Irregular shape as an independent predictor of prognosis in patients with primary intracerebral hemorrhage

The utility of noncontrast computed tomography markers in the prognosis of spontaneous intracerebral hemorrhage has been studied. This study aimed to investigate the predictive value of the computed tomography (CT) irregularity shape for poor functional outcomes in patients with spontaneous intracerebral hemorrhage. We retrospectively reviewed all 782 patients with intracranial hemorrhage in our stroke emergency center from January 2018 to September 2019. Laboratory examination and CT examination were performed within 24 h of admission. After three months, the patient's functional outcome was assessed using the modified Rankin Scale. Multinomial logistic regression analyses were applied to identify independent predictors of functional outcome in patients with intracerebral hemorrhage. Out of the 627 patients included in this study, those with irregular shapes on CT imaging had a higher proportion of poor outcomes and mortality 90 days after discharge (P < 0.001). Irregular shapes were found to be significant independent predictors of poor outcome and mortality on multiple logistic regression analysis. In addition, the increase in plasma D-dimer was associated with the occurrence of irregular shapes (P = 0.0387). Patients with irregular shapes showed worse functional outcomes after intracerebral hemorrhage. The elevated expression level of plasma D-dimers may be directly related to the formation of irregular shapes.

accompanied by the occurrence of neurological deterioration, is associated with poor functional outcomes in patients with intracerebral hemorrhage 4 . The heterogeneity of hematoma has been verified to be associated with hematoma expansion 5,6 . Although the CT angiography (CTA) spot sign is accepted as the gold standard for judging the risk of hematoma expansion 7 , the need for a contrast agent limits its availability and poses an additional risk for some patients. Therefore, the prompt and accurate identification of patients that at high risk of developing hematoma expansion and giving them individualized treatment measures may provide benefits in functional outcomes 8,9 . Compared with CTA, noncontrast computed tomography (NCCT) is an inexpensive and widely used examination tool that is more suitable for early diagnostic applications in patients with intracerebral hemorrhage 10 . Some NCCT biomarkers of intracerebral hemorrhage have received increasing attention from researchers for their good performance in predicting the poor prognosis of patients. Morotti et al. showed good predictive performance when the irregular shape was used as an NCCT marker for judging early intracerebral hemorrhage expansion 11 . However, this study was based on unadjusted pooled estimates, there was no independent collection of real patient medical records for analysis, and the factors associated with the appearance of irregular shapes have not been explored.
Therefore, in this study, we comprehensively collected the clinical data of patients with intracerebral hemorrhage and explored whether the irregular shape could serve as an independent predictor for judging patient prognosis. In our previous studies, it was confirmed that plasma D-dimer expression levels can predict the poor prognosis of patients with intracerebral hemorrhage 12 . On this basis, we further investigated whether there was a specific correlation between the appearance of the irregular shape of the hematoma and the elevated expression level of plasma D-dimer.

Patients and methods
Patients. The clinical data of 782 patients with ICH admitted to the emergency neurosurgery ward of the First Affiliated Hospital of Harbin Medical University from January 2018 to September 2019 were retrospectively analyzed. The inclusion and exclusion criteria were as follows: age ≥ 18 years; CT scans performed within 24 h of admission; and standard techniques used (CA-7000 Sysmex; Dade Behring) to analyze the concentration of D-dimer. Patients without CT images or follow-up CT scans within 24 h after the initial CT examination, patients with intracranial hemorrhage due to trauma or aneurysm, and patients with infratentorial hemorrhage were excluded.
To assess the clinical status, the Glasgow Coma Scale (GCS) score was measured by a neurosurgeon at the time of patient admission. To determine the functional outcome, patients were followed up at least three months after symptom onset, and their clinical outcomes were assessed by the modified Rankin scale (mRS; poor functional outcome: mRS score 4-6; favorable functional outcome: mRS score 0-3). In addition, all patients were followed for survival status until death or the end of the study. The last date of Study period (September 30, 2020) was used as an end date. The study protocol was approved by the Institutional Review Board of The First Affiliated Hospital of Harbin Medical University. Simultaneously, all methods were carried out in accordance with the relevant regulations and guidelines of Harbin Medical University. Image features. The imaging data of the patients were evaluated independently by two experienced neurologists. Hematoma volumes were measured on industry-standard DICOM images using 3D Slicer 4.10.2, an open-source software (SPL, Harvard Medical School, Boston, USA). For patients with severe symptoms, continuous deterioration, and suspected intracerebral hemorrhage expansion, the noncontrast CT scans will be examined within 24 h after admission, and the presence of hematoma expansion and irregular shapes will be judged. Hematoma expansion was defined as an absolute growth > 6 mL or a relative growth > 33% in hematoma size on follow-up nonenhanced CT scans compared with the initial CT examination after admission 13 . Irregular shapes were defined as the presence of two or more connected or separated irregular hematomas at the edge of the hematoma on the axial section with the largest hematoma cross-sectional area 14 (Supplemental Fig. 1).

Statistical analysis.
Statistical analysis and calculations were carried out using dedicated software (IBM SPSS Statistics 21.0; SPSS Inc, Armonk, NY, USA) or GraphPad Prism 8 software (version 8.0.3; GraphPad Software Inc, San Diego, CA). Baseline characteristics were compared between patients with and without hematoma expansion. Categorized variables were expressed as counts (percentages) and compared using the chi-square test or Fisher's exact test, while continuous variables were expressed as the mean ± SD or median (interquartile range, IQR) values and compared using independent sample Student's t test.
Variables associated with mortality at 90 days after discharge, poor prognosis, and hematoma enlargement were included in the univariate analysis. Variables with P values < 0.05 in the univariate analysis were included in multivariate logistic models to adjust for confounders. The scatter plot represents the relationship between irregular shape and plasma D-dimer expression levels. Statistical significance for all tests was defined as P < 0.05.

Results
According to the inclusion and exclusion criteria, 782 patients were included in our study. Figure 1 shows the study flow chart of patient inclusion for the analysis.  Fig. 2). Table 1 shows the comparison of clinical characteristics between patients with irregular shapes and the control group. The results showed that initial blood pressure, initial hematoma volume, hematoma expansion, presence of IVH, degree of midline shift, and elevated plasma D-dimer levels were significantly related to the occurrence of irregular shapes. Univariate logistic regression analysis was performed for poor outcome and mortality after 90 days of discharge in patients with ICH, and the factors associated with the development of hematoma expansion within 24 h of admission were explored ( Table 2). The analysis results indicated a significantly higher rate of early hematoma expansion in patients with irregular shapes (OR = 3.776, 95% CI 1.950-7.312, P < 0.001) and had a poor prognosis at 90 days (OR = 4.380, 95% CI 3.003-6.389, P < 0.001) and an expressively higher mortality rate (OR = 5.233, 95% CI 2.991-9.155, P < 0.001). The results also indicated that patients with elevated plasma D-dimer levels had a higher risk of mortality (OR = 3.518, 95% CI 2.096-5.905, P < 0.001) and poorer prognosis (OR = 2.809, 95% CI 1.928-4.092, P < 0.001). Multivariate logistic regression analyses were performed for significant variables in the  www.nature.com/scientificreports/  (Table 3). To explore whether the performance of different NCCT imaging markers in predicting poor prognosis of patients with intracerebral hemorrhage is superior to the irregular shape, we comprehensively interpreted the NCCT imaging markers, including the Island sign and Blend sign, in 627 spontaneous ICH patients included in our study. The predictive performance of different NCCT imaging markers for poor outcomes in ICH patients was analyzed with specificity and sensitivity, and the ROC curve was drawn. Our results demonstrated that irregular shape was better than other NCCT imaging markers for predicting the occurrence of unfavorable outcomes in ICH patients (area under the curve-AUC = 0.819, sensitivity 74.0%, specificity 78.7%, CI 0.781-0.856, Fig. 3A-C). Furthermore, in our study, the appearance of irregular shapes of the hematoma was a better predictor of poor patient functional outcome than the initial hematoma volume. (AUC = 0.808, Fig. 3D). All patients will be followed up until death or the end of the study. A total of 535 patients were included in the Kaplan-Meier survival analysis, except for 92 patients whose death information was missing. The longest follow-up among all patients reached 32 months, with a median of 20 months, and by the end of the study on September 30, 2020 (Fig. 4). The results showed that the irregular shape of the hematoma was significantly associated with shorter survival of patients with cerebral hemorrhage (P < 0.001). In addition, in the data analysis, we also found that the appearance of irregular shapes was accompanied by altered plasma D-dimer expression levels, so we speculated whether the generation of irregular shapes resulted from the increased plasma D-dimer expression. In the scatter plot, plasma D-dimer expression levels were significantly correlated with the appearance of irregular shapes (P = 0.039, Fig. 5).

Discussion
Our statistical results shows that irregular shape has predictive significance for the prognosis of patients with ICH. Though island sign, blend sign and initial hematoma volume also have some prognostic value for ICH, the predictive value of irregular shape for the prognosis of ICH is still more significant comparing with them.
As a severely disabling neurological disease, spontaneous intracerebral hemorrhage (ICH) is receiving more and more attention on how to quickly and accurately predict the occurrence of its adverse consequences, and then help clinicians to grade and manage patients. Regarding imaging, although there remain some controversies surrounding the clear definition and formation mechanism of the spot sign on CTA, it has been recognized as a well-established imaging marker with 51%-91% sensitivity and 58%-89% specificity for predicting the occurrence of hematoma expansion in the short term in ICH patients 7,15 , and the spot sign has also been identified as an independent predictor of poor prognosis in intracerebral hemorrhage. However, the need for contrast media for some patients (e.g., kidney failure, contrast allergy) limits CTA availability. The extravasation of contrast medium that occurs during CTA procedures may contribute to the occurrence of rebleeding 16 , posing an additional risk for some patients, which is not routinely performed in emergencies. Therefore, researchers have gradually explored the predictive value of NCCT markers, including the blend sign, irregular shape, and island www.nature.com/scientificreports/ sign, for intracerebral hemorrhage expansion, and the results showed equally good predictive performance 11,17 . Qi et al., in a retrospective analysis of 252 patients with intracerebral hemorrhage, showed that the island sign was an independent predictor of hematoma enlargement and poor outcomes in patients with ICH 18 . In a study conducted by Zhang et al. 1111 patients with ICH were recruited, and it was found that blend signs could help doctors grade the management of patients with intracerebral hemorrhage at admission 19 . In this study, we also analyzed the performance of different NCCT image markers in predicting poor functional prognosis in patients with cerebral hemorrhage, and the results showed good performance. In which the predictive power of irregular shapes was the best. In addition, many studies have demonstrated the significant value of hematoma volume for the prediction of prognostic situations in patients with intracerebral hemorrhage [20][21][22][23] . In our study, initial hematoma volume predicted poor functional outcomes in patients than irregularly shaped hematomas. In a previous study, Moratti et al. established a scoring system to predict the instability of the hematoma based on three indicators: presence or absence of the mixed sign, time to the appearance of the mixed sign, and reduction in the density of the hematoma at presentation 24 . However, these studies did not directly use the irregular shape of the hematoma to evaluate the prognosis of patients and did not explore the cause of NCCT imaging hallmarks.
The overall survival (OS) of patients with irregular shapes in our study was significantly shorter than that of patients without irregular shapes. During a follow-up of up to 32 months for all patients included in this study, Figure 3. Receiver operating characteristic curves of irregular shape, island sign, blend sign, and IHV with their corresponding areas under the curve (AUCs) for predicting poor prognosis. The best cut-off points were identified with their sensitivity, specificity, and confidence interval (CI), respectively.   www.nature.com/scientificreports/ The reason for the formation of irregular shape imaging signs was not deeply investigated in a large number of previous studies. In the present study, we showed for the first time that the appearance of irregular shapes was significantly associated with the elevation of plasma D-dimer (> 0.55 mg/L FEU). Plasma D-dimer has been shown to have good performance in the prediction of poor outcomes and mortality in patients with spontaneous ICH 12,25,26 . Acute brain injury resulting from hematoma formation is closely related to coagulation activation, while D-dimer, the final product of coagulation, is a fibrin degradation product released from plasmin degrading fibrin monomers 27 , and its expression quantity reflects the activation level of the systemic coagulation response. The massive release of thrombin, which induces inflammatory reactions in the perivascular tissues and then damages the blood-brain barrier (BBB), may lead to early hematoma expansion 28,29 . However, there was no significant statistical association between plasma D-dimer levels and hematoma expansion at admission in this study, which may be related to sample size and systematic error.
Our study has certain limitations. First, the patients included in this study were all hospitalized in the stroke emergency center of our hospital. Although other hospitals in the city have also established stroke centers, they are small and have low comprehensive diagnosis and treatment capacity, so patients with severe disease will more likely be treated in our hospital. This may contribute to a worse functional prognosis and even higher mortality in the patient population included in this study.
Second, some patients may have delayed going to the hospital for treatment, so the delay in time from symptom onset to admission and CT scans may have affected some variables in this study, for example, hematoma expansion, as they may have stabilized before the first CT examination. Previous studies have demonstrated that hematoma expansion is more likely to occur within the first hour of intracerebral hemorrhage symptom onset 30 . However, it is very difficult to obtain a precise judgment of the specific time of onset for each patient with intracerebral hemorrhage.
Finally, all patients in this study were Chinese, and no other races were included in the study, so the findings may not indicate the actual incidence in other regions. Although the morbidity of patients with nontraumatic ICH is directly due to cerebrovascular rupture, the main causes of ICH are not the same in Western countries as in Asia. Chronic hypertension is a major cause of ICH in China and other Asian countries, but cerebrovascular amyloidosis and oral anticoagulants play a more important role in the occurrence of ICH in Western countries 3,[31][32][33] . This may be because aging is more severe in Western developed countries, resulting in the deposition of amyloid-β peptide causing degeneration of cerebral vessels and the need for more anticoagulant or antiplatelet drug therapy in the elderly to combat the occurrence of ischemic diseases 34 . Hypertension can lead to massive expression of vascular extracellular matrix such as laminin and fibronectin, which will destroy the structure of cerebrovascular, leading to vascular sclerosis as well as increased permeability 35,36 . This is also the reason why cerebral hemorrhage shows a strong association with hypertensive disorders 37 . In China, the incidence of hypertension is 2.5 times higher than that in the United States, and more than 80% of stroke patients have been diagnosed with hypertension 36,38 . In addition, Chinese patients were more inclined to use a single blood pressure control drug when treating hypertension, and the proportion of polypharmacy combination therapy was only 21. 9%, which would result in less effective prevention and control of hypertension 39,40 . Therefore, the difference in the incidence of hypertension and the difference in medication strategies may be important factors leading to the difference in the causes of ICH between China and Western countries. Future studies using larger samples are needed to support our findings, which will help clinicians find a more rapid and accurate method to judge the occurrence and progression of patients with ICH.

Conclusion
CT irregular shape has an excellent performance in predicting functional outcomes in patients with spontaneous intracerebral hemorrhage. As an imaging marker of NCCT, the irregular shape can help doctors identify patients with rapid disease progression, poor prognosis, and high mortality. Moreover, this is the first study to confirm the statistical significance of D-dimer in predicting irregular shapes and has excellent potential to further explore the pathophysiological mechanism of D-dimer in hematoma enlargement. To improve prediction accuracy and sensitivity, plasma D-dimer level increases can be combined with other cerebral hemorrhage-related factors in subsequent studies.

Data availability
The data that support the findings of this study are available from the corresponding author upon reasonable request. www.nature.com/scientificreports/