Hsiao Yu-Yang, Fu Hung-Chun, Wu Chen-Hsuan, Lan Jui, Ou Yu-Che, Tsai Ching-Chou, Lin Hao
Department of Obstetrics and Gynecology, Kaohsiung Chang Gung Memorial Hospital and Chang Gung University College of Medicine, Kaohsiung 83341, Taiwan.
Department of Obstetrics and Gynecology, Chia-Yi Chang Gung Memorial Hospital, Chia-Yi 61363, Taiwan.
Diagnostics (Basel). 2022 Mar 23;12(4):790. doi: 10.3390/diagnostics12040790.
Background: Previous studies have shown that loss of progesterone receptor (PR) in endometrial cancer (EC) is associated with poor outcomes. Evaluating lymph node metastasis (LNM) is essential, especially before surgical staging. The aim of this study was to investigate the role of PR expression and other clinicopathological parameters in LNM and to develop a prediction model. Methods: We retrospectively evaluated endometrioid-type EC patients treated with staging surgery between January 2015 and March 2020. We analyzed PR status using immunohistochemical staining, and the expression was quantified using the H-score. We identified optimal cut-off values of H-score and CA125 for predicting LNM using receiver operating characteristic curves, and used stepwise multivariate logistic regression analysis to identify independent predictors. A nomogram for predicting LNM was constructed and validated using bootstrap resampling. Results: Of the 310 patients evaluated, the optimal cut-off values of PR H-score and CA125 were 162.5 (AUC 0.670, p = 0.001) and 40 U/mL (AUC 0.739, p < 0.001), respectively. Multivariate analysis showed that CA125 ≥ 40 U/mL (OR: 8.03; 95% CI: 3.44−18.77), PR H-score < 162.5 (OR: 5.22; 95% CI: 1.87−14.60), and tumor grade 2/3 (OR: 3.25; 95% CI: 1.33−7.91) were independent predictors. These three variables were incorporated into a nomogram, which showed effective discrimination with a concordance index of 0.829. Calibration curves for the probability of LNM showed optimal agreement between the probability as predicted by the nomogram and the actual probability. Our model gave a negative predictive value and a negative likelihood ratio of 98.4% and 0.14, respectively. Conclusions: PR H-score along with tumor grade and CA125 are helpful to predict LNM. In addition, our nomogram can aid in decision making with regard to lymphadenectomy in endometrioid-type EC.
既往研究表明，子宫内膜癌（EC）中孕激素受体（PR）缺失与预后不良相关。评估淋巴结转移（LNM）至关重要，尤其是在手术分期之前。本研究的目的是探讨PR表达及其他临床病理参数在LNM中的作用，并建立一个预测模型。方法：我们回顾性评估了2015年1月至2020年3月期间接受分期手术治疗的子宫内膜样型EC患者。我们采用免疫组织化学染色分析PR状态，并使用H评分对表达进行量化。我们使用受试者工作特征曲线确定预测LNM的H评分和CA125的最佳截断值，并使用逐步多因素逻辑回归分析确定独立预测因素。构建了一个预测LNM的列线图，并使用自助重抽样进行验证。结果：在评估的310例患者中，PR H评分和CA125的最佳截断值分别为162.5（AUC 0.670，p = 0.001）和40 U/mL（AUC 0.739，p < 0.001）。多因素分析显示，CA125≥40 U/mL（OR：8.03；95%CI：3.44−18.77）、PR H评分<162.5（OR：5.22；95%CI：1.87−14.60）和肿瘤分级2/3（OR：3.25；95%CI：1.33−7.91）是独立预测因素。这三个变量被纳入一个列线图，该列线图显示出有效的区分能力，一致性指数为0.829。LNM概率的校准曲线显示，列线图预测的概率与实际概率之间具有最佳一致性。我们的模型的阴性预测值和阴性似然比分别为98.4%和0.14。结论：PR H评分以及肿瘤分级和CA125有助于预测LNM。此外，我们的列线图可辅助子宫内膜样型EC患者淋巴结切除术的决策制定。
