The Diagnostic Value of Preoperative Hematology Makers in the Classication and Molecular Subtypes of Newly Diagnosed High-Grade Gliomas

Purpose: The purpose of our study is to explore the diagnostic value of the single and combined hematological maker for the classication and isocitrate dehydrogenase (IDH)-1/2 mutations (IDH) molecular subtypes of high-grade gliomas (HGGs). Methods: A total of 354 newly diagnosed HGGs patients were included in this study. Firstly, we compared the levels of hematology indicators in the classication and molecular subtypes of HGGs. Next, the correlation between the levels of hematology indicators with basic clinical features was analyzed. Finally, the diagnostic value of the single and combined hematology indicator for identifying the classication and molecular subtypes from HGGs was performed. Results: The level of brinogen (FIB) was higher in higher grade gliomas and glioblastoma multiforme IDH wild type (GBM IDH-wt). Nutrition-related indicators such as serum albumin (ALB), albumin/globulin ratio (AGR), and prognostic nutrition index (PNI) were negatively correlated with age, whereas FIB was positively associated with age. Compared with women, men with GBM had signicantly higher AGR and lower serum globulin (GLOB). We found that the best single and combined indicator for identifying GBM and GBM IDH-wt from HGGs were FIB [0.595 (0.519-0.672) and 0.615 (0.546-0.684)] and age+FIB [0.712 (0.642-0.783) and 0.726 (0.662-0.791)], respectively. Conclusions: Preoperative hematological indicators have high diagnostic value for GBM and GBM IDH-wt from HGGs, especially FIB combined age.


Introduction
High-grade gliomas (HGGs) are the most common and lethal primary type of brain tumors, presenting one of the highest rates of occurrence within the tumors of the central nervous system (CNS), de ned as Grade III (G3) and IV (G4) gliomas or glioblastoma multiforme (GBM) according to the 2016 World Health Organization (WHO) classi cation of CNS tumors 1,2 . HGGs individual therapies greatly vary across identifying the glioma grade and isocitrate dehydrogenase (IDH)-1/2 mutations (IDH) molecular subtypes con rmed by pathology after surgery 3,4 . Currently, the grade and IDH molecular subtypes of HGGs are determined by surgery or biopsy, which is invasive and traumatic to the human body. Therefore, it is necessary to establish a noninvasive method for the diagnosis of HGGs classi cation and IDH molecular subtypes.
At present, a large number of studies had shown that hematological indexes related to in ammation, coagulation, and nutrition had obviously predictive value in the prognosis of tumors [5][6][7] . Meanwhile, hematology related indicators had been demonstrated to be provided with remarkable predictive value for the prognosis of gliomas, including neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR) 8 , brinogen (FIB), prognostic nutrition index (PNI) and albumin-to-globulin ratio (AGR) 9 , etc. In addition, the different levels of hematological indicators had signi cant diagnostic value for gliomas 10 11 . Nevertheless, the current researches showed that the diagnostic value of hematological indicators had not yet fully included all indicators, and little is known about the grading diagnosis of HGGs.
In the present study, our goal is to identify the grade and IDH molecular subtypes of newly diagnosed HGGs based on preoperative hematological indicators related to in ammation, coagulation, and nutrition.
Moreover, we also explored the correlation between these hematological indicators and basic clinical characteristics, and compared the levels of hematological indicators between HGGs of different grade and IDH molecular subtypes. We have also established the receiver operating characteristic (ROC) curve analysis to determine the best diagnostic indicators for HGGs classi cation and IDH molecular subtypes.

Patients
The retrospective study included patients who were con rmed with histologically as newly diagnosed G3 and G4 gliomas from January 2017 to December 2018 in the First A liated Hospital of Zhengzhou University (FAHZZU). The following inclusion criteria were used: 1) patients with G3 and G4 gliomas con rmed by histopathology; 2) patients with complete data of blood routine, biochemical, and coagulation results; 3) patients with IDH mutations results of histopathology; 4) patients without previous malignancy or second primary tumors. The exclusion criteria were as follows: 1) patients who had clinical evidence of liver disease, acute infection, or chronic active in ammatory diseases; 2) patients who had autoimmune diseases, hematological disorders, or anticoagulation treatment; 3) patients who had no complete hematology results; 4) patients who had previous anti-tumor treatment before admission. According to the above inclusion and exclusion criteria, 354 patients were included in the study, eventually.

Data collection
The baseline characteristics of all enrolled HGGs patients came from the electronic medical record system of FAHZZU. Basic clinical information was collected including gender, age at diagnosis, preoperative Karnofsky performance status (pKPS) score, preoperative epilepsy (pEPI) status, histopathological grade, IDH mutation status, and hematology indexes, such as brinogen (FIB, g/L), serum albumin (ALB, g/L), serum globulin (GLOB, g/L). All blood sample tests were completed by the Clinical Laboratory of FAHZZU within 2 hours after collection. The related composite indexes were calculated as follows: NLR = absolute neutrophil counts/absolute lymphocyte counts ratio; PLR = absolute platelet counts/absolute lymphocyte counts ratio; MLR = absolute monocyte counts/ absolute lymphocyte counts ratio; AGR = serum albumin/ serum globulin ratio; and PNI = albumin (g/L)+5*total lymphocyte count (10 9 /L) 9 .

Page 4/16
The unpaired t-test was used to compare two groups of continuous variables, and the comparison of categorical variables was performed with the Chi-square test. Multi-group comparisons were evaluated by one-way analysis of variance (one-way ANOVA), and the least signi cant difference t-test (LSD-t) was used for Post Hoc Multiple Comparisons. The correlation between the hematology markers and clinical features was judged by the Pearson correlation coe cient for continuous variables and the unpaired ttest. To identify the diagnostic value of the hematology markers and their combinations, the ROC analysis was used to acquire the area under the curve (AUC). The categorical variables and continuous variables were shown as frequency (percentage) and mean ± standard deviation (SD), respectively. A 2tailed P value of <0.05 was considered to indicate statistical signi cance. All statistical analysis and drawing gures were performed using SPSS (version 24.0) and GraphPad Prism (version 6.01).

Clinical Characteristics
There were 354 HGGs patients included in this retrospective study according to the inclusion and exclusion criteria. Table 1 summarized detailed demographic information of the research subjects. Overall, 59 (16.7%) were G3 and 295 (93.3%) were G4 of the 354 patients, respectively. All clinical features were not statistically signi cant (p>0.05) except for IDH (p<0.001), age (p<0.001) and FIB (p=0.034). IDH mutations tended to occur in G3 and the elderly population was more susceptible to GBM. Moreover, the level of FIB was higher in G4 compared with G3. There were 27 (45.8%) women and 32 (54.2%) men in G3 To further understand the differences in the hematological indicators of IDH molecular subtypes, we divided patients into IDH wild type (IDH-wt) and IDH mutant type (IDH-mut) according to the IDH mutation status in the study. Compared with IDH-mut, the levels of NLR (t=2.315, p=0.035) and FIB (t=3.533, p<0.001) were both increased in IDH-wt ( Fig. 1a, b; Supplementary Table S1). Furthermore, we further classi ed patients into four subgroups based on their IDH mutation status, including G3 IDH mutant type (G3 IDH-mut), G3 IDH wild type (G3 IDH-wt), G4 IDH mutant type (G4 IDH-mut), and G4 IDH wild type (G4 IDH-wt). The results showed that only FIB was statistically signi cant (F=4.160, p=0.006, Fig. 1c), and further analysis showed that the levels of FIB were increased in G4 IDH-wt, compared with G3 IDH-mut (p=0.003) and G4 IDH-mut (p=0.040) (Supplementary Table S2).

Discussion
Although the current pathological diagnosis of gliomas grading was the gold standard, it needed to be con rmed after surgery. For the diagnosis of HGGs grading, we need a method of preoperative prediction to guide surgical strategies and treatment more accurately. Previous studies had indicated that in ammation, including NLR, PLR, LMR, FIB, ALB, AGR and PNI had the predictive and diagnostic value for the grading and prognosis of gliomas 8, 10-12 . Our research mainly analyzed the diagnostic value of preoperative hematological indicators related to in ammation, coagulation and nutrition in the classi cation and IDH molecular subtypes of HGGs. In addition, we explored their correlation with age, gender, pEPI, pKPS in different grade of HGGs. Furthermore, we evaluated their levels in IDH molecular subtypes of HGGs and different grade of HGGs.
The results of this study showed that the higher grade gliomas had higher levels of NLR, MLR, FIB, ALB and GLOB, and lower levels of PLR, AGR and PNI, respectively, This result was consistent with the previous related studies 11,13 , and declared more severe in ammation and worse nutritional status were present in higher grade gliomas. However, they were not statistically signi cant in the different grade of HGGs except FIB, which was contrary to these researches 11,13 . The possible reason for this was that there were too few G3 gliomas included in this study. Only the levels of NLR and FIB were associated with IDH mutation groups, indicating that there was a link between in ammation and coagulation 14 . Furthermore, we further analyzed the relationship between hematological indicators and IDH mutation subgroups, and found that only FIB had a signi cant difference between G4 IDH-wt and G3 IDH-mut or G4 IDH-mut. Previous study had shown that the immune response of the tumor microenvironment was more susceptible to the modulation of IDH mutation status in lower-grade gliomas 15 . This result suggested that there may be a connection between systemic and local in ammation and coagulation and immune regulation.
In our study, the levels of FIB and GLOB were positively correlated with age, and the levels of ALB, AGR and PNI were negatively correlated in the HGGs classi cation. However, the in ammation indicators were oppositely correlated in the HGGs grading groups. The reason for this result was the levels of FIB may re ect the critical function of FIB in an in ammatory tumor microenvironment that favors tumor progression 6 . When comparing hematological indicators in gender, we found that NLR and FIB had higher levels in men. Moreover, the level of PNI had the same result, consistent with the previous study 16 .
Different from other research results 13,17 , the levels of PLR in our research were opposite in this comparison. There was no signi cant difference in the levels of hematological indicators in the pKPS and pEPI subgroups.
In the analysis of the diagnostic value of hematological indicators, single and combined indicators were explored, respectively. The level of FIB had the highest diagnostic value among the 8 hematological indicators in distinguishing GBM and G4 IDH-wt from HGGs. Our results were different from studies that found that NLR was the best diagnostic indicator, and they studied the ability of hematological indicators in distinguish GBM and G4 IDH-wt from different diseases and three grade of gliomas 11,13 . In addition, as for combined indicators, age+PNI had the highest diagnostic value in the research of Wang et el 13 , however, the best combined diagnostic index was NLR+MLR in the study of Zheng et el 11 . The difference from them was that our research distinguished GBM and G4 IDH-wt from HGGs, and age+FIB had the highest diagnostic value among all the combined indicators. It was worth noting that grade and grade gliomas were included in their study, respectively.
This mechanism was temporarily unclear between hematological indicators and the grade of gliomas.
Previous studies had declared that immune suppression induced by neutrophils may promote the progression of glioma, and some T lymphocyte subsets may inhibit it by inducing cytotoxic cell death and cytokine production 18, 19 . Shao et el con rmed that platelet activation contributes to tumor angiogenesis, destroys extracellular matrix and releases adhesion molecules to promote cancer cell proliferation and metastasis, thus proving the importance of platelet activation in tumorigenesis 20 . The correlation between nutritional indicators and cancer may be related to the antioxidant effect of albumin on carcinogens such as nitrosamines and a atoxins, and the levels of GLOB may be related to the progression and metastasis of certain cancers 21,22 . Moreover, albumin re ected the state of systemic in ammation by down-regulating tumor necrosis factor alpha and interleukin 6 23 . Plasma brinogen was an important factor in in ammation and cancer expression, and had been demonstrated in vitro studies playing an important role in tumor cell proliferation, epithelial-mesenchymal transition, invasion, angiogenesis and hematological dissemination of tumor cells [24][25][26] .

Limitations
Our current research has some limitations. Firstly, this study may have a selection bias because it is a single center and retrospective analysis. Secondly, the sample size of the study included in this study was small, especially G3 gliomas, and grade and grade gliomas were not included. In addition, we did not have data for O6-methylguanine-DNA methyltransferase (MGMT) methylation status and telomerase reverse transcriptase (TERT) promoter status, both of which play an important role in patients with gliomas 27 . Therefore, multi-center prospective studies, further molecular pathology experimental analyses, large sample researches of HGGs are required.

Declarations
Funding No funding was received for this research.
Con icts of interest/Competing interests The authors declare that there is no con ict of interest or competing interests.