文献速递神经介入CTDSA

基于CTA的机器学习能否识别血管内卒中治疗成功但仍预后不良的患者

Can CTA-Based Machine Learning Identify Patients for Whom Successful Endovascular Stroke Therapy Is Insufficient?
PubMed 文献 · AJNR Am J Neuroradiol · 2025-12-04

在多中心前瞻性登记队列中,纳入大血管闭塞急性缺血性卒中且血管内治疗成功再灌注、最终梗死体积小于30 mL的连续患者,用预训练并微调的CTA深度学习模型预测90天改良Rankin量表0-2分。该模型预测性能优于仅用临床变量的逻辑回归模型(AUROC 0.81对0.492,P=0.006)。

要点

  • 成功再灌注且最终梗死体积小于30 mL的患者中,56.2%在90天时mRS仍为3-6分。
  • 基于治疗前CTA的深度学习模型预测90天良好功能结局的AUROC为0.81。
  • 仅用临床变量的逻辑回归模型预测90天良好功能结局的AUROC为0.492。

核心结论

  • 基于CTA的机器学习模型较标准临床变量能更可靠地预测大血管闭塞急性缺血性卒中患者血管内治疗成功再灌注且最终梗死体积较小后仍出现非预期不良功能结局,可能有助于预先识别仅靠血管内治疗再灌注不足以改善临床结局的患者。

研究设计与样本

  • 多中心前瞻性登记队列中连续纳入的大血管闭塞急性缺血性卒中并接受血管内治疗且成功再灌注的患者,仅纳入最终梗死体积小于30 mL并记录90天mRS者;模型为DeepsymNet-v3预训练后微调的DSN-CTA模型,主要终点为预测90天mRS 0-2的AUROC。

诊断标准与关键数值

  • DSN-CTA模型在1542例受试者上预训练,并在48例受试者中微调与交叉验证,所有受试者血管内治疗再灌注分级为TICI 2b-3;预测90天良好mRS(0-2)的AUROC为0.81,临床变量模型为0.492,P=0.006。

临床应用与阅片要点

  • 作者认为机器学习模型可能有助于预先识别仅靠血管内治疗再灌注不足以改善临床结局的患者。

局限性

  • 原文未报告局限性。

以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。

英文原文(摘要)· 点击展开

BACKGROUND AND PURPOSE: Despite advances in endovascular stroke therapy (EST) devices and techniques, many patients are left with substantial disability, even if the final infarct volumes (FIVs) remain small. Here, we evaluate the performance of a machine learning (ML) approach by using pretreatment CTA to identify this cohort of patients that may benefit from additional interventions.

MATERIALS AND METHODS: We identified consecutive subjects with large vessel occlusion (LVO) acute ischemic stroke (AIS) who underwent EST with successful reperfusion in a multicenter prospective registry cohort. We included only subjects with FIV < 30 mL and recorded 90-day outcome (mRS). A deep learning model was pretrained and then fine-tuned to predict 90-day mRS 0-2 by using pretreatment CTA images (DeepsymNet-v3 model pretrained on radiology reports and fine-tuned on detection of unexpected clinical outcomes using brain CTA images [DSN-CTA] model). The primary outcome was the predictive performance of the DSN-CTA model compared with a logistic regression model with clinical variables, measured by the area under the receiver operating characteristic curve (AUROC).

RESULTS: The DSN-CTA model was pretrained on 1542 subjects and then fine-tuned and cross-validated with 48 subjects, all of whom underwent EST with TICI 2b-3 reperfusion. Of this cohort, 56.2% of subjects had 90-day mRS 3-6 despite successful EST and FIV < 30 mL. The DSN-CTA model showed significantly better performance than a model with clinical variables alone when predicting good 90-day mRS (AUROC 0.81 versus 0.492; P = .006).

CONCLUSIONS: The CTA-based ML model was able to more reliably predict unexpected poor functional outcome after successful EST and small FIV for patients with LVO AIS compared with standard clinical variables. ML models may identify a priori patients in whom EST-based LVO reperfusion alone is insufficient to improve clinical outcomes.

来源:AJNR Am J Neuroradiol,查看原文。版权归原作者所有。

引用

  1. Jeevarajan JA, Dong Y, Ballekere A, et al. Can CTA-Based Machine Learning Identify Patients for Whom Successful Endovascular Stroke Therapy Is Insufficient? AJNR Am J Neuroradiol. 2025. PMID: 40533350. doi:10.3174/ajnr.A8885

供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。