| Objective:To retrospectively analyze the magnetic resonance imaing(MRI)signs and clinical characteristics of patients with placenta accreta spectrum(PAS)and use the optimal scale regression model to establish a quantitative scoring system for diagnosing and grading PAS,and explore the importance and quantitative assignment of MRI signs and clinical characteristics for diagnosing and grading PAS.Methods:A retrospective study was done form January 2016 to January2020 in our hospital.Total 102 patients were enrolled.Clinical data and MRI signs of 102 patients undergoing MRI examination due to clinical or ultrasound suspected of PAS.Clinical data collection:age,number of pregnancies,parities and abortions,cesarean section history,presence or absence of spontaneous delivery,other uterine cavity operations,prenatal vaginal bleeding,pathological examination,relationship between placenta and uterus in surgical records.MRI signs:placental location,the presence or absence of the placenta previa,heterogeneous signal intensity within the placenta,dark intraplacental bands on T2-weighted imaging(T2WI),abnormal dilated venous lakes within the placenta,focal thinning or the absence of myometrial wall at the placental site,placental bulge,loss of the thin T2 dark uteroplacental interface,and bladder changes.Statistical analysis was done by using SPSS version 22,single factor analysis to compare the differences between clinical characteristics and MRI signs in the non-placenta accreta spectrum(non-PAS)group,PAS group,and different implant depth of PAS groups.Then we performed the optimal scale regression analysis,the PAS grade as the dependent variable,and the above-mentioned statistically significant results as the independent variables to obtain the importance and quantification of different indicators for PAS and establish a scoring system.Results:1.A total of 102 patients were included in this study,including 24cases without placental implantation,30 cases with placental accreta(PA),39cases with placenta increta(PI),and 9 cases with placental percreta(PP).In the univariate analysis of clinical characteristics,the history of the previous cesarean section and the number of abortions≥2 were significantly different between the non-PAS group and the PAS group,P<0.05.There were no significant statistical differences between the PAS group and the non-PAS in terms of age,number of pregnancies,parities,spontaneous delivery,other uterine cavity operations and prenatal vaginal bleeding.The history of previous cesarean section was statistically significant different in the PA group,PI group and PP group,P<0.05.In the univariate analysis of MRI signs,placenta previa,placenta bulge,dark intraplacental bands on T2WI,abnormal dilated venous lakes within the placenta,loss of the thin T2 dark uteroplacental interface,focal thinning or the absence of uterin myometrial wall at the placental site,bladder changes were statistically significant in the non-PAS group,PAS group and different implant depth of PAS groups,P<0.05.The placental location,intraplacental hemorrhage,and heterogeneous signal intensity within the plac-enta were not statistically different between non-PAS group and the PAS group.2.In the optimal scale regression analysis,the five indicators of previous cesarean section history,placenta previa,placenta bulge,dark intraplacental bands on T2WI,and abnormal dilated venous lakes within the placenta were all statistically correlated with PAS and grade,P<0.01.The loss of the thin T2 dark uteroplacental interface interruption,uterine myometrium thinning,and bladder change had little effect on PAS,and there were no significant difference between the non-PAS group,PAS group,and different implant depth of PAS groups.Establish regression equation:PAS implantation degree=0.384×abnormal dilated venous lakes within the placenta+0.278×placental bulge+0.197×dark intraplacental bands on T2WI+0.128×previous cesarean section history+0.082×placenta previa,the regression equation had statistical significance,P<0.001,F=56.441,adjusted R~2 was 0.868.The impact of the above five indicators on the diagnosis of PAS were from large to small:34.1%of abnormal dilated venous lakes within the placenta,23.8%of placental bulge,12.4%of dark intraplacental bands on T2WI,7.9%of previous cesarean section history and 4.7%of placenta previa.According to the quantified value of the optimal scale regression analysis and 10 points conversion value were 0,2.5,7.8,10.when the score less than1.25,non-PAS was diagnosed,PA was diagnosed with a score of 1.25-5.15,PI was diagnosed with a score of 5.15-8.90,and PP was diagnosed with a score of>8.9.3.Comparing the PAS scoring system results with clinical or pathological results,the overall sensitivity of PAS diagnosis was 93.6%,specificity was 75%,positive predictive value was 92.4%,negative predictive value was 78.3%.The sensitivity of PA,PI,PP were 80.0%,82.1%and 77.8%,respectively,and the specificity were about 91.6%,89.1%and 95.7%.Conclusion:1.The history of previous cesarean section,placenta previa,placenta bulge,dark intraplacental bands on T2WI,abnormal dilated venous lakes within the placenta,focal thinning or the absence of myometrial wall at the placental site,loss of the thin T2 dark uteroplacental interface and bladder changes had certain value in diagnosing and grading PAS.2.The use of SPSS optimal scale regression analysis can quantify and score various indicators.The most valuable MRI signs were abnormal dilated venous lakes within the placenta,placental bulge and dark intraplacental bands on T2WI. |
