|
Prediction Value of Blood Markers in The Diagnosis of Glioma with Artificial Neural Network
|
|
| Author:
|
HAMID REZA KHAYAT KASHANI, MD, SHIRZAD AZHARI, MD, HASAN REZA MOHAMMADI, MD, KHALIL KOMLAKH, MD
|
| Abstract:
|
Purpose: Glioma is the most common brain tumour. Inflammatory mediators of glioma can change the blood cells and it can be used to differentiate healthy individuals from tumours. The objective of this study is to predict the grade of glioma tumour and differentiate the healthy person from the patient using statistical analysis and artificial neural network.
Methods: In this retrospective study, 157 glioma and 162 healthy individuals evaluated. Blood markers including leukocyte, lymphocyte, monocyte, platelets, NLR, PLR, MLR, dNLR were extracted. Predictive value of blood markers was determined with statistical analysis. Artificial neural network was also used to determine its accuracy in distinguishing between healthy and tumour cases with blood markers.
Results: A total 0f 157 glioma tumour (Grade I, 18; Grade II, 52; Grade III, 16; Grade IV; 71) and 162 healthy individuals were included in the study. Tumour group showed a significant increase in WBC, neutrophil, NLR and PLR compared to healthy controls and there was a significant decrease in lymphocyte, platelet and LMR. ROC curve represents that dNLR (AUC=0.729) and then NLR (AUC=0.715) are the most valuable prediction factor between tumour and healthy cases. Blood markers have a better prediction value when GBM and healthy individuals are compared.
This study reveals that ANN can differentiate healthy from tumour cases with 69.9% accuracy and this result is acceptable (AUC = 0.735).
Conclusions: Artificial neural network can predict the presence of glioma with 69.9% accuracy and it is a feasible and applicable method of using blood markers to determine preoperative glioma.
|
| Keyword:
|
Glioma; Glioblastoma; Neutrophil Lymphocyte Ratio (NLR); Biomarkers; Artificial Neural Network.
|
| EOI:
|
-
|
| DOI:
|
https://doi.org/10.31838/ijpr/2021.13.01.755
|
| Download:
|
Request For Article
|
|
|