基于临床及常规MRI征象Logistic回归模型列线图诊断胎盘植入

目的 评估基于临床特征及常规MRI征象的Logistic回归模型列线图预测胎盘植入的价值。方法 回顾性分析47例临床疑诊胎盘植入患者，其中18例发生胎盘植入、29例未植入，比较有无胎盘植入患者临床特征及MRI表现差异。采用Logistic回归方法构建预测胎盘植入模型，并制作列线图，绘制校正曲线，评估模型的预测效能。结果 有无胎盘植入患者孕次、剖宫产次数、胎盘位置、胎盘下血管异常差异均有统计学意义（P均<0.05），而年龄、流产史、阴道流血史、胎盘与子宫肌层分界面中断、胎盘局部膨出及胎盘信号差异均无统计学意义（P均>0.05）。剖宫产次>1及胎盘下血管异常是胎盘植入的独立危险因素，二者联合预测胎盘植入的准确率、灵敏度、特异度、阴性预测值分别为82.98%、77.78%、86.21%、86.21%，均高于单一参数（P均<0.05）；校正曲线显示模型预测胎盘植入概率与实际概率的一致性较好。结论 基于临床特征及常规MRI征象的Logistic回归模型列线图可作为术前预测胎盘植入的辅助工具。

Logistic regression model nomogram based on clinical and conventional MRI characteristics for diagnosis of placental accreta

Objective To observe the value of Logistic regression model nomogram based on clinical and conventional MRI characteristics for diagnosis of placental accreta. Methods Data of 47 patients with clinically suspected placental accreta were retrospectively analyzed. According to pathological results, 18 cases had placental accreta, while 29 cases had no placental accreta. Clinical characteristics and MRI manifestations of patients with or without placental accreta were compared and analyzed. Logistic regression method was used to construct the predictive model of placental accreta and made the nomogram, then calibration curve was drawn to evaluate the predictive performance of this model. Results There were significant differences of the times of pregnancy and cesarean section, placental position and subplacental vascularity were found between patients with or without placental accreta (all P<0.05), but not of age, the times of abortion, vaginal bleeding, interruption of the interface, local expansion of placenta nor placental signals (all P>0.05). The independent risk factors for placental accrete were the times of cesarean section >1 and subplacental vascularity. The accuracy, sensitivity, specificity and negative predictive value of combination of the above 2 parameters was 82.98%, 77.78%, 86.21% and 86.21%, respectively, all higher than those of each single parameter (all P<0.05). The calibration curve showed good agreement between the probability of placental accreta predicted by this model and the actual probability. Conclusion Logistic regression model nomogram based on clinical and conventional MRI characteristics could be used as an auxiliary tool for preoperative prediction of placental accreta.