检验医学 ›› 2025, Vol. 40 ›› Issue (12): 1146-1152.DOI: 10.3969/j.issn.1673-8640.2025.12.002

• 论著 • 上一篇    下一篇

基于外周血淋巴细胞亚群特征构建抑郁障碍患者预后列线图预测模型

高子康1, 陈晶2   

  1. 1.绍兴市第七人民医院心身障碍科,浙江 绍兴 312000
    2.绍兴市第七人民医院精神科,浙江 绍兴 312000
  • 收稿日期:2024-09-05 修回日期:2025-04-17 出版日期:2025-12-30 发布日期:2025-12-26
  • 作者简介:高子康,男,1996年生,学士,医师,主要从事精神疾病诊治工作。
  • 基金资助:
    浙江省医药卫生科技计划项目(2023KY1272)

Construction of a nomogram prediction model for the prognosis of patients with depressive disorder based on the characteristics of peripheral blood lymphocyte subsets

GAO Zikang1, CHEN Jing2   

  1. 1. Department of Psychosomatic Disorder,Shaoxing Seventh People's Hospital,Shaoxing 312000,Zhejiang,China
    2. Department of Psychiatry,Shaoxing Seventh People's Hospital,Shaoxing 312000,Zhejiang,China
  • Received:2024-09-05 Revised:2025-04-17 Online:2025-12-30 Published:2025-12-26

摘要:

目的 基于外周血淋巴细胞亚群构建抑郁障碍患者预后列线图预测模型,并评估其临床适用性。 方法 选取2021年5月—2022年6月绍兴市第七人民医院抑郁障碍患者166例作为建模集。所有患者均接受常规抗抑郁治疗,出院后随访1年,末次随访时采用大体功能评定量表(GAS)评定其预后情况。将GAS评分≤60分的77例患者纳入预后不良组,GAS评分>60分的89例患者纳入预后良好组。另选取2022年10月—2023年6月绍兴市第七人民医院抑郁障碍患者72例作为验证集,用于列线图预测模型的外部验证。收集所有患者的临床资料,并检测外周血T淋巴细胞亚群百分比(CD3+%、CD4+%、CD8+%)和硫酸脱氢表雄酮(DHEAS)水平,计算CD4/CD8比值。采用受试者工作特征(ROC)曲线评价各项指标判断抑郁障碍患者预后不良的效能。采用二元Logistic回归分析评估抑郁障碍患者预后的影响因素。采用校准曲线和决策曲线对列线图预测模型进行内部验证和效能评估。 结果 建模集预后不良组和预后良好组性别、CD3+%、CD4+%、CD8+%、CD4/CD8比值和DHEAS水平差异均有统计学意义(P<0.001),其他临床资料2个组之间差异均无统计学意义(P>0.05)。性别、CD3+%、CD4+%、CD8+%、CD4/CD8比值、DHEAS均为抑郁障碍患者预后不良的危险因素(P<0.05)。基于回归方程构建预测抑郁障碍患者预后的列线图模型,该模型判断抑郁障碍患者预后不良的曲线下面积(AUC)为0.964,C-index为0.963;模型的拟合度较好(Hosmer-Lemeshow χ2=11.507,P=0.175);当阈值概率>0.14时,可获得临床净收益,最大净收益为0.30。在验证集中,列线图预测模型判断抑郁障碍患者预后不良的AUC为0.920,模型的拟合度较好(Hosmer-Lemeshow χ2=6.782,P=0.553)。 结论 基于性别、CD3+%、CD4+%、CD8+%、CD4/CD8比值和DHEAS构建的列线图预测模型可有效评估抑郁障碍患者的预后情况,且临床适用性较好。

关键词: 淋巴细胞亚群, 抑郁障碍, 预后, 列线图模型

Abstract:

Objective To construct a nomogram prediction model for the prognosis of patients with depressive disorder based on peripheral blood lymphocyte subsets,and to evaluate its clinical applicability. Methods A total of 166 patients with depressive disorder from Shaoxing Seventh People's Hospital from May 2021 to June 2022 were enrolled as modeling set. All the patients received conventional antidepressant treatment,and they were followed up for 1 year after discharge. The prognosis of patients with depressive disorder was evaluated using the Global Assessment Scale(GAS) at the last follow-up. Totally,77 patients with a GAS score ≤60 were classified in poor prognosis group,and 89 patients with a GAS score >60 were classified in good prognosis group. Another 72 patients with depressive disorder from Shaoxing Seventh People's Hospital from October 2022 to June 2023 were enrolled as validation set for external validation of the nomogram prediction model. The clinical data were collected,and the percentages of peripheral blood T lymphocyte subset percentages(CD3+%,CD4+%,CD8+%) and dehydroepiandrosterone sulfate(DHEAS) levels were determined. The CD4/CD8 ratio was calculated. Receiver operating characteristic(ROC) curve was used to evaluate the efficacy of each index in predicting poor prognosis in patients with depressive disorder. Binary Logistic regression analysis was used to evaluate the influencing factors of prognosis in patients with depressive disorder. Calibration curves and decision curves were used to internally validate and evaluate the performance of the nomogram prediction model. Results There was statistical significance in gender,CD3+%,CD4+%,CD8+%,CD4/CD8 ratio and DHEAS level between poor prognosis group and good prognosis group in the modeling set(P<0.001),while there was no statistical significance in the other clinical data between the 2 groups(P>0.05). Gender,CD3+%,CD4+%,CD8+%,CD4/CD8 ratio and DHEAS were all risk factors for poor prognosis in patients with depressive disorder(P<0.05). A nomogram prediction model for predicting the prognosis of patients with depressive disorder was constructed based on the regression equation. The area under curve(AUC) of the nomogram prediction model for predicting poor prognosis in patients with depressive disorder was 0.964,and the C-index was 0.963. The model had a good fit(Hosmer-Lemeshow χ2=11.507,P=0.175). When the threshold probability was >0.14,clinical net benefit could be obtained,and the maximum net benefit was 0.30. In the validation set,the AUC of the nomogram prediction model for predicting poor prognosis in patients with depressive disorder was 0.920,and the model had a good fit(Hosmer-Lemeshow χ2=6.782,P=0.553). Conclusions The nomogram prediction model constructed based on gender,CD3+%,CD4+%,CD8+%,CD4/CD8 ratio and DHEAS can effectively evaluate the prognosis of patients with depressive disorder,and has good clinical applicability.

Key words: Lymphocyte subset, Depressive disorder, Prognosis, Nomogram model

中图分类号: