The Exploration On The Value Of MRI And MRI Combined With Relevant LncRNAs In Predicting Histological Grade Of HCC

Part I A study on the predictive value of enhanced MRI by Gd-BOPTA for HCC histological gradeBackgroundHepatocellular carcinoma(HCC),as a common primary malignant hepatic neoplasm,usually occurs in patients with chronic hepatic parenchymal disease,and it has high morbidity and mortality worldwide,especially in Africa and Asia.The prevalence rate in males is 2-4 times higher than in females,and HCC accounts for the second highest mortality rate among male tumor-related diseases.Surgical resection or liver transplantation is the main treatment for HCC.Histologic grade is a primary predictive factor of recurrence and metastasis in HCC patients after receiving liver transplantation and resection.Compared with well/moderatelydifferentiated HCC,poorly-differentiated HCC is more likely to have recurrence or metastasis after surgical resection.Therefore,it is very important to determine the histological grade of HCC for proper diagnosis and treatment in clinic.Biopsy of HCC patients is a relatively reliable method to evaluate the preoperative histological grade of HCC,but it is invasive and has some disadvantages such as bleeding,needle spread and biopsy sampling deviation.Therefore,it is particularly necessary to develop a non-invasive and relatively accurate method to evaluate the preoperative histological grade of HCC.Nowadays,with the continuous development of MR technology,MRI is more and more widely used in the diagnosis of abdominal diseases.Compared with CT and ultrasound images,MRI has superior soft-tissue contrast,and can be used for imaging with multiple parameters and arbitrary slice orientation.The sensitivity and accuracy of MRI are both higher than CT for HCC diagnosis.Especially,the introduction of hepatospecific contrast agents has further improved the diagnosis and differential diagnosis of HCC.Currently,enhanced MR with hepatospecific contrast agents has become the most sensitive imaging examination method to detect HCC.As two of the most widely used hepatospecific contrast agents,Gd-EOB-DTPA and Gd-BOPTA are reported to have similar diagnostic accuracy in HCC.Previous reports had attempted to use enhanced MRI with hepatospecific contrast agents or ADC value to evaluate HCC histological grade,but the consensus was not obtained.This study was to explore the value of quantitative MRI measure after Gd-BOPTA injection in distinguishing different histological grades of HCC.ObjectiveTo evaluate the predictive value of quantitative parameters of gadolinium beglumine(Gd-BOPTA)-enhanced MRI for histological grade of HCC.MethodsBetween January 2016 and January 2020,102 HCC patients underwent Gd-BOPTA-enhanced MRI in the Hospital were included in this study retrospectively.All patients were confirmed by histopathology examination after surgery,and histological grade was assessed based on Edmonson-Steiner grade system.The relationship between clinical,imaging features and histological grade was analyzed.The SPSS software 20.0(IBMSPSS,Chicago,IL)was used to perform data analysis.Categorical variables were compared by chi-square or Fisher’s exact test.Continuous variables were presented as mean value ±standard deviation(minimal value-maximal value)and compared by ANOVA.On the basis of clinical significance and related references,HCC histological grade Ⅰ～Ⅲ was defined as well/moderately-differentiated,and grade Ⅳ as poorly-differentiated.Receiver operating characteristic(ROC)curves were plotted for T1-relative intensity ratio(T1RIR),T2-relative intensity ratio(T2RIR),contrast-enhancement ratio in arterial phase(CERA)and contrast-noise-ratio(CNR)in hepatobiliary phase(CNRhbp),and area under the curve(AUC)was calculated to evaluate the diagnostic performance of each quantitative parameters on distinguishing between well/moderately-differentiated and poorly-differentiated HCC.P<0.05 was considered statistically significant.ResultsA total of 102 HCC patients were analyzed(grade Ⅰ,n=19;grade Ⅱ,n=30;grade Ⅲ,n=30;grade Ⅳ,n=23).There was no statistical difference regarding to clinical features among patients with different HCC histological grades(P>0.05),while the MRI findings of tumor necrosis(P<0.05),the capsule integrity(P<0.05),tumor emboli(P<0.01)were statistically significant.There were significant differences in CERA,CNRhbp,T1RIR,T2RIR and ADC values among patients with different HCC histological grades(all P<0.05).ROC curves were analyzed to compare the five MRI quantitative parameters.The cut-off value of ADC in distinguishing between well/moderately-differentiated and poorly-differentiated HCC was 0.771 ×10-3mm2/s(sensitivity 84.8%,specificity 82.6%,AUC value 0.862),the cut-off value of CNRhbp was 40.5(sensitivity 78.5%,specificity 91.3%,AUC value 0.916),the cut-off value of CERA was 1.01(sensitivity 54%,specificity 96%,AUC value 0.697),the cut-off value of T1RIR was 0.73(sensitivity 70%,specificity 78%,AUC value 0.746),the cut-off value of T2RIR was 2.06(sensitivity 24%,specificity 13%,AUC value 0.451).Conclusions1.MRI findings of tumor necrosis,capsule integrity,and tumor emboli could contribute to distinguishing HCC histological grade,and the presence of necrosis,tumor emboli or incomplete capsule might indicate a higher histological grade of HCC.2.The CNRhbp or ADC value on Gd-BOPTA-enhanced MRI might be important quantitative parameters to predict the HCC histological grade.CNRhbp value of 40.5 or ADC value of 0.771 ×10-3 mm2/s might have certain value in distinguishing between well/moderately-differentiated and poorly-differentiated HCC.Part Ⅱ A preliminary study on the predictive value of MRI combined with relevant lncRNAs for HCC histological gradeBackgroundHistological grade of HCC is an important predictor of recurrence and metastasis for patients who have received surgical resection or liver transplantation.Accurate evaluation of histological grade of HCC is crucial for development of treatment plan and prognosis assessment.Currently,the combination of functional MRI and hepatospecific contrast agents has been studied for preoperative histological grade evaluation of HCC,but no definite conclusion has been reached.In section 1 of this study,it is to verify the accuracy of quantitative parameters of Gd-BOPTA-enhanced MRI in predicting the histological grade of HCC based on the Part Ⅰ study’s results.As a traditional biomarker,alpha-fetoprotein(AFP)is widely used in clinic,which provides important clues for clinical diagnosis and prognosis assessment of HCC.However,there are still 20%-40%AFP-negative HCC patients.Hence,it is crucial to identify more potential diagnostic or prognostic biomarkers of HCC.As a group of non-protein coding RNAs with about 200 nucleotides in length,long noncoding RNAs(lncRNAs)are shown to affect a variety of biological processes via regulating the expression of relevant genes at different levels,thus involving in the occurrence and development of HCC.Moreover,since lncRNAs are relatively stable in body fluids,and detectable in both peripheral blood and tissues,many studies have proved that some lncRNAs can be used as potential biomarkers for HCC diagnosis,prognosis or histological grade.Therefore,in section 2 of this study,12 lncRNAs closely related to histological grade or invasiveness of HCC were selected by reviewing relevant literature,and the correlation between those lncRNAs and HCC histological grade was analyzed.With the development of magnetic resonance technology,and the deepening of basic medical research,there have been many studies of MRI combined with serology and molecular examination in evaluating diseases.Therefore,we preliminarily investigated whether the combination of lncRNA and MRI features of HCC could improve the diagnostic performance of preoperative evaluation of HCC histological grade.ObjectiveThis study aimed to preliminarily explore the feasibility of MRI combined with the relevant lncRNAs in predicting the histological grade of HCC.MethodsThis prospective study collected the data of 72 patients with suspected HCC who underwent gadolinium beglumine(Gd-BOPTA)-enhanced MR examination in the Hospital from April to December 2020.Blood samples were collected 1 to 3 days before surgery for lncRNAs extraction.According to inclusion and exclusion criteria,37 patients were eventually included in the study group.Meanwhile,blood samples of 19 healthy volunteers were selected as control group.Histological grade was assessed based on Edmonson-Steiner grade system.The SPSS software 20.0(IBMSPSS,Chicago,IL)was used to perform data analysis.The relationship between clinical,MRI features and histological grade was analyzed.The relationship between 12 lncRNAs(RP11-486012.2,RP11-863K10.6,LINC01093,RP11-273G15.2,FAM99A,FAM99B,LINC00261,SNHG12,SNHG7,SNHG1,GAS5,PVT1)with the histological grade of HCC was also evaluated.chi-square and Fisher’s exact test were used to compare the inter-group difference for categorical variables,and ANOVA was used to compare the inter-group difference for continuous variables.The differences of MRI parameters and lncRNA indicators in different HCC histological grades were compared by ANOVA.The HCC histological grades were taken as the dependent variable(Histological grade Ⅰ-Ⅲ was defined as well/moderately-differentiated,and grade Ⅳ as poorly-differentiated).The statistically significant variables of MRI parameters and lncRNA indicators were taken as independent variables for ROC curve analysis,and were considered for the multivariable logistic regression models,with MRI-related model as Model1,lncRNA related-model as Model2,MRI combined lncRNA-related model as Model3.Receiver operating characteristic(ROC)curves were plotted for the three Models,and area under the curve(AUC)was calculated to evaluate the diagnostic performance of each Model on distinguishing between well/moderately-differentiated and poorly-differentiated HCC.P<0.05 was considered statistically significant.Results1 Clinical and MRI features of HCC patients with different histological gradesA total of 37 HCC patients were analyzed,including 2 patients with grade Ⅰ,16 patients with grade Ⅱ,11 patients with grade Ⅲ,and 8 patients with grade Ⅳ.There was statistically significant for tumor size among different HCC histological grades,while no significant difference in other clinical characteristics(age,gender,lesion location,Child-Pugh grade,AFP and Ki67(%))(P>0.05).MRI features showed no significant difference for necrosis,hemorrhage,capsule,steatosis and liver cirrhosis(P>0.05).Fisher’s exact test was used to find that tumor emboli was statistically significant among different HCC histological grades(P<0.05).2 MRI quantitative parameters and lncRNA indicators of HCC patients with different histological gradesANOVA was used for MRI parameters in different HCC histological grades,and found that there were significant differences for CNRhbp(P<0.01)and ADC values(P<0.05)among patients with different HCC histological grades,while no significant differences for CERA,T1RIR,and T2RIR(all P<0.05).ANOVA was used for 12 lncRNAs in different HCC histological grades,and found that the mean values of LINC01093,RP11-273G15.2,FAM99A,LINC00261 and SNHG7 showed significant differences in HCC patients with different histological grades(P<0.05).3 ROC curve analysis of MRI and lncRNA indicators to predict HCC histological gradeThe sensitivity,specificity,cut-off value and AUC value for distinguishing between well/moderately-differentiated and poorly-differentiated HCC of the two MRI quantitative parameters(CNRhbp and ADC value)and the five lncRNAs(LINC01093,RP11-273G15.2,FAM99A,LINC00261,SNHG7)with statistical difference were calculated,respectively.The results showed that the cut-off value of CNRhbp in distinguishing between well/moderatelydifferentiated and poorly-differentiated HCC was 43.5(sensitivity 100%,specificity 79.3%,AUC value 0.916),the cut-off value of ADC was 0.702×10-3mm2/s(sensitivity 62.5%,specificity 93.1%,AUC value 0.765),the cut-off value of LINC01093 was 0.357(sensitivity 100%,specificity 13.8%,AUC value 0.504),the cut-off value of RP11-273G15.2 was 0.998(sensitivity 75%,specificity 72.4%,AUC value 0.655),the cut-off value of FAM99A was 1.919(sensitivity 100%,specificity 55.2%,AUC value 0.677),the cut-off value of LINC00261 was 0.586(sensitivity 87.5%,specificity 41.4%,AUC value 0.595),and the cut-off value of SNHG7 was 2.089(sensitivity 25%,specificity 96.6%,AUC value 0.530).4 Validation of the conclusion about MRI quantitative parameters in Part ⅠThe accuracy of CNRhbp at the cut-off value of 40.5 found in Part Ⅰ was 81.08%,and the accuracy of ADC at the cut-off value of 0.771 ×10-3mm2/s was 78.38%for predicting the histological grade of HCC.The cut-off values of CNRhbp and ADC values were compared between Part Ⅰ and Part Ⅱ.The absolute differences of CNRhbp and ADC at the cut-off value were 3 and 0.069×10-3mm2/s,respectively.5 Multivariate analysis and performance evaluation of MRI/lncRNA-related indicators for predicting HCC histological gradeMultivariate logistic regression models were constructed with MRI/lncRNA-related indicators.According to the results of multivariate logistic regression analysis,only CNRhbp was statistically different(P<0.05)in Modell and Model3.The ROC curves of the three models were plotted,and the AUC value of Model1(MRI-related variables)was 0.914,the AUC value of Model2(lncRNA-related variables)was 0.797,and the AUC value of Model3(MRI combined with lncRNA-related variables)was 0.974.Conclusions1.The CNRhbp value was relatively reliable for predicting HCC histological grade in patients with Gd-BOPTA-enhanced MRI.2.The combination of MRI with relevant lncRNAs might have certain value in improving the diagnostic performance of predicting histological grade of HCC.Innovations1.In this study,the relationship of HCC histological grade with clinical information,MRI features,MRI quantitative parameters was comprehensively evaluated,which concluded that some MRI features and quantitative parameters of MR could contribute to predicting HCC histological grade.2.In the part Ⅱ of this study,MRI quantitative parameters in the part Ⅰ were verified prospectively,and the conclusion was that CNRhbp value was relatively reliable for predicting HCC histological grade,which might provide proper guidence for establishing a reasonable scheme of diagnosis and treatment to the HCC patient.3.In this study,the diagnostic performance of MRI quantitative parameters combined with relevant lncRNAs was preliminarily investigated in predicting HCC histological grades,and the conclusion was that the combination might have some value in improving the diagnostic performance in predicting HCC histological grades.It provided some evidences for continuing to explore the underlying molecular mechanism in the relationship between MRI indicators and lncRNAs related to HCC histological grades,which had certain clinical and scientific value.Limitations1.In this study,this single-center study with small sample size could result in a low reliability for some results.2.In this study,the combination of MRI and lncRNA was used to evaluate the HCC histological grade,but the underlying molecular mechanism between the relevant lncRNA expression changes and MRI related indicators in HCC histological grade needs to be further studied.