Scholarlink

<p>Spontaneous thalamic hemorrhage (STH) is a severe type of hemorrhagic stroke with high mortality and disability rates. Identifying key risk factors for poor outcomes is crucial. This study developed and validated a clinical-radiomics nomogram to predict 180-day outcomes in STH patients. A total of 410 STH patients from the First Affiliated Hospital of Harbin Medical University were retrospectively included, with 287 in the training cohort and 123 in the internal validation cohort. The least absolute shrinkage and selection operator (LASSO) algorithm was used to select 6 of 107 extracted CT radiomics features, which were then analyzed for multicollinearity, and a Rad-score was calculated. LASSO-Logistic regression identified four clinical risk factors for poor prognosis, which were subsequently included in multicollinearity analyses. Three models: clinical, radiomics, and clinical-radiomics nomogram were constructed and validated. Model performance was evaluated using area under the curve (AUC), decision, and calibration curves, with DeLong tests for comparisons. Univariate and multivariate logistic regression analyses were conducted separately for the conservative and surgical treatment groups to identify independent prognostic factors in each group. The clinical-radiomics nomogram, incorporating age, GCS score, mGS score, rehabilitation therapy, and Rad-score, achieved high predictive performance (training cohort AUC: 0.899; internal validation: 0.889). Decision and calibration curves confirmed its clinical utility. The combined model outperformed standalone clinical or radiomics models. Subgroup analyses revealed that the Rad-score remained an independent risk factor for poor prognosis in both the conservative and surgical treatment groups. The AUC of the combined model was 0.898 and 0.828 in the conservative and surgical treatment groups, respectively. The clinical-radiomics nomogram we developed effectively predicts 180-day poor outcomes in STH patients and demonstrates superior predictive performance compared to the clinical and radiomics models. It offers a practical tool for clinicians to assess the prognosis and guide treatment decisions for high-risk patients.</p><p><?noindent??><b>Clinical trial number</b> Not applicable.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Development and validation of a clinical-radiomics nomogram for predicting 180-day functional outcomes in patients with spontaneous thalamic hemorrhage

  • 期刊论文

  • 发表日期：

    2025年6月7日

查看全文
  • Xi Zhang,
  • Yiwei Zhang,
  • Jie Zhang,
  • Yansong Liu,
  • Shang Gao,
  • Haopeng Zhang,
  • Zhaoxin Fan,
  • Yuyang Feng,
  • Aili Gao,
  • Hongsheng Liang

摘要

Spontaneous thalamic hemorrhage (STH) is a severe type of hemorrhagic stroke with high mortality and disability rates. Identifying key risk factors for poor outcomes is crucial. This study developed and validated a clinical-radiomics nomogram to predict 180-day outcomes in STH patients. A total of 410 STH patients from the First Affiliated Hospital of Harbin Medical University were retrospectively included, with 287 in the training cohort and 123 in the internal validation cohort. The least absolute shrinkage and selection operator (LASSO) algorithm was used to select 6 of 107 extracted CT radiomics features, which were then analyzed for multicollinearity, and a Rad-score was calculated. LASSO-Logistic regression identified four clinical risk factors for poor prognosis, which were subsequently included in multicollinearity analyses. Three models: clinical, radiomics, and clinical-radiomics nomogram were constructed and validated. Model performance was evaluated using area under the curve (AUC), decision, and calibration curves, with DeLong tests for comparisons. Univariate and multivariate logistic regression analyses were conducted separately for the conservative and surgical treatment groups to identify independent prognostic factors in each group. The clinical-radiomics nomogram, incorporating age, GCS score, mGS score, rehabilitation therapy, and Rad-score, achieved high predictive performance (training cohort AUC: 0.899; internal validation: 0.889). Decision and calibration curves confirmed its clinical utility. The combined model outperformed standalone clinical or radiomics models. Subgroup analyses revealed that the Rad-score remained an independent risk factor for poor prognosis in both the conservative and surgical treatment groups. The AUC of the combined model was 0.898 and 0.828 in the conservative and surgical treatment groups, respectively. The clinical-radiomics nomogram we developed effectively predicts 180-day poor outcomes in STH patients and demonstrates superior predictive performance compared to the clinical and radiomics models. It offers a practical tool for clinicians to assess the prognosis and guide treatment decisions for high-risk patients.

Clinical trial number Not applicable.