Li Wei, Fang Kun, Chen Jiaren, Deng Jian, Li Dan, Cao Hong
Department of Breast and Thyroid Surgery, The Second Affiliated Hospital, Hengyang Medical School, University of South China, Hengyang, China.
Department of Surgery, Yinchuan Maternal and Child Health Hospital, Yinchuan, China.
Front Genet. 2023 May 30;14:1133495. doi: 10.3389/fgene.2023.1133495. eCollection 2023.
We aimed to explore prognostic risk factors in patients with malignant phyllodes tumors (PTs) of the breast and construct a survival prediction model. The Surveillance, Epidemiology, and End Results database was used to collect information on patients with malignant breast PTs from 2004 to 2015. The patients were randomly divided into training and validation groups using R software. Univariate and multivariate Cox regression analyses were used to screen out independent risk factors. Then, a nomogram model was developed in the training group and validated in the validation group, and the prediction performance and concordance were evaluated. The study included 508 patients with malignant PTs of the breast, including 356 in the training group and 152 in the validation group. Univariate and multivariate Cox proportional hazard regression analyses showed that age, tumor size, tumor stage, regional lymph node metastasis (N), distant metastasis (M) and tumor grade were independent risk factors for the 5-year survival rate of patients with breast PTs in the training group ( < 0.05). These factors were used to construct the nomogram prediction model. The results showed that the C-indices of the training and validation groups were 0.845 (95% confidence interval [CI] 0.802-0.888) and 0.784 (95% CI 0.688-0.880), respectively. The calibration curves of the two groups were close to the ideal 45° reference line and showed good performance and concordance. Receiver operating characteristic and decision curve analysis curves showed that the nomogram has better predictive accuracy than other clinical factors. The nomogram prediction model constructed in this study has good predictive value. It can effectively assess the survival rates of patients with malignant breast PTs, which will aid in the personalized management and treatment of clinical patients.
我们旨在探讨乳腺恶性叶状肿瘤（PTs）患者的预后危险因素，并构建生存预测模型。利用监测、流行病学和最终结果数据库收集2004年至2015年乳腺恶性PTs患者的信息。使用R软件将患者随机分为训练组和验证组。采用单因素和多因素Cox回归分析筛选出独立危险因素。然后，在训练组中建立列线图模型，并在验证组中进行验证，评估其预测性能和一致性。该研究纳入了508例乳腺恶性PTs患者，其中训练组356例，验证组152例。单因素和多因素Cox比例风险回归分析显示，年龄、肿瘤大小、肿瘤分期、区域淋巴结转移（N）、远处转移（M）和肿瘤分级是训练组乳腺PTs患者5年生存率的独立危险因素（<0.05）。这些因素被用于构建列线图预测模型。结果显示，训练组和验证组的C指数分别为0.845（95%置信区间[CI]0.802-0.888）和0.784（95%CI0.688-0.880）。两组的校准曲线均接近理想的45°参考线，表现出良好的性能和一致性。受试者工作特征曲线和决策曲线分析曲线显示，列线图比其他临床因素具有更好的预测准确性。本研究构建的列线图预测模型具有良好的预测价值。它可以有效地评估乳腺恶性PTs患者的生存率，这将有助于临床患者的个性化管理和治疗。
