEx application software (version v7.1.0, www.lifexsoft.org) was used to obtain texture features. 3D ROI that covered the whole tumour was delineated in the images for each patient. Radiomic features were extracted from four phases (native, arterial, portal, and delay). A total of 497 radiomic features were extracted from each CECT phase.
Results or Findings: Tree radiomic features: CONVENTIONAL_HUKurtosis, DISCRETIZED_HUExcessKurtosis and GLZLM_SZE derived from native and arterial phases showed significant positive associations with the histologic grade (p<0,05) and were selected after multiple linear regression analysis. The sensitivity and specificity of radiomics-based model in detecting poor-differentiated HCC from well- and moderate- differentiated HCC were 87,5% and 94,7%, respectively (AUC 0,901±0,078 CI: 0,749-1,0).
Conclusion: The use of the CECT radiomics-based model reflects a better evaluating performance in the prediction of HCC grade, which may contribute to personalised treatment.
Limitations: The limitation of our research is the small number of included patients.
Ethics committee approval: Not applicable.
Funding for this study: Not applicable.

Author Block: R. Schmidt, C. Hamm, H. XU, V. H. Broukal, L. A. Gottwald, B. Gebauer, L. J. Savic; Berlin/DE
Purpose or Learning Objective: Radiogenomic venous invasion (RVI) is a set of imaging biomarkers indicative of the presence of microvascular invasion. This study aims to investigate the predictive value of RVI regarding response and survival of patients with HCC receiving LRT.
Methods or Background: This retrospective study included 95 patients with unresectable HCC, who received ablation using CT-guided high dose-rate brachytherapy alone (n=48) or in combination with transarterial chemoembolisation (TACE, n=47) between 01/2016-12/2017. Patients were stratified according to positive or negative RVI assessed on baseline contrast-enhanced MRI using two decision-tree-models: RVI (art) and RVI (ven) based on the presence of internal vessels in the arterial or portal-venous phase, respectively, and the absence of both a hypointense halo and a sharp tumour-liver-transition in native T1-weighted images. Primary endpoints were overall (OS), progression-free survival (PFS), and time to progression (TTP). Statistics included Fisher’s exact test and Kaplan-Meier analysis.
Results or Findings: Regarding brachytherapy alone, stratification according to RVI (art) achieved significant separation of OS (p=<0.001) and PFS (p=0.029) but not TTP (p=0.142), revealing poorer outcomes for RVI positive patients. RVI (ven) was predictive of TTP (p=0.032) and PFS (p=0.004) but not OS (p=0.078). On the contrary, both RVI types achieved no significant stratification for any endpoint following TACE/brachytherapy. In patients receiving brachytherapy alone, median OS, PFS, and TTP were shorter for RVI (ven) positive compared to negative patients (12.4 vs 40.4, 5.9 vs 13.2, 6.4 vs 11.8 months). In contrast, no difference could be observed for patients receiving TACE/brachytherapy.
Conclusion: While decisions for LRT are currently based on visual assessments of tumour enhancement on baseline MRI, the findings underscore the potential of RVI to identify HCC patients who would benefit from TACE before brachytherapy.
Limitations: This study was done retrospectively at a single site.
Ethics committee approval: Ethics committee approval was obtained.
Funding for this study: No funding was received for this study.