Establishment Of A Quantitative Scoring System For The Prediction Of The Severity Of Placenta Accrete Spectrum In Women With Placenta Previa

Background:Placenta accrete spectrum(PAS)disorder is the leading cause of disastrous bleeding and perinatal hysterectomy.The incidence of PAS increased in recent decades and will continue to rocket in the future decades.It is a serious challenge for obstetricians.There is no optimal strategy for treating it.The clinical outcomes could be improved by an accurate diagnosis and evaluation based on clinical and ultrosounographic features.Recently,researchers trying to develop a scoring system to improve the diagnostic accuracy employing multiple indicators.However,there are still two issues that need to be resolved: how to assign a reasonable value to each parameter in scoring system and how to distinguish the severities of PAS.Therefore,it is of great clinical significance to establish a scoring system for the purpose of accuracy diagnosis.Objectives:1.Aim to identify the ultrasonographic characteristics and clinical features that colsely related to depth of PAS by retrospective analysis.2.Aim to develop a quantitative scoring system by weighting each feature in the overall risk of PAS,and hereby to predict the severities of PAS by a prospective study.Materials and Methods:1.We retrospectively analysed 571 women with placenta previa in our hospital,between Jan 1,2012 to May 31,2016.(1)Patients were divided into two groups: non-deep PAS group(without PAS and accreta)and deep PAS group(increta and percreta)according to the surgical findings and pathological examination.(2)Clinical features and ultrasonographic characteristics that colsely related to deep PAS were analysed.2.Establishment of a quantitative scoring systemThe study was conducted on patients with placenta previa who received operation in the department of obstetrics of the Second Clinical College of the Army Medical University from June 12,2016 to June 30,2018.(1)A pre-scoring scale was established based on our retrospective results.(2)Patients were divided into four groups:without PAS,accreta,increta and percreta group,according to the surgical findings and pathological examination.(3)The parameters that strongly associated with the severity of PAS determined by analyzing and comparing the clinical features and ultrasounographic imagies.(4)The optimal scale regression model was employed to weight each parameter in different levels,then to establish a quantitative scoring system to predict to the severity of PAS.Statistical analysisIBM SPSS 22.0 was employed.Statistical analysis methods including univariate,multivariate and ROC curve analysis were adopted,and positive predictive value(PPV),negative predictive value(NPV)and false positive rate which showed the predictive accuracy were calculated.Results:1.Among 571 patients with placenta previa,493 and 78 were diagnosed of non-deep PAS and deep PAS.Compared with non-deep PAS group,the proportions of cases with heavy bleeding and massive transfusion during operation,further intervention requirement,bladder injury and hysterectomy in the deep PAS group were significantly higher;and the operation time in the deep PAS group was longer.2.The 12 variables which may associate with PAS were analysed by univariate and multivariate analysis.After removing the potential confounding variables,the four features were strongly associated with deep PAS: loss of hypoechoic retroplacental zone,presence of placental lacunas,hypervascularity at the uterus-bladder interface,and previous cesarean sections(CS),with the AUC of 0.79-0.83。3.Among 392 patients with placenta previa,232(60%),79(20%),53(13%),28(7%)were diagnosed of without PAS,accreta,increta and percreta respectively.There were no significant differences in maternal age,number of previous abortions,and prenatal vaginal bleeding between groups.The number of gravidity,the number of previous vaginal deliveries,and the number of previous CS were significantly different.4.All of the seven adopted parameters were significant different between groups,which were as follows: placental lacunas,placenta main attachment location,hypervascularity at the uterus-bladder interface,loss of bladder line,cervical length,the thickness of uterine wall and hypoechoic retroplacental zone,and the number of previous CS.5.Placenta main attachment location and the length of the cervix showed no significant difference(P > 0.05)in the multivariate analysis,which showed there is no effects on PAS severity.The other five variables were adopted into the optimal scale regression analysis.The weight of each variable in prediction severity of PAS was as follows(percentages): placental lacunas(25.7%),hypervascularity at the uterus-bladder interface(19.9%),the thickness of uterine wall and hypoechoic retroplacental zone(26.8%),bladder line(15.8%),and the previous CS(11.8%).The values of each variable in different levels weighted by the regression model,and then were converted to a 10-point system.The sum of these values yield the predictive score for degree of PAS.The predicted cut-off values for without PAS,accreta,increta and percreta were <2.25,2.25-6.2,6.2-8.95,and >8.95 points,respectively.The gaps between different levels of variables are not equal.6.The positive predictive value(PPV),negative predictive value(NPV),and false positive rate of the scoring system were: 96.68%,95.44%,and 3.32%,respectively.Conclusion:The number of CS and ultrasounographic characteristics are significantly associated with depth of PAS in women with placenta previa.The quantitative scoring system can effectively predict the severity of PAS for women with placenta previa.It may offer a new approaches to plan a treatment more targeted.