J Coll Physicians Surg Pak. 2026 Jan;36(1):43-50. doi: 10.29271/jcpsp.2026.01.43.
ABSTRACT
OBJECTIVE: To develop and validate a nomogram model based on LASSO-logistic regression to predict the 90-day functional prognosis of acute ischaemic stroke (AIS) patients after endovascular therapy (EVT).
STUDY DESIGN: A descriptive study. Place and Duration of the Study: Department of Radiology, Chongqing University Three Gorges Hospital, Chongqing, China, from January 2022 to October 2024.
METHODOLOGY: AIS patients who underwent EVT were retrospectively analysed. Patients were randomly split into the training (70%) and validation (30%) sets. Clinical, laboratory, and imaging data were collected. Independent prognostic factors were identified using LASSO regression (10-fold cross-validation to select optimal λ) followed by multivariate logistic regression. A nomogram was constructed by assigning scaled scores to each predictor based on its regression coefficients, summing these scores to generate a total score, and mapping the total score to predicted probabilities of 90-day functional outcomes.
RESULTS: Through LASSO-logistic regression analysis, ICH-24h-PO (OR = 5.879), THUAS (1.005), HSI (1.013), NLR (1.069), age (1.049), gender (2.553), and ASPECTS (0.769) were identified as independent influencing factors. The nomogram exhibited strong discriminative performance, achieving AUC values of 0.889 for the training cohort and 0.878 for the validation set. Probability calibration plots showed excellent concordance between predicted and observed event rates. The Hosmer-Lemeshow test results (p >0.05) and Brier score metrics (training: 0.134; validation: 0.146) further confirmed the model's reliability. Clinical decision curves demonstrated pronounced net benefits within the defined threshold probability ranges (training set: 0.1-1.0; validation set: 0.03-1.0).
CONCLUSION: A robust nomogram model integrating multi-dimensional data was developed to predict the 90-day functional outcomes in AIS patients after EVT, thereby serving as a promising tool for personalised decision-making and outcome optimisation.
KEY WORDS: Acute ischaemic stroke, Nomogram, Endovascular therapy, Logistic regression, Functional outcome.
PMID:41792065 | DOI:10.29271/jcpsp.2026.01.43