Stroke: Machine learning algorithms significantly improve the accuracy of CRP levels in predicting the prognosis of subarachnoid hemorrhage

Subarachnoid hemorrhage (SAH) refers to brain lesions or bottom surface of the brain blood vessels rupture, blood flow directly into a clinical syndrome caused by subarachnoid space, accounting for acute cerebral stroke 10%, it is a very serious common Disease
.

The WHO survey shows that the incidence rate in China is about 2.

Subarachnoid hemorrhage (SAH) refers to brain lesions or bottom surface of the brain blood vessels rupture, blood flow directly into a clinical syndrome caused by subarachnoid space, accounting for acute cerebral stroke 10%, it is a very serious common Disease

SAH The most common cause of SAH is intracranial aneurysm .

Bleeding is a serious complication of acute SAH, the mortality is about 50% or so
.

The risk of rebleeding within 24 hours after bleeding is the greatest, and the points of rebleeding within 1 month of

Bleeding is a serious complication of acute SAH, the mortality is about 50% of bleeding is a serious complication of acute SAH, the mortality is about 50% or so

CRP (C-reactive protein) is related to prognosis, CRP (C-reactive protein) is related to prognosis,

In order to assess whether CRP is an independent predictor of post-aSAH outcomes, develop new prognostic models that include CRP, and test whether these models can be improved by applying machine learning.
Recently, experts from the Department of Neurosurgery, South Cape Town General Hospital in the United Kingdom have carried out related research.
The results of the study were published in the journal Stroke
.

The researchers conducted an individual patient-level analysis of the data of patients within 72 hours after SAH in the previous two studies
.

A series of machine learning methods including logistic regression, random forest and support vector machine are used to evaluate the relationship between CRP and the modified Rankin scale (mRS)

A series of machine learning methods including logistic regression, random forest and support vector machines are used to evaluate the relationship between CRP and the modified Rankin scale (mRS)

CRP used in SAHIT can improve the accuracy of prognosis

CRP used in SAHIT can improve the accuracy of prognosis

The results showed that a total of 1117 patients were included in the analysis
.

In general, CRP on the first day after coma is an independent predictor of outcome

CRP on the first day after coma is an independent predictor of outcome

However, when the support vector machine (SVM) is further used (AUC=0.

Schematic diagram of SVM working principle

Schematic diagram of SVM working principle

It can be seen that CRP is an independent predictor of the prognosis of SAH
.

Incorporating it into a prognostic model can improve performance.

It can be seen that CRP is an independent predictor of the prognosis of SAH
.
It can be seen that CRP is an independent predictor of the prognosis of SAH
.

 

references:

CRP (C-Reactive Protein) in Outcome Prediction After Subarachnoid Hemorrhage and the Role of Machine Learning.
Stroke.
;0:STROKEAHA.
120.
030950 https://doi.
org/10.
1161/STROKEAHA.
120.
030950 .
 

CRP (C-Reactive Protein) in Outcome Prediction After Subarachnoid Hemorrhage and the Role of Machine Learning.
Stroke.
;0:STROKEAHA.
120.
030950 https://doi.
org/10.
1161/STROKEAHA.
120.
030950 Leave a message here