A Neural Network Model Using Pain Score Patterns to Predict the Need for Outpatient Opioid Refills Following Ambulatory Surgery: Algorithm Development and Validation.

BACKGROUND: Expansion of clinical guidance tools is crucial to identify patients at risk of requiring an opioid refill after outpatient surgery. OBJECTIVE: The objective of this study was to develop machine learning algorithms incorporating pain and opioid features to predict the need for outpatient opioid refills following ambulatory surgery. METHODS: Neural networks, regression, random forest, and a support vector machine were used to evaluate the data set. For each model, oversampling and undersampling techniques were implemented to balance the data set. Hyperparameter tuning based on k-fold cross-validation was performed, and feature importance was ranked based on a Shapley Additive Explanations (SHAP) explainer model. To assess performance, we calculated the average area under the receiver operating characteristics curve (AUC), F1-score, sensitivity, and specificity for each model. RESULTS: There were 1333 patients, of whom 144 (10.8%) refilled their opioid prescription within 2 weeks after outpatient surgery. The average AUC calculated from k-fold cross-validation was 0.71 for the neural network model. When the model was validated on the test set, the AUC was 0.75. The features with the highest impact on model output were performance of a regional nerve block, postanesthesia care unit maximum pain score, postanesthesia care unit median pain score, active smoking history, and total perioperative opioid consumption. CONCLUSIONS: Applying machine learning algorithms allows providers to better predict outcomes that require specialized health care resources such as transitional pain clinics. This model can aid as a clinical decision support for early identification of at-risk patients who may benefit from transitional pain clinic care perioperatively in ambulatory surgery.

An Ensemble Learning Approach to Improving Prediction of Case Duration for Spine Surgery: Algorithm Development and Validation.

BACKGROUND: Estimating surgical case duration accurately is an important operating room efficiency metric. Current predictive techniques in spine surgery include less sophisticated approaches such as classical multivariable statistical models. Machine learning approaches have been used to predict outcomes such as length of stay and time returning to normal work, but have not been focused on case duration. OBJECTIVE: The primary objective of this 4-year, single-academic-center, retrospective study was to use an ensemble learning approach that may improve the accuracy of scheduled case duration for spine surgery. The primary outcome measure was case duration. METHODS: We compared machine learning models using surgical and patient features to our institutional method, which used historic averages and surgeon adjustments as needed. We implemented multivariable linear regression, random forest, bagging, and XGBoost (Extreme Gradient Boosting) and calculated the average R2, root-mean-square error (RMSE), explained variance, and mean absolute error (MAE) using k-fold cross-validation. We then used the SHAP (Shapley Additive Explanations) explainer model to determine feature importance. RESULTS: A total of 3189 patients who underwent spine surgery were included. The institution's current method of predicting case times has a very poor coefficient of determination with actual times (R2=0.213). On k-fold cross-validation, the linear regression model had an explained variance score of 0.345, an R2 of 0.34, an RMSE of 162.84 minutes, and an MAE of 127.22 minutes. Among all models, the XGBoost regressor performed the best with an explained variance score of 0.778, an R2 of 0.770, an RMSE of 92.95 minutes, and an MAE of 44.31 minutes. Based on SHAP analysis of the XGBoost regression, body mass index, spinal fusions, surgical procedure, and number of spine levels involved were the features with the most impact on the model. CONCLUSIONS: Using ensemble learning-based predictive models, specifically XGBoost regression, can improve the accuracy of the estimation of spine surgery times.

Association of race and ethnicity in the receipt of regional anesthesia following mastectomy.

BACKGROUND: Regional anesthetic techniques have become increasingly used for the purpose of pain management following mastectomy. Although a variety of beneficial techniques have been described, the delivery of regional anesthesia following mastectomy has yet to be examined for racial or ethnic disparities. We aimed to examine the association of race and ethnicity on the delivery of regional anesthesia in patients undergoing surgical mastectomy using a large national database. METHODS: We used the American College of Surgeons-National Surgical Quality Improvement Program database to identify adult patients aged ≥18 years old who underwent mastectomy from 2014 to 2016. We reported unadjusted estimates of regional anesthesia accordingly to race and ethnicity and examined differences in sociodemographic characteristics and health status. Multivariable logistic regression was used to report the association of race and ethnicity with use of regional anesthesia. RESULTS: A total of 81 345 patients who underwent mastectomy were included, 14 887 (18.3%) of whom underwent regional anesthesia. The unadjusted rate of use of regional anesthesia was 18.9% for white patients, 16.8% for black patients, 15.6% for Asian patients, 16.5% for Native Hawaiian/Pacific Islander patients, 17.8% for American Indian or Alaska Native and 17.4% for unknown race (p<0.001). With respect to ethnicity, the unadjusted rate of regional anesthesia use was 18.4% for non-Hispanic patients vs 16.1% for Hispanic patients vs 18.6% for the unknown ethnicity cohort (p<0.001). On multivariable logistic regression analysis, the odds of receipt of regional anesthesia was 12% lower in black patients and 21% lower in Asian patients compared with white patients (p<0.001). The odds of regional anesthesia use were 13% lower in Hispanic compared with non-Hispanic patients (p<0.001). CONCLUSION: Black and Asian patients had lower odds of undergoing regional anesthesia following mastectomy compared with white counterparts. In addition, Hispanic patients had lower odds of undergoing regional anesthesia than non-Hispanic counterparts. These differences underlie the importance of working to deliver equitable healthcare irrespective of race or ethnicity.

Utilization of regional anesthesia for open reduction and internal fixation of clavicular fractures is associated with increased same-day discharge and not readmissions.

BACKGROUND: Regional anesthesia is becoming increasingly used for analgesia in clavicular surgery. The purpose of the study is to examine same-day discharge and readmissions in patients with clavicular fractures receiving open reduction and internal fixation, comparing those who received regional anesthesia to those that did not. METHODS: The National Surgical Quality Improvement Program registry was used to build a retrospective cohort with the primary and secondary outcomes of same-day discharge and readmission rates, respectively. The independent variable was type of anesthesia (ie, regional vs general). Unadjusted univariable followed by multivariable logistic regression analysis for the association of each covariate with the primary and secondary outcome was performed. A p<0.05 was considered statistically significant. RESULTS: A total of 2300 patients were included, in which 346 (15.0%) of received a nerve block. The only factor associated with same-day discharge on crude analysis was the use of a nerve block (OR 1.70, 95% CI 1.21 to 2.40, p<0.01). Factors not associated with same-day discharge included geriatric age (OR 0.39, 95% CI 0.26 to 0.57, p<0.01), diabetes mellitus (OR 0.54, 95% CI 0.33 to 0.89, p=0.01), dependent functional status (OR 0.27, 95% CI 0.11 to 0.69, p<0.01), hypertension (OR 0.39, 95% CI 0.29 to 0.52, p<0.01), bleeding disorder (OR 0.24, 95% CI 0.10 to 0.59, p<0.01) and an American Society of Anesthesiologists Physical Status score ≥3 (OR 0.33, 95% CI 0.25 to 0.45, p<0.01). On adjusted analysis, regional anesthesia was associated with increased odds of same-day discharge (OR 1.67, 95% CI 1.18 to 2.37, p<0.01), but not 30-day readmission (OR 0.57, 95% CI 0.13 to 2.52, p=0.46). CONCLUSION: Regional anesthesia use in patients with clavicular fractures is associated with higher odds for same-day discharge. There is no significant difference in the rates of hospital readmission within a 30-day period. LEVEL OF EVIDENCE: IV, retrospective case series.

Attenuation of brain response to vascular endothelial growth factor-mediated angiogenesis and neurogenesis in aged mice.

BACKGROUND AND PURPOSE: Alterations of neuroangiogenic response play important roles in the development of aging-related neurodisorders and affect gene-based therapies. We tested brain response to vascular endothelial growth factor (VEGF) in aged mice. METHODS: Adeno-associated viral vector (AAV)-VEGF, an adeno-associated viral vector expressing VEGF, was injected into the brain of 3-, 12-, and 24-month-old mice. AAV-LacZ-injected mice were used as controls (n=6). Before euthanasia at 6 weeks after vector injection, the mice were intraperitoneally injected with 5-bromodeoxyuridine for 3 consecutive days. The vascular density and the number of neuroprogenitors were analyzed. RESULTS: Injection of AAV-VEGF increased the vascular density in the brain of 3-, 12-, and 24-month-old mice by 22%+/-7% (AAV-VEGF: 320+/-15 per 10x field versus AAV-LacZ: 263+/-8, P<0.05), 20%+/-8 (AAV-VEGF: 300+/-9 versus AAV-LacZ: 250+/-11, P<0.05), and 7%+/-16% (AAV-VEGF: 257+/-27 versus AAV-LacZ: 236+/-13, P=0.283), respectively. There were more VEGF receptor-positive neuroprogenitors in the subventricular zone of AAV-VEGF-injected 3- (22+/-2) and 12-month-old mice (21+/-5) than that of 24-month-old mice (7+/-1). More 5-bromodeoxyuridine-positive endothelial cells and neuroprogenitors were detected around the injection site and subventricular zone of 3- (13+/-4) and 12-month-old mice (14+/-5) than that of 24-month-old mice (1+/-1). VEGF receptor 2 was upregulated in AAV-VEGF-injected brains of 3- and 12-month-old mice, but not in 24-month-old mice. CONCLUSIONS: The angiogenic and neurogenic response to VEGF stimulation is attenuated in the aged mouse brain, which may be due to reduced VEGF receptor activity.

Restoring transcription factor HoxA5 expression inhibits the growth of experimental hemangiomas in the brain.

Hemangiomas are angiogenesis-dependent benign vascular tumors that can rupture and cause intracranial hemorrhages. We previously showed that the transcription factor homeobox A5 (HoxA5), which is absent in activated angiogenic endothelial cells can block angiogenesis. Here, we investigated whether restoring expression of HoxA5 blocks hemangioma growth by transplanting mouse hemangioendothelioma endothelial cells (EOMA) or HoxA5-expressing EOMA cells into the brains of mice. The EOMA cells induced brain hemangiomas characterized by large cystlike spaces lined by thin walls of endothelial cells surrounded by scant smooth muscle cells. When HoxA5-expressing EOMA cells were injected, lesion volumes were reduced between 5- and 20-fold compared with the EOMA control group (p < 0.05). Restoration of HoxA5 was associated with increased thrombospondin-2, which inhibits angiogenesis and reduced hypoxia-inducible factor 1alpha expression. These data suggest that restoring HoxA5 can attenuate experimental brain hemangioma development.

Nonischemic cerebral venous hypertension promotes a pro-angiogenic stage through HIF-1 downstream genes and leukocyte-derived MMP-9.

Cerebral venous hypertension (VH) and angiogenesis are implicated in the pathogenesis of brain arteriovenous malformation and dural arteriovenous fistulae. We studied the association of VH and angiogenesis using a mouse brain VH model. Sixty mice underwent external jugular vein and common carotid artery (CCA) anastomosis (VH model), CCA ligation, or sham dissection (n=20). Hypoxia-inducible factor-1alpha (HIF-1alpha), vascular endothelial growth factor (VEGF) and stromal-cell-derived factor-1alpha (SDF-1alpha) expression, and matrix metalloproteinase (MMP) activity were analyzed. We found VH animals had higher (P<0.05) sagittal sinus pressure (8+/-1 mm Hg) than control groups (1+/-1 mm Hg). Surface cerebral blood flow and mean arterial pressure did not change. Hypoxia-inducible factor-1alpha, VEGF, and SDF-1alpha expression increased (P<0.05). Neutrophils and MMP-9 activity increased 10-fold 1 day after surgery, gradually decreased afterward, and returned to baseline 2 weeks after surgery. Macrophages began to increase 3 days after surgery (P<0.05), which coincided with the changes in SDF-1alpha expression. Capillary density in the parasagittal cortex increased 17% compared with the controls. Our findings suggest that mild nonischemic VH results in a pro-angiogenic stage in the brain by upregulating HIF-1 and its downstream targets, VEGF and SDF-1alpha, increasing leukocyte infiltration and MMP-9 activity.

Gene therapy in cerebrovascular diseases.

Several studies utilize gene therapy technology to study the molecular mechanisms and therapeutic approaches to a variety of neurological diseases. Although gene therapy for neuro-oncology and neurodegenerative diseases has advanced to the clinical arena, those of cerebrovascular nature have remained in the experimental stage and demonstrate promising results. Before gene therapy in cerebrovascular disorders can be successful, various obstacles need to be addressed. In this review, we focus on the fundamentals of gene therapy adapted to cerebrovascular diseases, which include cerebral ischemia, post-subarachnoid hemorrhagic vasospasm, arteriovenous malformations, etc. We discuss: 1) the utilization of gene delivery vehicle, through viral, non-viral, or cellular vectors; 2) the routes and specific delivery of the vector to its target cells or tissue in the central nervous system; 3) the studies that have utilized in vivo and ex vivo gene delivery approaches and their success in the attenuation of cerebrovascular injuries; and finally 4) the future of gene therapy in this field.