A Study On The Value Of Identifying Benign And Malignant Breast Masses Based On Differential Subsampling With Cartesian Ordering Dynamic Contrast-enhanced Magnetic Resonance Imaging And Synthetic Magnetic Resonance Imaging

Objective:To explore the enhanced features of K-space shared three-dimensional volumetric rapid dynamic enhanced magnetic resonance imaging（DISCO DCE-MRI）,the value of DISCO DCE-MRI and synthetic magnetic resonance imaging（sy MRI）derivative parameters in identifying benign and malignant breast masses,and to evaluate the diagnostic efficacy and clinical utility of multiparameter models.Method:A retrospective analysis was performed on 58 patients with mass-like enhanced lesions（BI-RADS category 3 and above,including 24 benign masses and 34malignant masses）who underwent DISCO dynamic enhancement and MAGIC sequence scan in the Second Affiliated Hospital of Nanchang University from June2022 to March 2023.Taking pathological results as the gold standard,the enhancement characteristics（morphology,edge,internal enhancement,time-signal intensity curve）and derivation parameters of DISCO DCE-MRI(MS,BAT,K^trans,K_ep,V_e)and relaxation time（T1,T2,PD,T1+,T2+,PD+）between benign and malignant groups were compared and analyzed.Based on the statistically different variables in univariate analysis,multi-parameter models was established.the calibration curve was used to evaluate the fit of the model.and the receiver operating characteristic curve（ROC）was used to analyze the diagnostic efficacy of each model to identify benign and malignant breast masses.The Delong test was used to compare the differences in diagnostic efficacy of each model,the clinical decision curve（DCA）was used to evaluate the clinical practicability of each model.Outcome:（1）In univariate analysis,there were statistical differences in morphology,edge,internal reinforcement and time-signal intensity curve（TIC）between benign and malignant groups（P<0.05）.The DISCO DCE-MRI-derived parameters MS,K^trans,and K_epin the malignant group were higher than those in the benign group,and the differences were statistically significant（Z=-3.821,4.642,and-4.531,P<0.001）.There was no statistical difference in BAT and V_ebetween the two groups（Z=-2.037,t=1.545,P=0.052,0.129）.Among MAGIC-derived parameters,except for T2（t=2.542,P=0.015）and T2+（t=3.437,P=0.001）,there was no statistical difference in other parameters（T1,PD,T1+,PD+）（P>0.05）.（2）Diagnostic performance and clinical validity of multiple models:Based on the statistically significant variables of the above differences,seven models were established:model1（including morphology,edge,internal reinforcement and TIC）,model2(including MS,K^trans,K_ep),model3（including T2 and T2+）,model4(including MS,K^trans,K_ep,T2 and T2+),model5(including morphology,edge,internal enhancement,TIC,MS,K^trans,K_ep),model6（including morphology,edges,internal reinforcement,TIC,T2,and T2+）,model7(including morphology,edges,internal reinforcement,TIC,MS,K^trans,K_ep,T2,and T2+).The calibration of each model was good,and the calibration curve showed that Model 7 had the best good fit.According to the ROC curve analysis,the AUC of the combined models（model4,model5,model6,and model7）were 0.964,0.972,0.957,and 0.988,respectively,which were larger than those of the single models（model1,model2,and model3）.The Delong test showed that there were statistically significant differences in the diagnostic efficacy between model3（AUC=0.756）and model4,model5,model6,and model7（P=0.001,P=0.002,P=0.001,P<0.001）,and model1（AUC=0.911）,model2（AUC=0.912）were statistically different from the AUC of model7（P=0.048,P=0.02）.There were no statistical differences among the other models.The decision curve（DCA）showed that model7 had the greatest clinical net benefit.Conclusion:（1）Compared with benign masses,the shape and edges of malignant breast lumps are more irregular,the internal strengthening is more uneven,TIC is more common in plateau type（type II）or outflow type（type III）,MS,K^trans,K_epof malignant mass are larger,and T2 and T2+are smaller.（2）Each multiparameter model has certain diagnostic value and clinical utility,and the single feature sy MRI parameters（T2,T2+）combined with DISCO DCE-MRI can significantly improve the diagnostic performance.Among them,the combination of sy MRI parameters and DISCO DCE-MRI enhancement characteristics（morphology,edge,internal reinforcement,TIC）and derivative parameters(MS,K^trans,K_ep)has the best diagnostic efficiency,and the clinical net benefit is the largest,which can provide a more accurate and objective non-invasive examination method for clinical identification of benign and malignant breast masses,which is helpful to assist clinical decision-making.