This study was approved by our local ethics committee. These case data were obtained retrospectively. From February 2018 to December 2020, the patients who had HICH as determined by computed tomography (CT) (GE Lightspeed 64, United States) scans and related examinations were enrolled. The selection criteria were as follows: (1) History of hypertension; (2) Altitude (1,500–4,000 m) (2,873.45 ± 745.99 m); (3) Time spent at these altitudes ≥ 5 years; (4) Hematoma volume 30–50 ml; (5) Good liver and kidney function; (6) Stable vital signs; (7) Hypertensive basal ganglia hemorrhage; and (8) Complete case data. Patients were excluded according to Figure 1. Twenty-one patients were treated with HBO after the operation (HBO group), and the other patients received conventional treatment (Control group) (Figure 1). The relevant data of the patients, including sex, age, medication history, hematoma size, hemoglobin concentration, and Glasgow Coma Scale (GCS), were collected and recorded in detail on admission. When the patient were admitted to the hospital, we asked the medical history. If the patient had used anticoagulant and antiplatelet drugs, we tested the patient’s coagulation function. If the patient were seriously affected by coagulation, we treated it conservatively and evaluated that it would not have a great impact on coagulation. We also conducted surgical treatment. All data were independently and blindly reviewed by two senior neurosurgeons.

Puncture and drainage hematoma: The largest hematoma plane was selected as the positioning point for drilling (Figures 2A, 3A). All patients were placed under local anesthesia, and a YL-1 puncture needle (Beijing Wan Fu Tie Medical Apparatus Co., Ltd., China) was used for drilling after selecting the positioning point (Xia et al., 2019; Figures 2D, 3D). After puncture and drainage, CT examination revealed that the hematoma was enlarged, and a craniotomy was performed to remove the hematoma. After 3–5 days of drainage, head CT was repeated and indicated that the hematoma had been reduced by 50–100%. The drainage tube was removed. Then, the patients in the HBO group were treated with HBO (100% oxygen at 2.2 ATA for 60 min for 10 days). In the control group, all treatments were the same, except HBO treatment was lacking. If the hematoma expanded more than 50 ml, we performed craniotomy to remove the hematoma. If the enlarged hematoma was less than 50 ml, but the GCS score decreased by 2 points, craniotomy was also considered.

Before puncture and drainage, CT angiography (CTA) was completed, and the result was normal (Figures 2G, 3G). CT perfusion (CTP) imaging was performed after written informed consent was obtained from the patients or their relatives. CTP imaging was performed twice. The first imaging scan was performed before HBO treatment for the HBO group and before standard treatment for the control group (Figures 2B,C, 3B,C); both groups underwent scans on the same number of days prior to the start of their respective treatments. The second CTP imaging scan was performed 10 days after HBO treatment for the HBO group and 10 days after the corresponding treatment for the control group (Figures 2E,F, 3E,F). A high-pressure syringe was used to inject iohexol (Yangtze River Pharmaceutical Group Co., Ltd., Production) through the anterior elbow vein (17.5 g in 50 ml, flow rate: 5 ml/s) (Wang et al., 2017). In the hematoma center, different regions of interest (ROIs) for the hematoma margin (defined as the area around the hematoma with a CT value at least 5 U less than of normal brain tissue or the area 1 cm away from the hematoma if the CT value was less than 5 U that of the brain tissue) and cortical area on the hematoma side (measuring the same length as the hematoma and edema) were selected for qualitative analysis. The mean perfusion parameters (cerebral blood volume (CBV) and (CBF) were calculated. The above parameters were also calculated for the control group. After 10 days, CTP imaging was performed again, and the parameters were recalculated. The above parameters were transferred to 4D Brain Perfusion (Vitrea, Japan) for ROI analysis. The formula ABC/2 was used to calculate the volume of the hematoma (Kothari et al., 1996). All data were independently and blindly reviewed by two senior neurosurgeons.

The CBV and CBF values of the two groups were measured and compared. A scatter diagram was plotted to demonstrate the relationship between the hemoglobin concentration on admission and the second set of CBV and CBF values. Receiver operating characteristic (ROC) curves were used to measure the cutoff values at which hematoma volume and hemoglobin concentration predicted a good outcome and the corresponding specificities and sensitivities. After a 6-month follow-up, the patients were scored using the Glasgow Outcome Scale (GOS). The scoring criteria were as follows: Good recovery (5), Moderate disability (4), Severe disability (3), Vegetative state (2), and Death (1). The patients were divided into 2 outcome groups: a good outcome group, which was defined as independent (GOS scores, 4-5), and a poor outcome group, which was defined as dependent (GOS scores, 3-1) (Dunatov et al., 2011).

All analyses were performed using SPSS 19 (IBM Corp., Armonk, NY, United States). Normally distributed data are expressed as the mean ± standard deviation. Quantitative data were analyzed between the two groups using separate T-tests and analysis of variance. Differences were considered significant at p < 0.05. A scattergram with a regression line was generated to determine the association between the hemoglobin concentration and CBV and CBF. ROC curve analysis was used to investigate the diagnostic relationship between risk factors and a good outcome. The Youden index was used to determine the optimal cutoff value as previously described (Wei et al., 2020).

A total of 41 patients were included in the study, of whom 21 received HBO treatment and 20 received conventional therapy. No significant differences in the basic characteristics of the two groups were noted (Table 1). Partial hematomas could not be extracted from two patients in the HBO group and three patients in the control group, resulting in success rates of 90.48 and 85%, respectively. The hematoma could not be extracted for several patients in the two groups, but their conditions were stable. Later, treatment was performed, and these data on the two groups were included. There were no significant differences in the success rate of the two groups (p > 0.05). No intracranial infection was found among the 41 patients.

No significant differences in the CBV and CBF values obtained from the first CTP scans were noted between the two groups (Table 2). CBV and CBF values of the hematoma centers and margins from the second scan in the HBO group were significantly higher than those in the control group (Table 2). In both groups, we analyzed the correlation between the hemoglobin concentration on admission and the CBV and CBF values extracted from the second CTP scan by generating a scatter diagram. The correlations between the hemoglobin concentration on admission and the CBV and CBF values from the hematoma center, margin and cortical area on the hematoma side were not significant in the HBO group (hematoma center: CBV: R² = 0.00, p = 0.96; CBF: R² = 0.00, P = 0.86; hematoma margin: CBV: R² = 0.03, p = 0.44; CBF: R² = 0.15, p = 0.07; cortical area on the hematoma side: CBV: R² = 0.09, p = 0.19 and CBF: R² = 0.03, p = 0.45) (Figures 4A–F). The correlations between the hemoglobin concentration on admission and the CBV and CBF values from the hematoma center, margin were significant in the control group (hematoma center: CBV: R² = 0.21, p = 0.04 and CBF: R² = 0.22, p = 0.04; hematoma margin: CBV: R² = 0.24, p = 0.03 and CBF: R² = 0.25, p = 0.02) (Figures 4A–D) group, but the correlations with the CBV and CBF values from the cortical area on the hematoma side were not significant in the control group (CBV: R² = 0.04, p = 0.38 and CBF: R² = 0.04, p = 0.39) (Figures 4E,F). ROC curve analysis showed that the AUC was 0.847 for hematoma volume and 0.653 for hemoglobin concentration (Table 3 and Figure 5A) in the HBO group. Similarly, the AUC was 0.844 for hematoma volume and 0.771 for hemoglobin concentration (Table 3 and Figure 5B) in the control group. ROC curve analysis showed that the specificity and sensitivity of hematoma volume were higher than those of hemoglobin concentration in the two groups (Figures 5A,B). The follow-up duration was 6 months. The GOS score of the HBO group was significantly higher than that of the control group (Figure 5C) (p < 0.05). However, there was no significant difference in the percentages of good outcomes between the two groups (Figure 5D).

This study was limited by the case load, its retrospective design, and the fact that the analysis was restricted to basal ganglia hemorrhages. These limitations might restrict the generalizability of our findings.