Browse

You are looking at 131 - 140 of 7,858 items for

  • User-accessible content x
Clear All
Free access

Andrew T. Hale, David P. Stonko, Amber Brown, Jaims Lim, David J. Voce, Stephen R. Gannon, Truc M. Le and Chevis N. Shannon

OBJECTIVE

Modern surgical planning and prognostication requires the most accurate outcomes data to practice evidence-based medicine. For clinicians treating children following traumatic brain injury (TBI) these data are severely lacking. The first aim of this study was to assess published CT classification systems in the authors’ pediatric cohort. A pediatric-specific machine-learning algorithm called an artificial neural network (ANN) was then created that robustly outperformed traditional CT classification systems in predicting TBI outcomes in children.

METHODS

The clinical records of children under the age of 18 who suffered a TBI and underwent head CT within 24 hours after TBI (n = 565) were retrospectively reviewed.

RESULTS

“Favorable” outcome (alive with Glasgow Outcome Scale [GOS] score ≥ 4 at 6 months postinjury, n = 533) and “unfavorable” outcome (death at 6 months or GOS score ≤ 3 at 6 months postinjury, n = 32) were used as the primary outcomes. The area under the receiver operating characteristic (ROC) curve (AUC) was used to delineate the strength of each CT grading system in predicting survival (Helsinki, 0.814; Rotterdam, 0.838; and Marshall, 0.781). The AUC for CT score in predicting GOS score ≤ 3, a measure of overall functionality, was similarly predictive (Helsinki, 0.717; Rotterdam, 0.748; and Marshall, 0.663). An ANN was then constructed that was able to predict 6-month outcomes with profound accuracy (AUC = 0.9462 ± 0.0422).

CONCLUSIONS

This study showed that machine-learning can be leveraged to more accurately predict TBI outcomes in children.

Free access

Anthony V. Nguyen, Elizabeth E. Blears, Evan Ross, Rishi R. Lall and Juan Ortega-Barnett

OBJECTIVE

Glioblastoma (GBM) and primary central nervous system lymphoma (PCNSL) are common intracranial pathologies encountered by neurosurgeons. They often may have similar radiological findings, making diagnosis difficult without surgical biopsy; however, management is quite different between these two entities. Recently, predictive analytics, including machine learning (ML), have garnered attention for their potential to aid in the diagnostic assessment of a variety of pathologies. Several ML algorithms have recently been designed to differentiate GBM from PCNSL radiologically with a high sensitivity and specificity. The objective of this systematic review and meta-analysis was to evaluate the implementation of ML algorithms in differentiating GBM and PCNSL.

METHODS

The authors performed a systematic review of the literature using PubMed in accordance with PRISMA guidelines to select and evaluate studies that included themes of ML and brain tumors. These studies were further narrowed down to focus on works published between January 2008 and May 2018 addressing the use of ML in training models to distinguish between GBM and PCNSL on radiological imaging. Outcomes assessed were test characteristics such as accuracy, sensitivity, specificity, and area under the receiver operating characteristic curve (AUC).

RESULTS

Eight studies were identified addressing use of ML in training classifiers to distinguish between GBM and PCNSL on radiological imaging. ML performed well with the lowest reported AUC being 0.878. In studies in which ML was directly compared with radiologists, ML performed better than or as well as the radiologists. However, when ML was applied to an external data set, it performed more poorly.

CONCLUSIONS

Few studies have applied ML to solve the problem of differentiating GBM from PCNSL using imaging alone. Of the currently published studies, ML algorithms have demonstrated promising results and certainly have the potential to aid radiologists with difficult cases, which could expedite the neurosurgical decision-making process. It is likely that ML algorithms will help to optimize neurosurgical patient outcomes as well as the cost-effectiveness of neurosurgical care if the problem of overfitting can be overcome.

Free access

Todd C. Hollon, Adish Parikh, Balaji Pandian, Jamaal Tarpeh, Daniel A. Orringer, Ariel L. Barkan, Erin L. McKean and Stephen E. Sullivan

OBJECTIVE

Pituitary adenomas occur in a heterogeneous patient population with diverse perioperative risk factors, endocrinopathies, and other tumor-related comorbidities. This heterogeneity makes predicting postoperative outcomes challenging when using traditional scoring systems. Modern machine learning algorithms can automatically identify the most predictive risk factors and learn complex risk-factor interactions using training data to build a robust predictive model that can generalize to new patient cohorts. The authors sought to build a predictive model using supervised machine learning to accurately predict early outcomes of pituitary adenoma surgery.

METHODS

A retrospective cohort of 400 consecutive pituitary adenoma patients was used. Patient variables/predictive features were limited to common patient characteristics to improve model implementation. Univariate and multivariate odds ratio analysis was performed to identify individual risk factors for common postoperative complications and to compare risk factors with model predictors. The study population was split into 300 training/validation patients and 100 testing patients to train and evaluate four machine learning models using binary classification accuracy for predicting early outcomes.

RESULTS

The study included a total of 400 patients. The mean ± SD patient age was 53.9 ± 16.3 years, 59.8% of patients had nonfunctioning adenomas and 84.7% had macroadenomas, and the mean body mass index (BMI) was 32.6 ± 7.8 (58.0% obesity rate). Multivariate odds ratio analysis demonstrated that age < 40 years was associated with a 2.86 greater odds of postoperative diabetes insipidus and that nonobese patients (BMI < 30) were 2.2 times more likely to develop postoperative hyponatremia. Using broad criteria for a poor early postoperative outcome—major medical and early surgical complications, extended length of stay, emergency department admission, inpatient readmission, and death—31.0% of patients met criteria for a poor early outcome. After model training, a logistic regression model with elastic net (LR-EN) regularization best predicted early postoperative outcomes of pituitary adenoma surgery on the 100-patient testing set—sensitivity 68.0%, specificity 93.3%, overall accuracy 87.0%. The receiver operating characteristic and precision-recall curves for the LR-EN model had areas under the curve of 82.7 and 69.5, respectively. The most important predictive variables were lowest perioperative sodium, age, BMI, highest perioperative sodium, and Cushing’s disease.

CONCLUSIONS

Early postoperative outcomes of pituitary adenoma surgery can be predicted with 87% accuracy using a machine learning approach. These results provide insight into how predictive modeling using machine learning can be used to improve the perioperative management of pituitary adenoma patients.

Free access

Aiko Terada, Masaki Komiyama, Tomoya Ishiguro, Yasunari Niimi and Hidenori Oishi

The authors performed a nationwide study in Japan to evaluate the annual detected rate of pediatric intracranial arteriovenous (AV) shunts such as brain AV malformations (BAVMs), pial AV fistulas (PAVFs), vein of Galen aneurysmal malformations (VGAMs), and dural AV fistulas (DAVFs). These rates were revealed for the first time and showed that VGAM, DAVF, and PAVF were relatively common but that BAVMs were extremely rare in neonates and infants.

Free access

Nikhil Paliwal, Prakhar Jaiswal, Vincent M. Tutino, Hussain Shallwani, Jason M. Davies, Adnan H. Siddiqui, Rahul Rai and Hui Meng

OBJECTIVE

Flow diverters (FDs) are designed to occlude intracranial aneurysms (IAs) while preserving flow to essential arteries. Incomplete occlusion exposes patients to risks of thromboembolic complications and rupture. A priori assessment of FD treatment outcome could enable treatment optimization leading to better outcomes. To that end, the authors applied image-based computational analysis to clinically FD-treated aneurysms to extract information regarding morphology, pre- and post-treatment hemodynamics, and FD-device characteristics and then used these parameters to train machine learning algorithms to predict 6-month clinical outcomes after FD treatment.

METHODS

Data were retrospectively collected for 84 FD-treated sidewall aneurysms in 80 patients. Based on 6-month angiographic outcomes, IAs were classified as occluded (n = 63) or residual (incomplete occlusion, n = 21). For each case, the authors modeled FD deployment using a fast virtual stenting algorithm and hemodynamics using image-based computational fluid dynamics. Sixteen morphological, hemodynamic, and FD-based parameters were calculated for each aneurysm. Aneurysms were randomly assigned to a training or testing cohort in approximately a 3:1 ratio. The Student t-test and Mann-Whitney U-test were performed on data from the training cohort to identify significant parameters distinguishing the occluded from residual groups. Predictive models were trained using 4 types of supervised machine learning algorithms: logistic regression (LR), support vector machine (SVM; linear and Gaussian kernels), K-nearest neighbor, and neural network (NN). In the testing cohort, the authors compared outcome prediction by each model trained using all parameters versus only the significant parameters.

RESULTS

The training cohort (n = 64) consisted of 48 occluded and 16 residual aneurysms and the testing cohort (n = 20) consisted of 15 occluded and 5 residual aneurysms. Significance tests yielded 2 morphological (ostium ratio and neck ratio) and 3 hemodynamic (pre-treatment inflow rate, post-treatment inflow rate, and post-treatment aneurysm averaged velocity) discriminants between the occluded (good-outcome) and the residual (bad-outcome) group. In both training and testing, all the models trained using all 16 parameters performed better than all the models trained using only the 5 significant parameters. Among the all-parameter models, NN (AUC = 0.967) performed the best during training, followed by LR and linear SVM (AUC = 0.941 and 0.914, respectively). During testing, NN and Gaussian-SVM models had the highest accuracy (90%) in predicting occlusion outcome.

CONCLUSIONS

NN and Gaussian-SVM models incorporating all 16 morphological, hemodynamic, and FD-related parameters predicted 6-month occlusion outcome of FD treatment with 90% accuracy. More robust models using the computational workflow and machine learning could be trained on larger patient databases toward clinical use in patient-specific treatment planning and optimization.

Free access

Clinton J. Devin, Mohamad Bydon, Mohammed Ali Alvi, Panagiotis Kerezoudis, Inamullah Khan, Ahilan Sivaganesan, Matthew J. McGirt, Kristin R. Archer, Kevin T. Foley, Praveen V. Mummaneni, Erica F. Bisson, John J. Knightly, Christopher I. Shaffrey and Anthony L. Asher

OBJECTIVE

Back pain and neck pain are two of the most common causes of work loss due to disability, which poses an economic burden on society. Due to recent changes in healthcare policies, patient-centered outcomes including return to work have been increasingly prioritized by physicians and hospitals to optimize healthcare delivery. In this study, the authors used a national spine registry to identify clinical factors associated with return to work at 3 months among patients undergoing a cervical spine surgery.

METHODS

The authors queried the Quality Outcomes Database registry for information collected from April 2013 through March 2017 for preoperatively employed patients undergoing cervical spine surgery for degenerative spine disease. Covariates included demographic, clinical, and operative variables, and baseline patient-reported outcomes. Multiple imputations were used for missing values and multivariable logistic regression analysis was used to identify factors associated with higher odds of returning to work. Bootstrap resampling (200 iterations) was used to assess the validity of the model. A nomogram was constructed using the results of the multivariable model.

RESULTS

A total of 4689 patients were analyzed, of whom 82.2% (n = 3854) returned to work at 3 months postoperatively. Among previously employed and working patients, 89.3% (n = 3443) returned to work compared to 52.3% (n = 411) among those who were employed but not working (e.g., were on a leave) at the time of surgery (p < 0.001). On multivariable logistic regression the authors found that patients who were less likely to return to work were older (age > 56–65 years: OR 0.69, 95% CI 0.57–0.85, p < 0.001; age > 65 years: OR 0.65, 95% CI 0.43–0.97, p = 0.02); were employed but not working (OR 0.24, 95% CI 0.20–0.29, p < 0.001); were employed part time (OR 0.56, 95% CI 0.42–0.76, p < 0.001); had a heavy-intensity (OR 0.42, 95% CI 0.32–0.54, p < 0.001) or medium-intensity (OR 0.59, 95% CI 0.46–0.76, p < 0.001) occupation compared to a sedentary occupation type; had workers’ compensation (OR 0.38, 95% CI 0.28–0.53, p < 0.001); had a higher Neck Disability Index score at baseline (OR 0.60, 95% CI 0.51–0.70, p = 0.017); were more likely to present with myelopathy (OR 0.52, 95% CI 0.42–0.63, p < 0.001); and had more levels fused (3–5 levels: OR 0.46, 95% CI 0.35–0.61, p < 0.001). Using the multivariable analysis, the authors then constructed a nomogram to predict return to work, which was found to have an area under the curve of 0.812 and good validity.

CONCLUSIONS

Return to work is a crucial outcome that is being increasingly prioritized for employed patients undergoing spine surgery. The results from this study could help surgeons identify at-risk patients so that preoperative expectations could be discussed more comprehensively.

Free access

Krishna Chaitanya Joshi, Ignacio Larrabide, Ahmed Saied, Nada Elsaid, Hector Fernandez and Demetrius K. Lopes

OBJECTIVE

The authors sought to validate the use of a software-based simulation for preassessment of braided self-expanding stents in the treatment of wide-necked intracranial aneurysms.

METHODS

This was a retrospective, observational, single-center study of 13 unruptured and ruptured intracranial aneurysms treated with braided self-expanding stents. Pre- and postprocedural angiographic studies were analyzed. ANKYRAS software was used to compare the following 3 variables: the manufacturer-given nominal length (NL), software-calculated simulated length (SL), and the actual measured length (ML) of the stent. Appropriate statistical methods were used to draw correlations among the 3 lengths.

RESULTS

In this study, data obtained in 13 patients treated with braided self-expanding stents were analyzed. Data for the 3 lengths were collected for all patients. Error discrepancy was calculated by mean squared error (NL to ML −22.2; SL to ML −6.14, p < 0.05), mean absolute error (NL to ML 3.88; SL to ML −1.84, p < 0.05), and mean error (NL to ML −3.81; SL to ML −1.22, p < 0.05).

CONCLUSIONS

The ML was usually less than the NL given by the manufacturer, indicating significant change in length in most cases. Computational software-based simulation for preassessment of the braided self-expanding stents is a safe and effective way for accurately calculating the change in length to aid in choosing the right-sized stent for optimal placement in complex intracranial vasculature.

Free access

Victor E. Staartjes, Carlo Serra, Giovanni Muscas, Nicolai Maldaner, Kevin Akeret, Christiaan H. B. van Niftrik, Jorn Fierstra, David Holzmann and Luca Regli

OBJECTIVE

Gross-total resection (GTR) is often the primary surgical goal in transsphenoidal surgery for pituitary adenoma. Existing classifications are effective at predicting GTR but are often hampered by limited discriminatory ability in moderate cases and by poor interrater agreement. Deep learning, a subset of machine learning, has recently established itself as highly effective in forecasting medical outcomes. In this pilot study, the authors aimed to evaluate the utility of using deep learning to predict GTR after transsphenoidal surgery for pituitary adenoma.

METHODS

Data from a prospective registry were used. The authors trained a deep neural network to predict GTR from 16 preoperatively available radiological and procedural variables. Class imbalance adjustment, cross-validation, and random dropout were applied to prevent overfitting and ensure robustness of the predictive model. The authors subsequently compared the deep learning model to a conventional logistic regression model and to the Knosp classification as a gold standard.

RESULTS

Overall, 140 patients who underwent endoscopic transsphenoidal surgery were included. GTR was achieved in 95 patients (68%), with a mean extent of resection of 96.8% ± 10.6%. Intraoperative high-field MRI was used in 116 (83%) procedures. The deep learning model achieved excellent area under the curve (AUC; 0.96), accuracy (91%), sensitivity (94%), and specificity (89%). This represents an improvement in comparison with the Knosp classification (AUC: 0.87, accuracy: 81%, sensitivity: 92%, specificity: 70%) and a statistically significant improvement in comparison with logistic regression (AUC: 0.86, accuracy: 82%, sensitivity: 81%, specificity: 83%) (all p < 0.001).

CONCLUSIONS

In this pilot study, the authors demonstrated the utility of applying deep learning to preoperatively predict the likelihood of GTR with excellent performance. Further training and validation in a prospective multicentric cohort will enable the development of an easy-to-use interface for use in clinical practice.

Free access

Nicholas Dietz, Mayur Sharma, Ahmad Alhourani, Beatrice Ugiliweneza, Dengzhi Wang, Miriam A. Nuño, Doniel Drazin and Maxwell Boakye

OBJECTIVE

There is increasing emphasis on patient-reported outcomes (PROs) to quantitatively evaluate quality outcomes from degenerative spine surgery. However, accurate prediction of PROs is challenging due to heterogeneity in outcome measures, patient characteristics, treatment characteristics, and methodological characteristics. The purpose of this study was to evaluate the current landscape of independently validated predictive models for PROs in elective degenerative spinal surgery with respect to study design and model generation, training, accuracy, reliability, variance, and utility.

METHODS

The authors analyzed the current predictive models in PROs by performing a search of the PubMed and Ovid databases using PRISMA guidelines and a PICOS (participants, intervention, comparison, outcomes, study design) model. They assessed the common outcomes and variables used across models as well as the study design and internal validation methods.

RESULTS

A total of 7 articles met the inclusion criteria, including a total of 17 validated predictive models of PROs after adult degenerative spine surgery. National registry databases were used in 4 of the studies. Validation cohorts were used in 2 studies for model verification and 5 studies used other methods, including random sample bootstrapping techniques. Reported c-index values ranged from 0.47 to 0.79. Two studies report the area under the curve (0.71–0.83) and one reports a misclassification rate (9.9%). Several positive predictors, including high baseline pain intensity and disability, demonstrated high likelihood of favorable PROs.

CONCLUSIONS

A limited but effective cohort of validated predictive models of spine surgical outcomes had proven good predictability for PROs. Instruments with predictive accuracy can enhance shared decision-making, improve rehabilitation, and inform best practices in the setting of heterogeneous patient characteristics and surgical factors.

Free access

Katie L. Krause, James T. Obayashi, Kelly J. Bridges, Ahmed M. Raslan and Khoi D. Than

OBJECTIVE

Common interbody graft options for anterior cervical discectomy and fusion (ACDF) include structural allograft and polyetheretherketone (PEEK). PEEK has gained popularity due to its radiolucency and its elastic modulus, which is similar to that of bone. The authors sought to compare the rates of pseudarthrosis, a lack of solid bone growth across the disc space, and the need for revision surgery with the use of grafts made of allogenic bone versus PEEK.

METHODS

The authors retrospectively reviewed 127 cases in which patients had undergone a 1-level ACDF followed by at least 1 year of radiographic follow-up. Data on age, sex, body mass index, tobacco use, pseudarthrosis, and the reoperation rate for pseudarthrosis were collected. These data were analyzed by performing a Pearson’s chi-square test.

RESULTS

Of 127 patients, 56 had received PEEK implants and 71 had received allografts. Forty-six of the PEEK implants (82%) were stand-alone devices. There were no significant differences between the 2 treatment groups with respect to patient age, sex, or body mass index. Twenty-nine (52%) of 56 patients with PEEK implants demonstrated radiographic evidence of pseudarthrosis, compared to 7 (10%) of 71 patients with structural allografts (p < 0.001, OR 9.82; 95% CI 3.836–25.139). Seven patients with PEEK implants required reoperation for pseudarthrosis, compared to 1 patient with an allograft (p = 0.01, OR 10.00; 95% CI 1.192–83.884). There was no significant difference in tobacco use between the PEEK and allograft groups (p = 0.586).

CONCLUSIONS

The results of this study demonstrate that the use of PEEK devices in 1-level ACDF is associated with a significantly higher rate of radiographically demonstrated pseudarthrosis and need for revision surgery compared with the use of allografts. Surgeons should be aware of this when deciding on interbody graft options, and reimbursement policies should reflect these discrepancies.