Although significant progress has been made in the diagnosis and treatment of pituitary tumors, there is currently a lack of comprehensive preoperative evaluation tools to properly classify patients with pituitary tumors to guide individualized treatment options. WHO classification is an important prognostic tool based on histological criteria, but it cannot fully predict tumor progression. There are substantial differences in the same grade of the tumor. Postoperative pathology and genetic examination play a major role in tumor recurrence or progression. However, most of these detection methods are invasive and expensive, and cannot fully define the spatial heterogeneity of tumors. A large number of studies have shown that hematological parameters have great potential in disease prognosis stratification(18-21). However, before our investigation, no relevant reports based on blood-related markers to predict the biological behavior of pituitary tumors have been collected. Using clinical and hematological information from patients with pituitary tumors, we carried out a quantitative analysis to build a predictive survival model and generate a nomogram to further predict the prognosis.
In this study, we collected a wide range of hematological parameters to assess their association with pituitary tumor progression. We used these parameters to construct a BRS model with prognostic value, and its predictive performance was moderate, especially in long-term prediction (AUC = 0.788).
To further improve the predictive ability and applicability of BRS, we incorporated clinical prognostic factors into BRS and constructed a clinical prognostic model. We validated the model in the test set. The results showed that the area under the ROC curve at 1-, 3-, and 5 years after surgery was 0.718,0.852, and 0.864, respectively. The area under the curve at 1-, 3-, and 5 years was 0.600,0.889, and 0.660, respectively. Based on this, a nomogram was constructed. The nomogram corrected the C index 0.732 (training set) and 0.704 (validation set). We found that this model has good predictive accuracy in predicting pituitary tumor progression based on multiple hematological parameters and multiple clinical prognostic factors, and no other similar models have been found.
Ki-67 proliferation index is often used as a proliferation marker in routine pathological studies and is highly correlated with tumor cell proliferation and growth. Data on the relationship between the Ki-67 proliferation index and PA recurrence have been controversial (22, 23). Zakir et al concluded from the study of patients with a giant pituitary tumor that the Ki-67 proliferation index was not related to the tumor progression of giant PA (24). Similarly, YOgawa analyzed the results of 33 patients with Ki-67 > 3 % pituitary adenomas and concluded that high Ki-67 levels were not associated with the duration of disease stabilization in these patients (25).In contrast, some experts say that the Ki-67 proliferation index is closely related to the prognosis of patients. According to the World Health Organization ( WHO ), the fourth edition of pituitary tumor classification, Ki-67 > 3 % is used to evaluate tumor proliferation and invasion, but the view of the 3 % value varies from person to person. In the study of Kim et al., a Ki-67 index higher than 3 % is associated with PA recurrence. (26) Hasanov et al.reported that there was no difference in the risk of recurrence in the high Ki-67 group (>3 % ). The high Ki-67 index had a critical point of 2.5 % and sponge infiltration was a reliable marker of PA recurrence (27). R.et al.reported that there was no difference in the risk of recurrence in the high Ki-67 group (>3 % ), the critical point of high Ki-67 index was 2.5 % and sponge infiltration was a reliable marker of PA recurrence. (28) The expression threshold level of the Ki-67 index was > 2. % Predict progress with high specificity (29).In this study, Ki-67 HR was 1.31, and the higher the value, the higher the probability of recurrence or progression.
Tumor size is associated with a variety of factors. For large and giant nonfunctional pituitary adenomas, the maximum diameter of preoperative solid tumors is an important predictor of recurrence or progression (30). In a study with an average tumor size of 23 mm, it was concluded that the larger the tumor, the greater the risk of recurrence. The risk of recurrence of large tumors is 1.04 times that of small tumors (31). Tumor size (HR = 11.06,95 % CI: 6.11-20.03, P < 0.0001) was also reported. Preoperative pituitary tumor size was an independent predictor of prognosis (32). In this study, preoperative pituitary tumor size (HR = 1.04,95 % CI: 1.01-1.08) was an independent predictor of pituitary tumor recurrence or progression.
Ecto5 ' -nucleotidase is an enzyme that binds glycosylphosphatidylinositol ( GPI )(33). It is currently believed that disorders exist in most types of human cancers(34-36). The main function of this enzyme is to synthesize extracellular nucleosides (35, 37), It is an indispensable key enzyme in nucleotide catabolism. 5 ' -Nucleotidase can decompose adenosine-5 ' -phosphate into phosphate and nucleosides, thereby breaking the 3 ', 5 ' -phosphate diester bond between nucleosides and phosphate to complete the decomposition of nucleotides. High levels of adenosine have been found in solid tumors. As revealed by Spychala et al., adenosine has also been shown to promote tumor development and inhibit cytokines, immune cell adhesion, natural killer cells, and macrophages(38). As early as decades ago, 5 ' -nucleotidase has been shown to have a synergistic effect on T lymphocyte proliferation. Arias et al. were involved in the binding of lymphocytes to endothelial cells by using a monoclonal antibody ( mAb ) 4G4, which exerts anti-inflammatory and anti-tumor effects by inhibiting ecto-5 ' -nucleotidase activity in peripheral blood lymphocytes (39).According to the results of Hu et al., high expression of Ecto-5 ' -nucleotidase is an independent factor for poor prognosis in patients with gastric cancer and can be used as a reliable prognostic biomarker for gastric cancer. The high concentration of Ecto-5 ' -nucleotide expression in the study of Hu et al.is an independent poor prognostic factor for patients with gastric cancer, so it can be used as a reliable prognostic biomarker for gastric cancer (34). The expression of NT5E also increased in thyroid carcinoma. Tumor microinvasion and lymph node metastases are linked to high expression of NT5E(36). Following a comparison of the blood NT5 E levels between colon cancer patients and the healthy control group, Wang et al. The malignant development and clinical prognosis of colorectal cancer are directly related to the elevated expression of serum NT5E in patients with colorectal cancer. (35).
In this study, we further found that increased NT5 E concentration was a risk factor for tumor recurrence and progression, with an HR value of 1.18,95 % confidence interval: 1.06-1.31. Consistent with previous studies, high NT5E values are associated with poor clinical outcomes.
A tumor microenvironment (TME) is a special ecosystem formed by the interaction between cells in the tumor, and the continuous interaction and interference within it determine its overall ecological state. High potassium ion concentration in the microenvironment prevents tumor growth (40). High levels of extracellular potassium concentration limit nutrient uptake, inhibit the effector program of T cells and stimulate autophagy. This effect reduces histone acetylation levels at effectors and depletion sites, resulting in CD8 + T cells with more durable, pluripotent, and tumor-clearing capacity (40). As the core regulator of TAM ( tumor-associated macrophages ) function, the presence of inward rectifier K + channel Kir2.1 is very important for the anti-tumor state of TAM. The conditional lack of Kir2.1 allows TAM to regain an anti-tumor state and improve local anti-tumor immunity. However, if the channel is deficient, it will lead to electrochemically dependent glutamine uptake disorder, which will lead to the transformation of TAM metabolism from oxidative phosphorylation to glycolysis(41). T-cell function and potassium-mediated suppression of Akt-mTOR signaling are influenced by the serine/threonine phosphatase PP2A activity. The expression of potassium channel Kv1.3 improves effector function in vitro and in vivo, increases the K efflux of tumor-specific T cells, and increases the tumor clearance rate and survival rate of melanoma mice (42). The high concentration of extracellular potassium ions is closely related to hydrogen ions. To maintain the balance of internal and external potassium ions, the H + / K + exchange channel is opened to make potassium ions influx and compensate for hydrogen ions outflow, increasing to extracellular H + concentration and an increase in an extracellular acidic environment. Studies have shown that long-term exposure to acidic environments impairs methionine absorption and metabolism by down-regulating SLC7A5, thereby altering the deposition of H3K27me3 in the key T cell stem gene promoter. These changes promote the maintenance of cell stemness and improve anti-tumor efficacy (43). Moreover, lactate can boost anti-tumor immunity and CD8+ T cell survival. CD8 + T cell-dependent tumor growth was inhibited by subcutaneous injection of sodium lactate in tumor-bearing mice. Tumor growth was also inhibited when lactate-pretreated CD8 + T cells were transferred to tumor-bearing mice in vitro(44). Similarly, the BRS constructed in this study confirmed that the good prognosis of pituitary adenoma with a high K level was related, and the average PFS of high K was 1139.64 days. The average PFS of the low potassium group was 771.88 days. It provides more reference indicators for the prognosis evaluation of pituitary tumors. However, the mechanism of the effect on the prognosis of pituitary tumors needs to be further explored.
Cholinesterase (ChE) is a kind of glycoprotein synthesized by hepatocytes and then released into the blood at a certain rate. It can be used as one of the markers of liver protein synthesis. The enzyme is widely present in the liver, central nervous system, and peripheral nervous system(45). However, decreased ChE levels were observed in some cases, such as acute and chronic liver injury, cirrhosis, and liver metastasis, as well as protein-energy malnutrition, stress, and inflammation(46). According to this passage, the increase of serum cholinesterase activity is mainly related to kidney disease, obesity, hyperthyroidism, fatty liver, and other diseases, and may also be related to hemolytic anemia, schizophrenia, megaloblastic anemia, and other diseases. Therefore, serum cholinesterase levels can be used to evaluate the severity and prognosis of the disease and is a useful clinical indicator. A variety of malignancies have bad prognoses when cholinesterase levels are reduced, such as prostate cancer(47)、colorectal cancer (48)、non-small cell lung cancer (49)、carcinoma of stomach (50). It is crucial in determining which illnesses are benign and which ones are malignant. Compared with the benign gastric disease group, the serum cholinesterase level in the gastric cancer group was significantly lower (51). Studies have also shown that overexpression of cholinesterase directly disrupts the metabolism of acetylcholine, leading to brain neurotransmission disorders and even Alzheimer's disease and cancer. It has been reported that a significant increase in ChE activity was observed in tumor cells and glioma (52). It is shown that the concentration of ChE in patients with brain tumors is higher than that in healthy people. The poor prognosis of pituitary tumors in this study was associated with high concentrations of ChE. This conclusion is contrary to the role of ChE in other tumors in the literature. There are few studies on ChE in intracranial tumors, and there is no other literature for reference. The analysis may be due to the small sample size, resulting in accidental results.
There are some limitations to what we tried to study. First, because we conducted a single-center retrospective study, there may be bias in the analysis and selection process. Second, we have a relatively small number of patients in several subgroups. Clinical and biochemical data are limited. Thirdly, our model has a poor performance in the prediction of short-term recurrence, so it is necessary to further improve the prediction model to improve the accuracy of short-term prognosis prediction.