Abstract
Background
Accurate prognostication in deep supratentorial intracerebral hemorrhage (ICH) remains challenging. Hematoma volume is a key determinant of outcome and a component of the ICH Score. Whether radiomics-derived shape features provide prognostic information beyond volumetric burden and clinical risk scores remains unclear. We investigated the association between radiomic shape characteristics of basal ganglia hemorrhage (BGH) and intraventricular hemorrhage (IVH) with in-hospital mortality and unfavorable functional outcome.
Methods
We retrospectively analyzed 50 patients with deep supratentorial ICH and IVH. Semi-automated segmentations were used to extract radiomic features (roundness, elongation, flatness, Feret diameter, surface area) and volumetric parameters. Logistic regression models were constructed for in-hospital mortality and unfavorable functional outcome (mRS 5–6 at discharge; secondary definition mRS 4–6), with collinearity assessment guiding feature selection. Model discrimination was evaluated using receiver operating characteristic curves with DeLong comparisons. Secondary models incorporated the ICH Score. Calibration and internal validation were assessed using the Hosmer–Lemeshow test and bootstrap resampling (1000 iterations).
Results
For in-hospital mortality, BGH roundness showed discrimination comparable to BGH volume (AUC 0.742 vs 0.738). In multivariable analysis, lower roundness remained independently associated with mortality after adjustment for the ICH Score (OR 0.20 per 0.1 increase; 95% CI 0.04–0.66; p = 0.021), whereas volume was no longer significant. The combined model achieved an AUC of 0.866 versus 0.813 for the ICH Score (DeLong p = 0.28). IVH-derived shape parameters were not associated. For unfavorable outcome, BGH volume remained strongest predictor.
Conclusions
Radiomic hemorrhage roundness may provide complementary prognostic information beyond hematoma volume and the ICH Score in deep ICH with IVH.
Supplementary Information
The online version contains supplementary material available at 10.1007/s00415-026-13875-1.
Keywords: Intracerebral hemorrhage, Basal ganglia hemorrhage, Intraventricular hemorrhage, Hematoma morphology, ICH score
Introduction
Spontaneous intracerebral hemorrhage (ICH) is associated with high in-hospital mortality and substantial long-term disability [16]. Accurate early prognostication remains a central challenge in clinical practice, as it directly influences treatment decisions. Established predictors of outcome include age, admission neurological status, hematoma volume, and the presence of intraventricular hemorrhage [19]. These variables are incorporated into widely used clinical scores such as the ICH Score [7].
Hematoma volume has consistently emerged as one of the strongest predictors of mortality and functional outcome [12]. Volume reflects the overall hemorrhagic burden and correlates with mass effect and tissue destruction. However, radiologically, hemorrhages of similar size often differ substantially in their geometric configuration. Such differences may reflect variations in bleeding dynamics, spatial expansion patterns, or secondary disruption of surrounding tissue [10, 11]. These structural aspects are not fully captured by volume-based metrics alone, as incorporated in the ICH score, where volume is dichotomized at a threshold of 30 cm3 based on the ABC/2 formula, without any further refinement or adjustment [7].
Quantitative imaging approaches now allow objective characterization of lesion morphology beyond simple volumetric assessment [8]. Radiomic shape descriptors, including roundness, elongation, and surface characteristics, provide reproducible measures of hemorrhage geometry [3]. While radiomic methods have been increasingly investigated in oncologic imaging and ischemic stroke, data on their prognostic relevance in spontaneous ICH remain limited [8, 20]. In particular, it remains unclear whether morphologic shape characteristics provide prognostic value beyond established clinical predictors such as hematoma volume and the ICH Score.
Deep supratentorial ICH, especially involving the basal ganglia, represents a relatively homogeneous anatomical subgroup in which mass effect and intraventricular extension frequently coexist [4]. This anatomical subgroup allows a focused evaluation of whether geometric features of hemorrhage are independently associated with outcome. By restricting the cohort to patients with concomitant intraventricular hemorrhage, we aimed to further reduce clinical heterogeneity and focus on a high-risk subgroup in which both parenchymal and ventricular components contribute to outcome.
In this study, we evaluated the association between radiomic shape characteristics of BGH and clinical outcome in patients with deep supratentorial ICH. We hypothesized that hemorrhage roundness, as a quantitative measure of geometric regularity, would be independently associated with in-hospital mortality and provide prognostic information beyond hematoma volume as represented in the dichotomized ICH Score.
Methods
Study design and patient population
This retrospective single-center cohort study included adult patients admitted between 2020 and January 2026 with spontaneous deep supratentorial ICH. Inclusion required a primary BGH with concomitant IVH on initial non-contrast computed tomography (CT). Patients with secondary hemorrhage due to vascular malformations, intracranial tumors, trauma, or hemorrhagic transformation of ischemic stroke were excluded. The study was approved by the Ethics Committee of Leipzig University (reference number: 073/26-ek) in accordance with institutional and national guidelines and regulations. The requirement for informed consent was waived due to the retrospective study design and anonymized data analysis.
Clinical data collection and outcome definition
Baseline demographic and clinical variables were extracted from electronic medical records, including age, sex, admission Glasgow Coma Scale (GCS), and variables required to calculate the ICH Score according to established criteria [7]. The ICH Score was derived from admission clinical and radiologic parameters. The primary endpoint was in-hospital mortality. The secondary endpoint was unfavorable functional outcome at discharge, defined as modified Rankin Scale (mRS) score 5–6. Functional status was determined based on documented neurological assessment at discharge. An additional exploratory analysis was performed using an alternative definition of unfavorable outcome (mRS 4–6).
Imaging acquisition and hematoma segmentation
All patients underwent non-contrast CT imaging at admission according to institutional stroke protocols. Hematoma segmentation was performed using 3D Slicer (version 5.6.2; www.slicer.org), as illustrated in Fig. 1. A semi-automated threshold-based approach was applied, followed by manual refinement where necessary to ensure anatomical accuracy. Separate segmentations were generated for BGH and IVH, enabling independent volumetric and morphologic analyses of both compartments. Volumes were calculated directly from the segmented regions and reported in cubic millimeters and cubic centimeters.
Fig. 1.
Representative segmentation of deep supratentorial intracerebral hemorrhage with intraventricular extension. A Axial and B coronal non-contrast CT images demonstrating basal ganglia hemorrhage (BGH, red) and intraventricular hemorrhage (IVH, orange). C 3D reconstruction of the segmented hemorrhage compartments. D Sagittal CT slice illustrating the spatial configuration of the hemorrhage
Radiomic shape feature extraction
Shape-based radiomic features were extracted from segmented hemorrhage volumes using geometric descriptors implemented in 3D Slicer. Roundness was defined as a measure of similarity to a sphere, calculated as the ratio between the surface area of a sphere derived from the Feret diameter and the actual surface area of the segmented object, with a value of 1 indicating a perfect sphere. Representative examples of hematomas with high and low roundness are provided in Supplementary Figure S1. Elongation and flatness were calculated from the principal moments of inertia of the segmented volume, reflecting the relative proportions of the principal axes and deviations from isotropic geometry. The Feret diameter was defined as the diameter of the smallest sphere enclosing the entire segmented volume. Surface area corresponded to the total area of the three-dimensional hemorrhage boundary. To limit dimensionality and reduce the risk of overfitting, only shape features were analyzed. These included roundness, elongation, flatness, Feret diameter, and surface area. No intensity- or texture-based features were considered. For regression analyses, roundness values were scaled by a factor of 10 to facilitate interpretation of regression coefficients, corresponding to changes per 0.1 unit on the original scale.
Statistical analysis
Continuous variables are reported as mean ± standard deviation or median (IQR), as appropriate; categorical variables as counts and percentages. Associations with in-hospital mortality were evaluated using logistic regression, reporting odds ratios (ORs) with 95% confidence intervals (CIs). Correlations between volumetric and morphologic parameters were assessed using Pearson coefficients. Multicollinearity was examined using variance inflation factors, and collinear variables were not included simultaneously in multivariable models. Two prespecified multivariable models were constructed. The first included BGH roundness and BGH volume to assess whether geometric characteristics were independently associated with mortality beyond volumetric burden. The second model additionally incorporated the ICH Score to evaluate incremental prognostic value. Although the ICH Score includes dichotomized hematoma volume (> 30 cm3), continuous BGH volume was modeled to retain granularity [7]. Given the limited number of outcome events (n = 17), the number of predictors was restricted. Additional sensitivity analyses were performed including IVH volume and sex as covariates. For exploratory analyses, patients were stratified into high and low BGH roundness groups based on the median value. Model discrimination was assessed using receiver operating characteristic (ROC) analysis with comparison of AUCs by DeLong’s test. Calibration was evaluated using the Hosmer–Lemeshow test. Internal validation was performed by bootstrap resampling (1000 iterations). Analogous analyses were conducted for unfavorable functional outcome (mRS 5–6; mRS 4–6). All analyses were performed using R (version 4.4.2). A two-sided p < 0.05 was considered statistically significant.
Results
Cohort characteristics
A total of 50 patients with BGH and concomitant IVH were included. In-hospital mortality occurred in 17 patients (34%), and unfavorable functional outcome at discharge (mRS 5–6) was observed in 28 patients (56%). Baseline demographic, clinical, and imaging characteristics are summarized in Table 1. The median age was 71.0 years (IQR 62.0–78.0), and 29 patients (58%) were male. The median admission Glasgow Coma Scale score was 6.0 (IQR 3.0–11.0). The median ICH Score was 3 (IQR 2–4). Median length of stay was 13 days (IQR 7–23), ranging from 1 to 61 days. Selected comorbidities were present in the cohort, including diabetes (22%), heart failure (26%), chronic obstructive pulmonary disease (10%), and malignancy (12%), while dementia and peripheral arterial disease were less frequent (each 6%).
Table 1.
Baseline demographic, clinical, and imaging characteristics of patients with BGH and concomitant IVH. Continuous variables are presented as median (IQR), and categorical variables as number (percentage)
| Characteristic | N = 50 |
|---|---|
| Age | 71 (62–78) |
| Sex – Female | 21 (42%) |
| Sex – Male | 29 (58%) |
| Admission GCS | 6 (3–11) |
| ICH Score (median, IQR) | 3 (2–4) |
| ICH Score—1 | 4 (8%) |
| ICH Score – 2 | 16 (32%) |
| ICH Score – 3 | 17 (34%) |
| ICH Score – 4 | 12 (24%) |
| ICH Score – 5 | 1 (2%) |
| BGH volume (cm3) | 16.4 (8.5–40.7) |
| IVH volume (cm3) | 8.7 (4–22.8) |
| In-hospital mortality | 17 (34%) |
| Unfavorable outcome (mRS 5–6) | 28 (56%) |
| Unfavorable outcome (mRS 4–6) | 41 (82%) |
GCS Glasgow Coma Scale, ICH intracerebral hemorrhage, IVH intraventricular hemorrhage; mRS modified Rankin scale, BGH basal ganglia hemorrhage
Univariable screening of morphologic parameters
All extracted shape parameters of the BGH and IVH were evaluated in univariable logistic regression models for in-hospital mortality. The full set of univariable associations is displayed in Fig. 2.
Fig. 2.
Univariable associations of morphologic radiomic parameters with in-hospital mortality. Forest plot displaying odds ratios and 95% confidence intervals from univariable logistic regression analyses of shape-derived radiomic features of basal ganglia hemorrhage and intraventricular hemorrhage. The vertical dashed line indicates an odds ratio of 1. Surface area variables are scaled per 100 mm2
Among BGH-derived features, roundness (OR 0.21 per 0.1 increase, 95% CI 0.06–0.54; p = 0.004), Feret diameter (OR 1.07, 95% CI 1.03–1.13; p = 0.002), and surface area (OR 1.03 per 100 mm2, 95% CI 1.01–1.06; p = 0.005) were significantly associated with mortality, whereas elongation and flatness were not. IVH Feret diameter (OR 1.03, 95% CI 1.00–1.06; p = 0.044) and IVH surface area (OR 1.01 per 100 mm2, 95% CI 1.00–1.03; p = 0.032) also demonstrated significant associations. Given the conceptual overlap between volumetric and geometric size descriptors, correlation analysis was performed. As shown in Fig. 3, BGH surface area (r = 0.99) and BGH Feret diameter (r = 0.87) demonstrated very strong correlation with BGH volume, whereas roundness showed only moderate correlation (r = − 0.40). Variance inflation factor analysis confirmed relevant multicollinearity for surface area (VIF 4.92) and Feret diameter (VIF 5.13), while roundness exhibited low collinearity (VIF 1.39). Based on these findings, roundness was selected as the primary morphologic parameter for multivariable modeling. Baseline characteristics stratified by high versus low BGH roundness are provided in Supplementary Table S1. No significant differences in key clinical or radiological variables were observed between groups.
Fig. 3.
Correlation matrix of basal ganglia morphologic parameters. Pearson correlation matrix of volumetric and geometric descriptors of BGH. Strong correlations were observed between volume, surface area, and Feret diameter, whereas roundness showed only moderate correlation with size-related parameters
Morphologic model and in-hospital mortality
In univariable analysis, lower BGH roundness (OR 0.21 per 0.1 increase; 95% CI 0.06–0.54; p = 0.004) and larger BGH volume (OR 1.04 per cm3 increase; 95% CI 1.01–1.08; p = 0.010) were significantly associated with mortality. In a multivariable model including BGH roundness and BGH volume, both parameters remained independently associated with mortality (Table 2). Lower roundness was associated with increased mortality (OR 0.21; 95% CI 0.05–0.65; p = 0.016), and larger hematoma volume remained significant (OR 1.04; 95% CI 1.01–1.08; p = 0.028).
Table 2.
Logistic regression models for in-hospital mortality. Univariable and multivariable analyses evaluating associations between BGH morphologic radiomic features and mortality. Odds ratios are reported with 95% confidence intervals
| Variable | OR | 95% CI | p—Value |
|---|---|---|---|
| Morphologic model | |||
| BGH roundness (per 0.1 increase) | 0.21 | 0.05–0.65 | 0.016 |
| BGH volume (per cm3) | 1.04 | 1.01–1.08 | 0.028 |
| Full model (incl. ICH Score) | |||
| ICH Score (per point) | 3.42 | 1.29–11.73 | 0.025 |
| BGH volume (per cm3) | 1.02 | 0.98–1.06 | 0.308 |
| BGH roundness (per 0.1 increase) | 0.20 | 0.04–0.66 | 0.021 |
CI confidence interval, ICH intracerebral hemorrhage, OR odds ratio, BGH basal ganglia hemorrhage. The morphologic model includes BGH roundness and BGH volume. The fully adjusted model additionally includes the ICH Score
The distribution of BGH roundness stratified by survival status is illustrated in Fig. 4, demonstrating significantly lower roundness values among non-survivors (Wilcoxon p = 0.005).
Fig. 4.
BGH roundness stratified by in-hospital mortality. Distribution of BGH in survivors and non-survivors. Boxes represent the interquartile range with the median; dots represent individual observations
Prognostic contribution beyond the ICH score
The ICH Score was strongly associated with mortality in univariable analysis (OR 4.15 per point; 95% CI 1.87–11.32; p = 0.002). In the fully adjusted model including ICH Score, BGH volume, and BGH roundness, both the ICH Score and roundness remained independently associated with mortality, whereas BGH volume lost statistical significance (Table 2). The ICH Score alone demonstrated good discrimination for mortality (AUC 0.813). The combined model including ICH Score, BGH volume, and BGH roundness yielded an AUC of 0.866 (Fig. 5). Although this represented a numerical increase, the difference did not reach statistical significance (DeLong p = 0.28).
Fig. 5.
Receiver operating characteristic (ROC) curves comparing discrimination for in-hospital mortality between the ICH Score alone and the combined model including the ICH Score, BGH volume, and BGH roundness. Area under the curve (AUC) values are indicated
The adjusted predicted probability of mortality across the range of basal ganglia roundness is shown in Fig. 6, demonstrating a steep increase in mortality risk with decreasing roundness, independent of ICH Score and hematoma volume. Model performance was supported by a Nagelkerke R2 of 0.52. Hosmer–Lemeshow testing indicated good calibration (χ2 = 4.14, p = 0.84). Internal validation using bootstrap resampling (1000 iterations) confirmed coefficient stability, with the percentile-based 95% confidence interval for the roundness coefficient not crossing zero (− 4.45 to − 0.39). In an additional sensitivity analysis including IVH volume and sex as covariates, the association between BGH roundness and mortality remained robust (OR 0.21 per 0.1 increase, 95% CI 0.03–0.74). IVH volume was also independently associated with mortality, whereas BGH volume remained non-significant. Sex was not independently associated with mortality and did not materially alter the effect estimates.
Fig. 6.
Adjusted predicted probability of in-hospital mortality according to BGH roundness. Predicted probability curve derived from the multivariable logistic regression model including the ICH Score and BGH volume. The shaded area represents the 95% confidence interval. Tick marks along the x-axis indicate observed patient values
Functional outcome
For unfavorable functional outcome at discharge (mRS 5–6), larger BGH volume was significantly associated with poor outcome (OR 1.08 per cm3 increase; p = 0.003), whereas roundness did not reach statistical significance (OR 0.48 per 0.1 increase; 95% CI 0.20–1.02; p = 0.073). The distribution of hematoma volume stratified by functional outcome is illustrated in Fig. 7, demonstrating substantially larger volumes among patients with unfavorable outcome compared with those with favorable outcome at discharge. In an additional analysis using mRS 4–6 as the definition of unfavorable outcome, a similar distribution pattern was observed. However, in multivariable analysis, neither roundness nor hematoma volume was independently associated with outcome after adjustment for the ICH Score.
Fig. 7.
Basal ganglia hemorrhage volume by functional outcome at discharge using two outcome definitions. A Unfavorable outcome defined as mRS 5–6. B Unfavorable outcome defined as mRS 4–6. Patients with unfavorable outcome exhibited larger hematoma volumes compared with those with favorable outcome. Boxes represent the interquartile range with the median; dots represent individual observations
Discussion
In this retrospective cohort of patients with deep supratentorial ICH and concomitant IVH, lower hemorrhage roundness was independently associated with in-hospital mortality. This association remained after adjustment for continuous hematoma volume and the ICH Score. In contrast, hematoma volume showed the strongest association with unfavorable functional outcome at discharge, whereas roundness was not independently associated with functional outcome at discharge. These findings suggest that morphologic irregularity may be more strongly associated with early mortality risk than with short-term functional outcome, which may be influenced by additional factors and includes both severely disabled survivors and deceased patients.
Morphology beyond volume
Hematoma volume is a well-established predictor of outcome in spontaneous ICH and a core component of the ICH Score [7, 18]. However, volume reflects overall hemorrhage size without describing geometric configuration. Furthermore, the measurement is based on the ABC/2 formula, which may introduce significant bias in the results. In our cohort, roundness showed discrimination comparable to volume in univariable analysis and remained independently associated with mortality in multivariable models.
Previous imaging studies have suggested that hemorrhage shape may reflect underlying bleeding dynamics and mechanical expansion patterns [15]. However, quantitative evaluation of geometric descriptors in spontaneous ICH has been limited. Our findings extend this concept by demonstrating that a reproducible shape metric remains associated with mortality after adjustment for both continuous volume and the ICH Score.
These results indicate that geometric irregularity may capture structural characteristics not reflected by size alone [9]. Reduced sphericity has previously been associated with poorer postoperative recovery in dorsally located spinal meningiomas and with postoperative cranial nerve function and proliferative activity in cranial meningiomas [21, 22]. Moreover, radiomic shape metrics have demonstrated prognostic relevance in subdural hematoma cohorts, including the KEPPRA and PROMISE analyses [5, 6].
Prognostic information beyond the ICH score
Whether radiomic shape features add information beyond established clinical risk scores is clinically relevant. The ICH Score incorporates hematoma volume in dichotomized form and remains a robust predictor of mortality [17]. Although modeling continuous volume provides a more detailed measure of hemorrhage extent, geometric parameters such as roundness may reflect structural characteristics not fully captured by size alone [15].
The persistence of an association between shape metrics and mortality after adjustment for established predictors suggests partial independence from conventional risk factors [23]. However, the absence of a significant improvement in overall discrimination indicates that radiomic morphology does not replace established clinical scores. Instead, shape-based descriptors may add information beyond established clinical scores. Overall, volume and morphology appear to describe different aspects of hemorrhage geometry [10]. While volume quantifies extent, configuration may relate to spatial expansion patterns and structural complexity.
Mortality versus functional outcome
Morphologic and volumetric parameters showed different associations with outcome type. Roundness was independently associated with in-hospital mortality but not with unfavorable functional outcome at discharge. This may reflect that functional outcome at discharge is a more heterogeneous endpoint influenced by factors beyond hemorrhage characteristics. Furthermore, combining mRS 5 and 6 may dilute associations more closely related to mortality risk. In an additional analysis using mRS 4–6, neither roundness nor hematoma volume was independently associated with outcome after adjustment for the ICH Score, suggesting that broader functional outcome definitions at discharge may dilute associations more closely related to mortality. In contrast, BGH volume demonstrated strong discrimination for functional outcome. This difference may reflect that early mortality in deep ICH is often related to acute mass effect and rapid neurological decline, whereas functional status at discharge is more closely linked to overall tissue damage and ventricular extension [2, 13, 14, 24]. Accordingly, geometric configuration may be more relevant for early fatal course, while total hemorrhage extent appears more closely related to short-term disability. Given the limited sample size, differences between outcome definitions should be interpreted with caution, as some observed effects may reflect statistical variability rather than true biological differences.
Clinical and methodological implications
Quantitative shape analysis can be derived from standard non-contrast CT imaging using semi-automated segmentation and predefined geometric descriptors. Restricting the analysis to a limited number of shape features reduces dimensionality and limits the risk of overfitting compared with high-dimensional radiomics approaches. Roundness is directly interpretable and reproducible. In contrast to texture-based features, geometric descriptors are less sensitive to scanner variability and reconstruction parameters, which may support more consistent application across centers [1]. Radiomic shape metrics should therefore be considered as an adjunct to established clinical predictors rather than a replacement.
Limitations
This study has several limitations. Its retrospective single-center design may limit generalizability, and external validation in independent cohorts is required before clinical translation. Due to the retrospective design, information on withdrawal of care or treatment limitations was not systematically available; therefore, in-hospital mortality could not be differentiated with respect to treatment decisions. Follow-up beyond discharge was not available, and the analysis focused on in-hospital outcomes. The sample size was modest, with 17 mortality events. Although predictor selection was restricted and internal bootstrap validation was performed, some degree of overfitting cannot be excluded. Hematoma segmentation was semi-automated with manual refinement. While conducted according to a standardized protocol, interobserver variability was not formally assessed. The analysis was limited to predefined shape-based features. This approach reduced dimensionality and improved interpretability but did not explore potential contributions from texture- or intensity-based radiomics. Functional outcome was assessed at hospital discharge rather than at 90 days, which is more commonly used in stroke research. Early mRS may not fully reflect longer-term recovery. Finally, the cohort was restricted to deep supratentorial ICH with intraventricular extension, and the findings may not be generalizable to lobar hemorrhages or cases without ventricular involvement.
Conclusion
In deep supratentorial ICH with IVH, hemorrhage roundness was independently associated with in-hospital mortality after adjustment for hematoma volume and the ICH Score. In contrast, hematoma volume showed a stronger relationship with unfavorable functional outcome. These findings suggest that geometric configuration and volumetric extent reflect different aspects of hemorrhage severity. Shape-based analysis derived from routine CT imaging may add information beyond established predictors. Larger studies with external validation are required to determine the role of geometric descriptors in prognostic modeling of spontaneous ICH.
Supplementary Information
Below is the link to the electronic supplementary material.
Funding
Open Access funding enabled and organized by Projekt DEAL. This publication was supported by the Open Access Publishing Fund of Leipzig University.
Data availability
The datasets generated and/or analysed during the current study are not publicly available due to institutional data protection regulations but are available from the corresponding author on reasonable request.
Declarations
Conflicts of interest
The authors declare that they have no competing interests.
Ethical approval
This study was conducted in accordance with the principles of the Declaration of Helsinki. Ethical approval was obtained from the Ethics Committee of the University of Leipzig (protocol number 073/26-ek; date of approval: 13 April 2026).
References
- 1.Barras CD, Tress BM, Christensen S, MacGregor L, Collins M, Desmond PM, Skolnick BE, Mayer SA, Broderick JP, Diringer MN, Steiner T, Davis SM (2009) Density and shape as CT predictors of intracerebral hemorrhage growth. Stroke 40:1325–1331 [DOI] [PubMed] [Google Scholar]
- 2.Christensen MC, Morris S, Vallejo-Torres L, Vincent C, Mayer SA (2012) Neurological impairment among survivors of intracerebral hemorrhage: the FAST Trial. Neurocrit Care 16:224–231 [DOI] [PubMed] [Google Scholar]
- 3.Dierksen F, Tran AT, Zeevi T, Maier IL, Qureshi AI, Sanelli PC, Werring DJ, Malhotra A, Falcone GJ, Sheth KN, Payabvash S (2024) Peri-hematomal edema shape features related to 3-month outcome in acute supratentorial intracerebral hemorrhage. Eur Stroke J 9:383–390 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Eslami V, Tahsili-Fahadan P, Rivera-Lara L, Gandhi D, Ali H, Parry-Jones A, Nelson LS, Thompson RE, Nekoobakht-Tak S, Dlugash R, McBee N, Awad I, Hanley DF, Ziai WC (2019) Influence of intracerebral hemorrhage location on outcomes in patients with severe intraventricular hemorrhage. Stroke 50:1688–1695 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Guranda A, Richter A, Wach J, Güresir E, Vychopen M (2025) KEPPRA: key epilepsy prognostic parameters with radiomics in acute subdural hematoma before craniotomy. Brain Sci. 10.3390/brainsci15020204 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Guranda A, Richter A, Wach J, Güresir E, Vychopen M (2025) PROMISE: prognostic radiomic outcome measurement in acute subdural hematoma evacuation post-craniotomy. Brain Sci. 10.3390/brainsci15010058 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Hemphill JC 3rd, Bonovich DC, Besmertis L, Manley GT, Johnston SC (2001) The ICH score: a simple, reliable grading scale for intracerebral hemorrhage. Stroke 32:891–897 [DOI] [PubMed] [Google Scholar]
- 8.Huang X, Wang D, Zhang Q, Ma Y, Zhao H, Li S, Deng J, Ren J, Yang J, Zhao Z, Xu M, Zhou Q, Zhou J (2022) Radiomics for prediction of intracerebral hemorrhage outcomes: a retrospective multicenter study. Neuroimage Clin 36:103242 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Ji Z, Hao Y, Gao B, Zhang X, Zhang Y, Jia J, Xia X, Guo Y, Li S, Wu J, Kang K, Zhao X (2025) Quantitative shape irregularity and density heterogeneity predict hematoma expansion in patients with intracerebral hemorrhage. Ann Clin Transl Neurol 12:2068–2076 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Jianbo C, Ting X, Yihao C, Xiaoning W, Hong S, Qinghua Z, Zeju Y, Xingong W, Fengxuan T, Jianjun C, Wenbin M, Junji W, Ming F, Yao J, Renzhi W (2021) The patterns of morphological change during intracerebral hemorrhage expansion: a multicenter retrospective cohort study. Front Med (Lausanne) 8:774632 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Liu C, Zhang H, Wang L, Jiang Q, Lu E, Yuan C, Liang Y, Sun Z, Xiang H, Xu X, Sun J, Fu B, Zhao B, Zhang D, Chen X, Wang N, Wang L, Yang G (2022) Irregular shape as an independent predictor of prognosis in patients with primary intracerebral hemorrhage. Sci Rep 12:8552 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.LoPresti MA, Bruce SS, Camacho E, Kunchala S, Dubois BG, Bruce E, Appelboom G, Connolly ES Jr. (2014) Hematoma volume as the major determinant of outcomes after intracerebral hemorrhage. J Neurol Sci 345:3–7 [DOI] [PubMed] [Google Scholar]
- 13.Lord AS, Gilmore E, Choi HA, Mayer SA (2015) Time course and predictors of neurological deterioration after intracerebral hemorrhage. Stroke 46:647–652 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Mayer SA, Thomas CE, Diamond BE (1996) Asymmetry of intracranial hemodynamics as an indicator of mass effect in acute intracerebral hemorrhage. a transcranial doppler study. Stroke 27:1788–1792 [DOI] [PubMed] [Google Scholar]
- 15.Oge DD, Topcuoglu MA, Gocmen R, Arsava EM (2020) The dynamics of hematoma surface regularity and hematoma expansion in acute intracerebral hemorrhage. J Clin Neurosci 74:160–163 [DOI] [PubMed] [Google Scholar]
- 16.Phung JYH, Yogendrakumar V, Dowlatshahi D (2025) Acute spontaneous intracerebral hemorrhage management update: five new things you should know. Ann Indian Acad Neurol 28:155–158 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Rodríguez-Fernández S, Castillo-Lorente E, Guerrero-Lopez F, Rodríguez-Rubio D, Aguilar-Alonso E, Lafuente-Baraza J, Gómez-Jiménez FJ, Mora-Ordóñez J, Rivera-López R, Arias-Verdú MD, Quesada-García G, Arráez-Sánchez M, Rivera-Fernández R (2018) Validation of the ICH score in patients with spontaneous intracerebral haemorrhage admitted to the intensive care unit in Southern Spain. BMJ Open 8:e021719 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Rotzel HB, Serrano Lázaro A, Aguillón Prada D, Mesejo Arizmendi A, Sanchís Piqueras C, Rodriguez Gimillo M, Blasco Cortés ML, Carbonell Monleón MN, Díaz Barina O (2015) Hematoma volume and other prognostic factors with mortality in spontaneous intracerebral hemorrhage. Intensive Care Med Exp. 10.1186/2197-425x-3-s1-a981 [Google Scholar]
- 19.Safatli DA, Günther A, Schlattmann P, Schwarz F, Kalff R, Ewald C (2016) Predictors of 30-day mortality in patients with spontaneous primary intracerebral hemorrhage. Surg Neurol Int 7:S510-517 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Tenhoeve SA, Findlay MC, Cole KL, Gautam D, Nelson JR, Brown J, Orton CJ, Bounajem MT, Brandel MG, Couldwell WT, Rennert RC (2024) The clinical potential of radiomics to predict hematoma expansion in spontaneous intracerebral hemorrhage: a narrative review. Front Neurol 15:1427555 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Vychopen M, Arlt F, Wilhelmy F, Seidel C, Barrantes-Freer A, Güresir E, Wach J (2023) Association of quantitative radiomic shape features with functional outcome after surgery for primary sporadic dorsal spinal meningiomas. Front Surg 10:1303128 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Wach J, Naegeli J, Vychopen M, Seidel C, Barrantes-Freer A, Grunert R, Güresir E, Arlt F (2023) Impact of shape irregularity in medial sphenoid wing meningiomas on postoperative cranial nerve functioning, proliferation, and progression-free survival. Cancers (Basel). 10.3390/cancers15123096 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Wang CW, Liu YJ, Lee YH, Hueng DY, Fan HC, Yang FC, Hsueh CJ, Kao HW, Juan CJ, Hsu HH (2014) Hematoma shape, hematoma size, Glasgow coma scale score and ICH score: which predicts the 30-day mortality better for intracerebral hematoma? PLoS ONE 9:e102326 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Zhou HX, Hao N, Xu XL (2018) Related factors of early mortality in young adults with cerebral hemorrhage. Open Med (Wars) 13:214–220 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The datasets generated and/or analysed during the current study are not publicly available due to institutional data protection regulations but are available from the corresponding author on reasonable request.