Higher Serum Levels of Lactate Dehydrogenase Before Microsurgery Predict Poor Outcome of Aneurysmal Subarachnoid Hemorrhage

Deep learning and machine learning predictive models for neurological function after interventional embolization of intracranial aneurysms

Objective

The objective of this study is to develop a model to predicts the postoperative Hunt-Hess grade in patients with intracranial aneurysms by integrating radiomics and deep learning technologies, using preoperative CTA imaging data. Thereby assisting clinical decision-making and improving the assessment and prognosis of postoperative neurological function.

Methods

This retrospective study encompassed 101 patients who underwent aneurysm embolization surgery. 851 radiomic features were extracted from CTA images. 512 deep learning features are extracted from last layer of ResNet50 deep convolutional neural network model. The feature screening process pipeline encompassed intraclass correlation coefficient analysis, principal component analysis, U test, spearman correlation analysis, minimum redundancy maximum relevance algorithm and Lasso regression, to identify features most correlated with postoperative Hunt-Hess grading. In the model construction phase, three distinct models were constructed: radiomics feature-based model (RSM), deep learning feature-based model (DLM), and deep learning-radiomics feature fusion model (DLRSCM). The study also calculated the radiomics score and combined it with clinical data to construct a Nomogram for predictive modeling. DLM, RSM and DLRSCM model was constructed by 9 base algorithms and 1 ensemble learning algorithm - Stacking ensemble model. Model performance was evaluated based on the area under the Receiver Operating Characteristic (ROC) curve (AUC), Matthews Correlation Coefficient (MCC), calibration curves, and decision curves analysis.

Results

5 significant radiomic feature and 4 significant deep learning features were obtained through the feature selection process. These features were utilized for model construction. Bootstrap resampling method was used for internal validation of the models. In terms of model evaluation, the DLM model, the stacking ensemble algorithm results achieved an AUC of 0.959 and MCC of 0.815. In the RSM model, the stacking ensemble model AUC was 0.935 and MCC was 0.793. The stacking ensemble model in DLRSCM outperformed others, with an AUC of 0.968 and MCC of 0.820. Results indicated that the ANN performed optimally among all base models, while the stacked ensemble learning model exhibited the highest predictive performance.

Conclusion

This study demonstrates that the combination of radiomics and deep learning is an effective approach to predict the postoperative Hunt-Hess grade in patients with intracranial aneurysms. This holds significant value in the early identification of postoperative neurological complications and in enhancing clinical decision-making.

Blood inflammatory biomarkers predict in-hospital pneumonia after endovascular treatment of aneurysm in patients with aneurysmal subarachoid hemorrhage

The systemic inflammatory response index (SIRI) is a well-known marker of systemic inflammation reflecting the body's inflammatory/immune state. The study aimed to evaluate the relationship between the SIRI on admission and aneurysmal subarachnoid hemorrhage (aSAH)-associated pneumonia and compare with other currently used bio-markers. We reviewed 562 successive patients with aneurysmal SAH who underwent endovascular treatment between January 2019 and September 2021. ASAH-associated pneumonia was diagnosed using the modified Centers for Disease Control and Prevention criteria. The SIRI on admission was calculated as monocyte count × neutrophil count / lymphocyte count. Multiple logistic regression models were used for data analysis. A total of 158 (28.11%) patients developed aSAH-associated pneumonia. Using the Multiple logistic regression analysis, a notable dose-response association was found between the elevated SIRI (fourth quartile) and aSAH-associated pneumonia (adjusted odds ratio = 6.759; 95% confidence interval [CI], 3.280-13.930; p < 0.001 [p for trend < 0.001]). The SIRI (0.701, 95% CI: 0.653-0.749) presented a higher area under the curve (AUC) than systemic immune- inflammation index (SII) (0.669, 95% CI: 0.620-0.718) (p = 0.089); neutrophil-to-lymphocyte ratio (NLR) (0.665, 95% CI: 0.616-0.714) (p = 0.035) and platelet-lymphocyte ratio (PLR) (0.587, 95% CI: 0.534-0.641) (p < 0.001). A higher SIRI on admission was associated with aSAH-associated pneumonia, which may guide further clinical trials of prophylactic antibiotic therapy.

Serum LDH levels may predict poor neurological outcome after aneurysmal subarachnoid hemorrhage

INTRODUCTION

Serum lactate dehydrogenase (LDH) levels are often elevated in cardiovascular diseases. Their prognostic role after subarachnoid hemorrhage (SAH) remains poorly evaluated.

METHODS

This is a retrospective single-center study of patients with non-traumatic SAH admitted to the intensive care unit (ICU) of an University Hospital from 2007 to 2022. Exclusion criteria were pregnancy and incomplete medical records or follow-up data. Baseline information, clinical data, radiologic data, the occurrence of neurological complications as well as serum LDH levels during the first 14 days of ICU stay were collected. Unfavorable neurological outcome (UO) at 3 months was defined as a Glasgow Outcome Scale of 1-3.

RESULTS

Five hundred and forty-seven patients were included; median serum LDH values on admission and the highest LDH values during the ICU stay were 192 [160-230] IU/L and 263 [202-351] IU/L, respectively. The highest LDH value was recorded after a median of 4 [2-10] days after ICU admission. LDH levels on admission were significantly higher in patients with UO. When compared with patients with favorable outcome (FO), patients with UO had higher serum LDH values over time. In the multivariate logistic regression model, the highest LDH value over the ICU stay (OR 1.004 [95% CI 1.002 - 1.006]) was independently associated with the occurrence of UO; the area under the receiving operator (AUROC) curve for the highest LDH value over the ICU stay showed a moderate accuracy to predict UO (AUC 0.76 [95% CI 0.72-0.80]; p < 0.001), with an optimal threshold of > 272 IU/L (69% sensitivity and 74% specificity).

CONCLUSIONS

The results in this study suggest that high serum LDH levels are associated with the occurrence of UO in SAH patients. As a readily and available biomarker, serum LDH levels should be evaluated to help with the prognostication of SAH patients.

A 90-Day Prognostic Model Based on the Early Brain Injury Indicators after Aneurysmal Subarachnoid Hemorrhage: the TAPS Score

This study aimed to establish a new scoring model based on the early brain injury (EBI) indicators to predict the 90-day functional outcomes in patients with aneurysmal subarachnoid hemorrhage (aSAH). We retrospectively enrolled 825 patients and prospectively enrolled 108 patients with aSAH who underwent surgical clipping or endovascular coiling (derivation cohort = 640; validation cohort = 185; prospective cohort = 108) in our institute. We established a logistic regression model based on independent risk factors associated with 90-day unfavorable outcomes. The discrimination of the prognostic model was assessed by the area under the curve in a receiver operating characteristic curve analysis. The Hosmer-Lemeshow goodness-of-fit test and a calibration plot were used to evaluate the calibration of the prediction model. The developed scoring model named "TAPS" (total score, 0-7 points) included the following admission variables: age > 55 years old, WFNS grade of 4-5, mFS grade of 3-4, Graeb score of 5-12, white blood cell count > 11.28 × 10⁹/L, and surgical clipping. The model showed good discrimination with the area under the curve in the derivation, validation, and prospective cohorts which were 0.816 (p < 0.001, 95%CI = 0.77-0.86), 0.810 (p < 0.001, 95%CI = 0.73-0.90), and 0.803 (p < 0.001, 95%CI = 0.70-0.91), respectively. The model also demonstrated good calibration (Hosmer-Lemeshow goodness-of-fit test: X² = 1.75, df = 8, p = 0.988). Compared with other predictive models, TAPS is an easy handle tool for predicting the 90-day unfavorable outcomes of aSAH patients, which can help clinicians better understand the concept of EBI and quickly identify those patients at risk of poor prognosis, providing more positive treatment strategies. Trial registration: NCT04785976. Registered 5 March 2021-retrospectively registered, http://www.clinicaltrials.gov .

Lactate dehydrogenase predicting mortality in patients with aneurysmal subarachnoid hemorrhage

Objective

Lactate dehydrogenase (LDH) has been reported to be associated with outcomes after surgery in patients with aneurysmal subarachnoid hemorrhage (aSAH), but it is unclear if this is independent from other biomarkers and across all aSAH treatments. This study aims to assess whether LDH is an independent predictor of mortality in patients with aSAH and test whether the inclusion of LDH in a well‐established prediction model can improve discrimination and reclassification.

Methods

This was a retrospective observational study at a tertiary academic medical center. This study measured baseline LDH levels taken at admission and longitudinal LDH levels (up to a month postadmission) to assess median, max, and trajectory LDH levels. The primary outcome was mortality at 90 days. Multivariable regression analyses were used to evaluate associations between LDH and outcomes. The full original Subarachnoid Hemorrhage International Trialists' (SAHIT) model was used as the reference model.

Results

In total, 3524 patients with aSAH were included. LDH at admission was independently associated with mortality at 90 days (quartile 4 vs. 1: odds ratio 1.60; 95% CI 1.08–2.37) and mortality at the longest follow‐up (quartile 4 vs. 1: hazard ratio1.72; 95% CI 1.34–2.20). Compared with the SAHIT model, the addition of three LDH (admission, max, and median) levels to the SAHIT model significantly improved the area under the curve and categorical net reclassification improvement for prediction mortality.

Interpretation

In patients with aSAH, LDH level is an independent predictor of all‐cause mortality. The incorporation of LDH into a well‐established prediction model improved the ability to predict the risk of death in patients with aSAH.

Clinical Value and Prognosis of C Reactive Protein to Lymphocyte Ratio in Severe Aneurysmal Subarachnoid Hemorrhage

Objective

To investigate the relationship between CLR and disease severity and clinical prognosis of aSAH.

Methods

The authors retrospectively analyzed the clinical data of 221 patients with aSAH, who were admitted to the intensive care unit from January 2017 to December 2020. The indicators of inflammatory factors in the first blood routine examination within 48 h of bleeding were obtained. The prognosis was evaluated by mRS score at discharge, mRS>2 was a poor outcome. Through the receiver operating characteristic (ROC) curve, the area under the curve was calculated and the predicted values of inflammatory factors (CLR, CRP, WBC, and neutrophils) were compared. Univariate and multivariable logistic regression analyses were used to evaluate the relationship between CLR and the clinical prognosis of patients. ROC curve analysis was performed to determine the optimal cut-off threshold, sensitivity, and specificity of CLR in predicting prognosis at admission.

Results

According to the mRS score at discharge, 139 (62.90%) patients were classified with poor outcomes (mRS>2). The inflammatory factor with the best predictive value was CLR, which had an optimal cut-off threshold of 10.81 and an area under the ROC curve of 0.840 (95%CI.788–0.892, P < 0.001). Multivariable Logistic regression analysis showed that the Modified Fisher grade, Hunt-Hess grade, and CLR at admission were independent risk factors for poor outcomes of patients with aSAH (P < 0.05). According to Hunt-Hess grade, patients were divided into a mild group (Hunt-Hess ≤ 3) and a severe group (Hunt-Hess > 3), and the CLR value was significantly higher in severe patients with aSAH than in mild patients. The optimal cut-off threshold of CLR in the severe group was 6.87, and the area under the ROC curve was 0.838 (95% CI.752–0.925, P < 0.001).

Conclusions

The CLR value at the admission of patients with aSAH was significantly associated with Hunt-Hess grade, The higher Hunt-Hess grade, the higher the CL R-value, and the worse the prognosis. Early CLR value can be considered as a feasible biomarker to predict the clinical prognosis of patients with aSAH.

Serum Lactate Dehydrogenase to Phosphate Ratio as an Independent Predictor for Adverse Outcome of Microsurgical Clipping for Ruptured Intracranial Aneurysm: A Propensity-Score Matching Analysis

Objective: In this study, we assessed the correlation between the lactate dehydrogenase (LDH) to phosphate ratio and the prognosis of microsurgical clippings for ruptured intracranial aneurysm (rIA) to test the hypothesis that the serum LDH to phosphate ratio could be a predictor of the outcome of microsurgical clipping for rIA. Methods: Records of rIA patients between 2012 and 2018 were retrospectively collected. Age, sex, Hunt-Hess grade, Fisher grade, medical history, aneurysm location, hydrocephalus, laboratory data including serum LDH, phosphate, and LDH to phosphate ratio, related complications, and the outcomes in 3 months were recorded. Results: A total of 1608 rIA patients in our institution were collected, and 856 patients treated by microsurgical clipping were enrolled. On admission, a significantly higher LDH-phosphate ratio was observed in patients with poor outcomes at 3 months (median ± SD, 200.175 ± 107.290 for mRS 0−2 vs. 323.826 ± 219.075 for mRS score 3−6; p = 0.000). An LDH to phosphate ratio of 226.25 in the receiver operating characteristic (ROC) curve was the optimal cutoff value to discriminate between good and poor outcomes at 3 months. The LDH to phosphate ratio ≥ 226.25 on admission was independently correlated with poor outcomes in rIA patients. In addition, Hunt and Hess grade, Fisher grade, pneumonia, and DIND were also independently correlated with poor outcomes. After removing the bias in essential clinical variables between patients with LDH to phosphate, ratio ≥ 226.25 versus <226.25 by PSM, the number of patients with poor outcomes at 3 months increased in patients with an LDH to phosphate ratio of ≥226.25 (p = 0.005). Conclusions: The LDH to phosphate ratio was a potential biomarker and could predict the unfavorable outcome of microsurgical clipping for rIA in 3 months, related to neuronal damage, cerebral hypoxia, and early brain injury after aneurysm ruptures.

Association of laboratory markers and cerebral blood flow among sickle cell anemia children

BACKGROUND

Stroke is one of the highest complications of sickle-cell anemia (SCA). The Transcranial Doppler (TCD) has been adopted worldwide as a gold standard method for detecting alterations in the blood velocity in cerebral arteries. In this study, we investigated the association between laboratory parameters and increased cerebral blood flow velocity in Brazilian SCA pediatric patients.

METHODS

The study included 159 pediatric patients with SCA, submitted to TCD velocity screening, and the time-averaged maximum mean velocity (TAMMV) was determined in the middle cerebral artery (MCA), anterior cerebral artery (ACA), and distal intracranial internal carotid artery (ICA). We compared cerebral blood flow in patients stratified by the following: TCD1-defined as normal, with TAMMV inferior to 170 cm/s; TCD2-conditional, with TAMMV above 170 cm/s, but less than 199 cm/s; TCD3-altered, with TAMMV greater than or equal to 200 cm/s.

RESULTS

TAMMV was negatively correlated with age and weight (p < 0.05). Moreover, TAMMV was associated or correlated with reductions in HbF, RBC, hemoglobin, hematocrit, HDL, and haptoglobin and, increases in MCV, MCH, RDW, reticulocytes, WBC, lymphocytes, monocytes, eosinophils, total and indirect bilirubin, LDH, AST, ALT, glucose, ferritin, and AAT (p < 0.05).

CONCLUSION

The current study highlights the importance of the investigation of hemolytic and inflammatory biomarkers for monitoring the clinical outcome of SCA pediatric patients, to avoid acute or chronic stroke. Moreover, glucose and HDL-C appear useful for predicting higher TAMMV.