Objective To explore the performance of a new multimodal feature fusion classification model based on distance matching and discriminative representation learning for differentiating high-grade glioma(HGG)from solitary brain metastasis(SBM).Methods We collected multi-parametric magnetic resonance imaging(MRI)data from 61 patients with HGG and 60 with SBM,and delineated regions of interest(ROI)on T1WI,T2WI,T2-weighted fluid attenuated inversion recovery(T2_FLAIR)and post-contrast enhancement T1WI(CE_T1WI)images.The radiomics features were extracted from each sequence using Pyradiomics and fused using a multimodal feature fusion classification model based on distance matching and discriminative representation learning to obtain a classification model.The discriminative performance of the classification model for differentiating HGG from SBM was evaluated using five-fold cross-validation with metrics of specificity,sensitivity,accuracy,and the area under the ROC curve(AUC)and quantitatively compared with other feature fusion models.Visual experiments were conducted to examine the fused features obtained by the proposed model to validate its feasibility and effectiveness.Results The five-fold cross-validation results showed that the proposed multimodal feature fusion classification model had a specificity of 0.871,a sensitivity of 0.817,an accuracy of 0.843,and an AUC of 0.930 for distinguishing HGG from SBM.This feature fusion method exhibited excellent discriminative performance in the visual experiments.Conclusion The proposed multimodal feature fusion classification model has an excellent ability for differentiating HGG from SBM with significant advantages over other feature fusion classification models in discrimination and classification tasks between HGG and SBM.

Objective To predict microvascular invasion(MVI)in hepatocellular carcinoma(HCC)using a model based on multi-phase dynamic-enhanced CT(DCE-CT)radiomics feature and hierarchical fusion of multiple classifiers.Methods We retrospectively collected preoperative DCE-CT images from 111 patients with pathologically confirmed HCC in Guangzhou First People's Hospital between January,2016 and April,2020.The volume of interest was outlined in the early arterial phase,late arterial phase,portal venous phase and equilibrium phase,and radiomics features of these 4 phases were extracted.Seven classifiers based on different algorithms were trained using the filtered feature subsets to obtain multiple base classifiers under each phase.According to the hierarchical fusion strategy,a multi-criteria decision-making-based weight assignment algorithm was used for fusing each base classifier under the same phase with the model after extracting the phase information to obtain the prediction model.The proposed model was evaluated using a 5-fold cross-validation and assessed for area under the ROC curve(AUC),accuracy,sensitivity,and specificity.The prediction model was also compared with the fusion models using a single phase or multiple phases,models based on a single phase with a single classifier,models with different base classifier diversities,and 8 classifier models based on other ensemble methods.Results The experimental results showed that the performance of the proposed model for predicting HCCMVI was optimal after incorporating the 4 phases and 7 classifiers,with AUC,accuracy,sensitivity,and specificity of 0.828,0.766,0.877,and 0.648,respectively.Comparative experiments showed that this prediction model outperformed the models based on a single phase with a single classifier and other ensemble models.Conclusion The proposed prediction model is effective for predicting MVI in HCC with superior performance to other models.

Objective To explore the performance of a new multimodal feature fusion classification model based on distance matching and discriminative representation learning for differentiating high-grade glioma(HGG)from solitary brain metastasis(SBM).Methods We collected multi-parametric magnetic resonance imaging(MRI)data from 61 patients with HGG and 60 with SBM,and delineated regions of interest(ROI)on T1WI,T2WI,T2-weighted fluid attenuated inversion recovery(T2_FLAIR)and post-contrast enhancement T1WI(CE_T1WI)images.The radiomics features were extracted from each sequence using Pyradiomics and fused using a multimodal feature fusion classification model based on distance matching and discriminative representation learning to obtain a classification model.The discriminative performance of the classification model for differentiating HGG from SBM was evaluated using five-fold cross-validation with metrics of specificity,sensitivity,accuracy,and the area under the ROC curve(AUC)and quantitatively compared with other feature fusion models.Visual experiments were conducted to examine the fused features obtained by the proposed model to validate its feasibility and effectiveness.Results The five-fold cross-validation results showed that the proposed multimodal feature fusion classification model had a specificity of 0.871,a sensitivity of 0.817,an accuracy of 0.843,and an AUC of 0.930 for distinguishing HGG from SBM.This feature fusion method exhibited excellent discriminative performance in the visual experiments.Conclusion The proposed multimodal feature fusion classification model has an excellent ability for differentiating HGG from SBM with significant advantages over other feature fusion classification models in discrimination and classification tasks between HGG and SBM.

Objective To predict microvascular invasion(MVI)in hepatocellular carcinoma(HCC)using a model based on multi-phase dynamic-enhanced CT(DCE-CT)radiomics feature and hierarchical fusion of multiple classifiers.Methods We retrospectively collected preoperative DCE-CT images from 111 patients with pathologically confirmed HCC in Guangzhou First People's Hospital between January,2016 and April,2020.The volume of interest was outlined in the early arterial phase,late arterial phase,portal venous phase and equilibrium phase,and radiomics features of these 4 phases were extracted.Seven classifiers based on different algorithms were trained using the filtered feature subsets to obtain multiple base classifiers under each phase.According to the hierarchical fusion strategy,a multi-criteria decision-making-based weight assignment algorithm was used for fusing each base classifier under the same phase with the model after extracting the phase information to obtain the prediction model.The proposed model was evaluated using a 5-fold cross-validation and assessed for area under the ROC curve(AUC),accuracy,sensitivity,and specificity.The prediction model was also compared with the fusion models using a single phase or multiple phases,models based on a single phase with a single classifier,models with different base classifier diversities,and 8 classifier models based on other ensemble methods.Results The experimental results showed that the performance of the proposed model for predicting HCCMVI was optimal after incorporating the 4 phases and 7 classifiers,with AUC,accuracy,sensitivity,and specificity of 0.828,0.766,0.877,and 0.648,respectively.Comparative experiments showed that this prediction model outperformed the models based on a single phase with a single classifier and other ensemble models.Conclusion The proposed prediction model is effective for predicting MVI in HCC with superior performance to other models.

Objective To evaluate the performance of magnetic resonance imaging(MRI)multi-sequence feature imputation and fusion mutual model based on sequence deletion in differentiating high-grade glioma(HGG)from low-grade glioma(LGG).Methods We retrospectively collected multi-sequence MR images from 305 glioma patients,including 189 HGG patients and 116 LGG patients.The region of interest(ROI)of T1-weighted images(T1WI),T2-weighted images(T2WI),T2 fluid attenuated inversion recovery(T2_FLAIR)and post-contrast enhancement T1WI(CE_T1WI)were delineated to extract the radiomics features.A mutual-aid model of MRI multi-sequence feature imputation and fusion based on sequence deletion was used for imputation and fusion of the feature matrix with missing data.The discriminative ability of the model was evaluated using 5-fold cross-validation method and by assessing the accuracy,balanced accuracy,area under the ROC curve(AUC),specificity,and sensitivity.The proposed model was quantitatively compared with other non-holonomic multimodal classification models for discriminating HGG and LGG.Class separability experiments were performed on the latent features learned by the proposed feature imputation and fusion methods to observe the classification effect of the samples in two-dimensional plane.Convergence experiments were used to verify the feasibility of the model.Results For differentiation of HGG from LGG with a missing rate of 10%,the proposed model achieved accuracy,balanced accuracy,AUC,specificity,and sensitivity of 0.777,0.768,0.826,0.754 and 0.780,respectively.The fused latent features showed excellent performance in the class separability experiment,and the algorithm could be iterated to convergence with superior classification performance over other methods at the missing rates of 30%and 50%.Conclusion The proposed model has excellent performance in classification task of HGG and LGG and outperforms other non-holonomic multimodal classification models,demonstrating its potential for efficient processing of non-holonomic multimodal data.

AIM:To explore the efficacy of lenvatinib(Len)in enhancing the therapeutic effects of immune checkpoint inhibitor for hepatocellular carcinoma(HCC)and to delve into its immunomodulatory mechanisms within the tumor microenvironment.METHODS:The effects of various concentrations of Len on the migration of human umbilical vein endothelial cells(HUVECs)and the secretion of CXC chemokine ligand 10(CXCL10)were investigated,and the mechanism by which Len modulates CXCL10 secretion was validated.An orthotopic HCC model was established,and the mice bearing tumors were randomly allocated into 4 groups:PBS group,BMS-202(PD-1/PD-L1 inhibitor)group,Len group,and Len/BMS-202 group.The progression of the orthotopic liver tumors was monitored with small animal in vivo im-aging techniques.On the 13th day after the treatment,mice were sacrificed and tumor tissues were harvested for analysis.Immunofluorescence was employed to identify apoptosis,vascular architecture,and hypoxic status within the tumor tis-sue.The expression levels of proliferation marker Ki67,transforming growth factor-β(TGF-β),and the infiltration de-grees of CD4+T cells and CD8+T cells in the tumor tissue were monitored with immunohistochemistry.The secretion of im-mune factors interferon-γ(IFN-γ),CXCL10 and TGF-α in the mouse serum was quantified with ELISA.Above all data were followed by statistical analysis.RESULTS:(1)Len could facilitate endothelial cell migration within a specific range and potentiated the response of tumor cells to IFN-γ by blocking fibroblast growth factor receptor(FGFR),thereby increasing the secretion of CXCL10 from the tumor cells.(2)Compared with PBS group,tumor growth was slower in all treatment groups,with Len/BMS-202 group showing the most significant inhibition of tumor growth in tumor-bearing mice(P<0.05).(3)Compared with PBS group and monotherapy groups,Len/BMS-202 significantly promoted tumor tissue apoptosis and inhibited tumor cell proliferation(P<0.05).(4)Compared with PBS group and BMS-202 group,both Len group and Len/BMS-202 group manifested a substantial enhancement in pericytes coverage rate(P<0.01),concomitantly showing a marked improvement in hypoxic conditions(P<0.01).(5)Compared with PBS group and monotherapy groups,Len/BMS-202 group showed a significant increase in the infiltration of CD4+T cells and CD8+T cells within the tumor(P<0.01),along with a marked decrease in the expression of TGF-β(P<0.01).(6)Compared with PBS group,all treatment groups collectively induced varying degrees of secretion of IFN-γ,CXCL10 and TGF-α in mouse serum(P<0.05),with Len/BMS-202 group demonstrating the most pronounced effects(P<0.01).CONCLUSION:Lenvatinib may augment the therapeutic efficacy of BMS-202 in HCC by facilitating tumor vascular normalization,alleviating hypoxic conditions,and enhancing the secretion of CXCL10,thereby synergistically activating the tumor immune microenvironment.

Objective To evaluate the performance of magnetic resonance imaging(MRI)multi-sequence feature imputation and fusion mutual model based on sequence deletion in differentiating high-grade glioma(HGG)from low-grade glioma(LGG).Methods We retrospectively collected multi-sequence MR images from 305 glioma patients,including 189 HGG patients and 116 LGG patients.The region of interest(ROI)of T1-weighted images(T1WI),T2-weighted images(T2WI),T2 fluid attenuated inversion recovery(T2_FLAIR)and post-contrast enhancement T1WI(CE_T1WI)were delineated to extract the radiomics features.A mutual-aid model of MRI multi-sequence feature imputation and fusion based on sequence deletion was used for imputation and fusion of the feature matrix with missing data.The discriminative ability of the model was evaluated using 5-fold cross-validation method and by assessing the accuracy,balanced accuracy,area under the ROC curve(AUC),specificity,and sensitivity.The proposed model was quantitatively compared with other non-holonomic multimodal classification models for discriminating HGG and LGG.Class separability experiments were performed on the latent features learned by the proposed feature imputation and fusion methods to observe the classification effect of the samples in two-dimensional plane.Convergence experiments were used to verify the feasibility of the model.Results For differentiation of HGG from LGG with a missing rate of 10%,the proposed model achieved accuracy,balanced accuracy,AUC,specificity,and sensitivity of 0.777,0.768,0.826,0.754 and 0.780,respectively.The fused latent features showed excellent performance in the class separability experiment,and the algorithm could be iterated to convergence with superior classification performance over other methods at the missing rates of 30%and 50%.Conclusion The proposed model has excellent performance in classification task of HGG and LGG and outperforms other non-holonomic multimodal classification models,demonstrating its potential for efficient processing of non-holonomic multimodal data.

Objective To explore the CT dominant phase and optimal classification model in differenting clear cell renal cell carcinoma (ccRCC) from fat‐poor angiomyolipoma (fpAML) through quantitative multiple‐phase CT radiomic features analysis. Methods Clinical and imaging data of 195 cases pathologically confirmed ccRCC (n=131) and fpAML (n=64) were retrospectively studied. All the patients underwent non‐contrast enhanced CT scans and dynamic multi‐phase (corticomedullary phase, medullary phase and excretion phase) contrast‐enhanced CT scans. Regions of interest (ROIs) were manually delineated based on the selected image slices with the maximal diameter of the lesion using ITK‐SNAP software, followed by the acquisition of candidate CT radiomic feature sets from each phase with statistically significant differences by using Mann‐Whitney U test. Then, using the synthetic minority oversampling technique (SMOTE), 232 classification models which are composed of 29 different feature selection algorithms (top 10 features were chosen by the backward elimination method) and 8 different classifiers were constructed. Employing the 5‐fold cross‐validation method, the performance of each classification models for each phase was evaluated using accuracy (ACC), sensitivity (SEN), specificity (SPE) and area under receiver operating characteristic curve (AUC), to acquire dominant CT phases and the optimal classification models for distingushing ccRCC and fpAML, along with the key imaging radiomic features. Results In this study, the mean maximal diameter of ccRCC and fpAML lesions were (3.9±1.4) cm, and (3.5±1.7) cm, respectively, and there was no statistically significant difference in the size of the tumor between two groups (P>0.05). From 102 initial imaging feature sets, the total number of candidate imaging feature sets (P<0.05) were:non‐enhanced phase (n=26), corticomedullary phase (n=71), medullary phase (n=68), excretion phase (n=62). Among the 232 classification models through different combination of classifiers and feature selectors, the amount of classification models which achieved the maximum of AUC value (AUCmax) from different CT phases were: non‐enhanced phase (n=106, 45.7%), corticomedullary phase (n=94, 40.5%), medullary phase (n=23, 9.9%), excretion phase (n=9, 3.9%). Imaging features from non‐enhanced phase and corticomedullary phase yielded higher performance compared with medullary phase and excretion phase, with the corresponding optimal prediction models were SVM‐fisher_score (AUC: 0.897, ACC: 83%, SEN: 84%, SPE:80%) and Logistic Regression‐RFS (AUC: 0.891, ACC: 83%, SEN: 81%, SPE: 89%), respectively. Conclusions The quantitative imaging features from non‐enhanced and corticomedullary phase have better performance among proposed classification models than that from medullary phase and excretion phase. Furthermore, it is feasible to acquire proper combination of feature selection and classifiers to achieve high performance in identifying ccRCC and fpAML.

[Objective]To explore the MRI features of the mucinous breast carcinoma and the correlation with biological prognos?tic factors.[Methods]MRI features of 35 pure and 15 mixed mucinous carcinomas were retrospectively analyzed. MR images were reviewed for shape,margin,the signal intensity,enhancement patterns of tumors and DWI features. All the patients were detected by immunohistochemical staining with expression of ER,PR,CerbB-2,Ki-67 and Her-2. Correlations between the pure and mixed mucinous breast carcinoma and prognostic factors were analyzed.[Results]16 oval masses(16/35,45.7%)and 10 circular masses (10/35,28.6%)were found in 35 pure mucinous breast carcinomas with clear boundary(26/35,74.3%)and lobulated shape(31/35,88.6%);9 irregular masses(9/15,60%)were found in mixed mucinous breast carcinomas with unclear boundary(13/15, 86.7%). Very high signal intensity on T2-weighted images was found in 33 pure mucinous carcinomas(33/35,94.3%)and 11 mixed mucinous carcinomas showed mixed signal intensity(11/15,73.3%). Early enhancement rate was(114.7 ± 9.1)% for pure muci?nous carcinomas and(165.6 ± 14.3)%for mixed mucinous carcinomas. 28 pure mucinous tumors demonstrated persistent enhancing pattern on time-signal intensity curve ,7 pure mucinous tumors demonstrated plateau pattern and 7 mixed mucinous carcinomas showed plateau pattern and washout pattern respectively. Mean ADC value was(1.91 ± 0.06)×10-3 mm2/s for pure mucinous carcino?mas and(1.13±0.08)×10-3mm2/s for mixed mucinous carcinomas. There was significant difference with morphology,boundary,T2WI signal,early enhancement rate,time-signal intensity curve,ADC value between pure and mixed mucinous breast carcinoma(P <0.05). There was significant difference between pure and mixed mucinous breast carcinoma with Her-2 and Ki-67 expression(P <0.05).[Conclusion]MRI could identify PMBC and MMBC from the shape,the signal intensity,dynamic enhancement and ADC val?ue,and PMBC had distinctive MRI features. The prognosis of MMBC is worse than that of PMBC form correlation between biological prognostic factors and mucinous breast carcinoma.

Objective To investigate the diagnostic value of the texture analysis derived from conventional MR imaging in differentiating glioblastomas from solitary brain metastases. Methods Thirty-four patients with pathological diagnoses of glioblastomas and 34 patients with pathological diagnoses of solitary brain metastases were enrolled in our study. All patients underwent conventional MR imaging including axial T1WI, T2WI, fluid attenuated inversion recovery (FLAIR) and contrast-enhanced T1WI before surgery. Texture features were calculated from manually drawn ROIs by using MaZda software. The feature selection methods included mutual information (MI), Fishers coefficient, classification error probability combined with average correlation coefficients (POE+ACC) and the combination of the above three methods. These methods were used to identify the most significant texture features in discriminating glioblastomas from metastases. Then the statistical methods including raw data analysis (RDA), principal component analysis (PCA), linear discriminant analysis (LDA) and nonlinear discriminant analysis (NDA) were used to distinguish glioblastomas from metastases. The results were shown by misclassification rate. Meanwhile, two senior radiologists (who had 5 and 9 years of experience in neuroimaging diagnosis, respectively) analysed the data of the 68 patients. Chi-square test was used to compare the differences in the results between the radiologists' analysis and the texture analysis. Results In the four kinds of sequences, the texture features for differentiating glioblastomas from solitary brain metastases were mainly from T2WI which had the lowest misclassification rate, 8.82% (6/68). The misclassification rates of the feature selection methods were similar in MI, Fisher's coefficient and POE + ACC (10.29%-27.94% for MI;11.76%-44.12% for Fisher's coefficientand 8.82%-38.24% for POE+ACC). However, the misclassification rate of the combination of the three methods (8.82%-33.83% for FPM) was lower than that of any other kind of method. In the statistical methods, NDA (8.82%-11.76% ) had lower misclassification rate than RDA (26.47%-39.71% ), PCA (27.94%-39.71%) and LDA (13.24%-44.12%). Misclassification rate of the radiologists' analysis 14.71%(10/68) was higher than that of the texture analysis, but there was no statistically difference between them (χ2= 10.993, P=0.287). Conclusion Texture analysis of conventional MR imaging can provide reliably objective basis for differentiating glioblastoma from solitary brain metastasis.