The Study Of The Correlation Between Multi-parameters Of Ultrasound And Micro Vessel Density In Peripheral Lung Cancer

Objective:The aim of this study was to investigate the correlation between ultrasound parameters and the MVD in peripheral lung cancer,including the size of the lesion,blood flow grades,parameters of contrast-enhanced ultrasound.And then,we made a conjoint analysis of these parameters that correlate with the MVD in peripheral lung cancer,to find if this method could improve the ability of ultrasound to reflect the MVD in peripheral lung cancer,to lay a foundation for the follow-up research.Method:From October 2019 to February 2021,we made ultrasound and contrast-enhanced ultrasound（CEUS）exam for patients who came to our hospital due to space-occupying lesions in lung,to gain the multiple parameters of ultrasound and CEUS,and then we made ultrasound-guided biopsy to the lesions,to make pathological as well as CD34 immunohistochemical exam,and then,we screened out the ultrasound and pathological data of cases which was confirmed lung cancer by pathology exam.The methods are as follows:Firstly,we made routine ultrasound examination and record the size and blood grade of the lesion,the blood flow grade was divided into 0 to III grades refer to Adle′rs methods.And then,we made contrast-enhanced ultrasound（CEUS）exam,recording and saving video for 5 minutes,then starting the MFE procedure,the MFE pattern was mainly classified“dead wood”,“vascular”,“cotton”and undefined patterns.At the same time,the region of interested（ROI）was chosen and we thus gained the quantitative parameters of CEUS,such as Arrive Time（AT）,Time to Peak（TTP）,Peak Intensity（PI）,Ascending Slope（AS）,Descending Slope（DS）,Decreasing Time from Peak to One Half（DT/2）,Area Under the Curve（AUC）.Next,we made ultrasound or contrast-enhanced ultrasound guided lung biopsy,the target site was consistent with ROI of CEUS.We gain the sample by Bard automatic biopsy needle,the sample were then taken away to the Pathology Department to make pathological examination and CD34 immunohistochemical examination,the MVD counts refer to Weidner′s methods.The size,blood flow grades,MFE patterns,and quantitative parameters of CEUS were analyzed to find the correlation of MVD between them.And then we combined the parameters which correlates with MVD,to find to find if this method could improve the ability of ultrasound to reflect the MVD of peripheral lung cancer by using multiple linear regression.Results:A total of 108 cases were included in the study ultimately,among them,there were 84 males and 24 females.The pathological result and the number of the cases were listed as follows:adenocarcinoma（54 cases）,squamous cell carcinoma（35cases）,adeno-squamous carcinoma（2 cases）,small cell lung cancer（7 cases）,neuroendocrine carcinoma（7 cases）,lymphoepithelioma-like carcinoma（3 cases）.1.All the case were divided into two gro≤u5pcs（m group and>5cm group）according to the maximum diameter of the lesion.Kendall’s tau-b test showed that there were not statistical significant different between the maximum diameter of the lesion and MVD,Kendall’s tau-b=-0.014,P=0.858;A Mann-Whitney U test was run to determine if there were differences in MVD between≤5cm group and>5cm group,the result showed that there were not statistical significant different between the two groups,U=1268.5,Z=-0.179,P=0.858.2.The Spearman correlation analyze showed that there were positive correlation between the blood flow grades and MVD（r_s=0.349,P<0.001）.Simple Liner Regression showed that there were liner correlation between blood grade and MVD,F（1,105）=14.553（P<0.001）,regression equation:MVD=18.349+4.990×blood flow grade（r²=0.122）.3.The MVD between the three groups（dead wood pattern,vascular pattern and cotton pattern）were compared pairwise by Mann-Whitney U test.MVD for cotton pattern（mean rank=42.02）and vascular pattern（mean rank=48.19）were not statistically significantly different（U=852.00,Z=-1.126,P=0.26）;MVD for cotton pattern（mean rank=36.54）and dead wood pattern（mean rank=22.17）were statistically significantly different（U=228.00,Z=-2.780,P<0.016）;MVD for vascular pattern（mean rank=35.84）and dead wood pattern（mean rank=19.44）were statistically significantly different（U=179.00,Z=-3.291,P<0.016）;（P<0.016 was considered the difference was statistical significant）.Taking dead wood pattern as reference,a Multiple Liner Regression was run to determine if there were differences between MFE and MVD,the result showed that the model was statistical significant different,F（2,103）=5.778（P<0.05）;liner equation:MVD=20.50+（4.543×cotton pattern）+（3.960×vascular pattern）（r²=0.101）.4.The Spearman correlation analyze showed that PI and AUC were positively correlate with MVD,r_s（PI）=0.537,P<0.001,r_s（AUC）=0.325,P<0.05;whereas AT,TTP,AS and DS were not correlate with MVD.Simple Liner Regression showed that PI and AUC have liner correlation with MVD,F _PI（1,106）=37.80（P<0.001）、F _AUC（1,106）=12.23（P<0.05）,regression equation:MVD_PI=-7.753+1.024×PI（r²=0.263）,MVD _AUC=10.370+0.004（r²=0.103）.5.By means of above mentioned statistical methods,we have selected four parameters（blood grade,MFE patterns,PI and AUC）which associated with MVD,then we made a conjoint analysis by Multiple Liner Regression,and the result showed that the regression model was statistical significant,F（5,100）=11.566（P<0.001）;regression equation:MVD=-10.004-（0.003×AUC）+（1.441×PI）+（2.344×blood flow grade）+（3.856×Cotton pattern）+（2.958×Vascular pattern）（r²=0.366）.Conclusion1.Among the multi-parameters of ultrasound,blood flow grades,MFE patterns,PI and AUC were correlate with the MVD in peripheral lung cancer,and all of them may reflect MVD of peripheral lung cancer to a certain degree.2.After the conjoint analysis of blood grades,MFE patterns,PI and AUC,the result showed that,the r²of the conjoint analysis was higher than the parameters which was used independently,indicating that the conjoint analysis could improve the ability of ultrasound to reflect the MVD,and probably it could be used for assess the MVD in peripheral lung cancer non-invasively.