Online Dynamic Nomogram for Predicting 90‐Day Prognosis of Patients With Primary Basal Ganglia Cerebral Hemorrhage After Microscopic Keyhole Craniotomy for Hematoma Removal

ABSTRACT

Objective

Primary basal ganglia cerebral hemorrhage (PBGCH) is the most common type of hypertensive intracerebral hemorrhage. Microscopically removing the hematoma via keyhole or microbone window craniotomy remains the most common surgical method in many hospitals across China for treating cases of primary basal ganglia hemorrhage exceeding 30 mL. The aim of this study was to establish a new practical evaluation system based on preoperative clinical and imaging factors to predict the short‐term prognosis of PBGCH after microscopic keyhole craniotomy for hematoma removal (MKCHR), providing a reference for clinicians and patients' families in deciding whether to proceed with surgery.

Methods

A retrospective analysis was performed on 74 cases of PBGCH treated with MKCHR. Patient prognosis was assessed at 90 days postsurgery using the modified Rankin Scale. This study employed R software to conduct both univariate and multivariate logistic regression analyses aimed at identifying preoperative factors that influence short‐term prognosis following MKCHR. Additionally, a web‐based interactive nomogram was developed to forecast outcomes for PBGCH patients receiving MKCHR treatment. Model robustness was gauged using the concordance index ( C‐index) and receiver operating characteristic (ROC) curve. Internal validation involved bootstrap resampling and calibration. Clinical utility was assessed via decision curve analysis (DCA), clinical impact curve (CIC), and net reduction interventions (NRI).

Results

Glasgow Coma Scale (GCS) score ≤ 6, hemorrhagic volume > 102 mL, brain herniation, age > 58 years ( p < 0.05) were independent risk factors for poor prognosis after MKCHR. The online dynamic nomogram website is https://sjwkalg.shinyapps.io/DynNomapp/. The model's C‐index and area under the ROC are both 0.899 (95% confidence interval [CI], 0.817–0.980). Following 1000 bootstrap resamples, the calibration curve indicates that the dynamic nomogram's predicted values closely match the observed values.

The models of DCA, CIC, and NRI show good clinical application.

Conclusion

The online dynamic nomogram developed in this study demonstrates high predictive accuracy. This platform is characterized by its noninvasive and convenient nature, which facilitates the formulation of clinical treatment strategies. It offers a reliable data reference for preoperative surgical decision‐making in patients with PBGCH, thereby aiming to achieve beneficial outcomes.

Abstract

In this study, we created the first online dynamic nomogram to predict the 90‐day prognosis of patients with primary basal ganglia cerebral hemorrhage following microscopic keyhole craniotomy for hematoma removal. This tool assists in surgical strategy selection and postoperative prognosis evaluation. A graphic abstract was created using Figdraw.com.