Stanford Neurosurgical
Artificial Intelligence and Machine Learning Laboratory
The safety and efficacy of surgery is our number one priority. As we enter the era of big data, the onus is upon us to utilize what we have learned to improve medical care for future generations. Using novel, cutting edge artificial intelligence and machine learning techniques, our goal is to mine through millions of patient records to predict outcomes following all types of surgery. Imagine a virtual algorithm, that is capable of providing a true estimate of surgical risk, outcome, and efficacy – based specifically on your characteristics. This is the very foundation of precision medicine, and will empower physicians to deliver outstanding care.
Research
The focus of my laboratory is to utilize precision medicine techniques to improve the diagnosis and treatment of neurologic conditions. From traumatic brain injury to spinal scoliosis, the ability to capture detailed data regarding clinical symptoms and treatment outcomes has empowered us to do better for patients. Utilize data to do better for patients, that’s what we do.
Publications
-
Telehealth in Neurosurgery: 2021 Council of State Neurosurgical Societies National Survey Results.
World neurosurgery
2022
More
Abstract
OBJECTIVE: Telehealth was rapidly adopted during the COVID-19 pandemic. A survey was distributed to neurosurgeons in the United States (US) to understand its use within neurosurgery, what barriers exist, unique issues related to neurosurgery, and opportunities for improvement.METHODS: A survey was distributed via email and used the SurveyMonkey platform. The survey was sent to 3,828 practicing neurosurgeons within the US 404 responses were collected between Oct. 30, 2021, through Dec. 4, 2021.RESULTS: During the pandemic, telehealth was used multiple times per week by 60.65% and used daily by an additional 12.78% of respondents. A supermajority (89.84%) of respondents felt that evaluating patients across state lines with telemedicine is beneficial. Most respondents (95.81%) believed that telehealth improves patient access to care. The major criticism of telehealth was the inability to perform a neurological exam.CONCLUSIONS: Telehealth has been widely implemented within the field of neurosurgery during the COVID-19 pandemic and has increased access to care. It has allowed patients to be evaluated remotely, including across state lines. While certain aspects of the neurological exam are suited for video evaluation, sensation and reflexes cannot be adequately assessed. Neurosurgeons believe that telehealth adds value to their ability to deliver care.
View details for DOI 10.1016/j.wneu.2022.09.126
View details for PubMedID 36202339
-
Comparison of Deep Learning and Classical Machine Learning Algorithms to Predict Post-operative Outcomes for Anterior Cervical Discectomy and Fusion Procedures with State-of-the-art Performance.
Spine
2022
More
Abstract
STUDY DESIGN: Retrospective cohort.OBJECTIVE: Due to Anterior cervical discectomy and fusion (ACDF) popularity, it is important to predict post-operative complications, unfavorable 90-day readmissions, and 2-year re-operations to improve surgical decision making, prognostication and planning.SUMMARY OF BACKGROUND DATA: Machine learning has been applied to predict post-operative complications for ACDF; however, studies were limited by sample size and model type. These studies achieved 0.70 AUC. Further approaches, not limited to ACDF, focused on specific complication types, and resulted in AUC between 0.70-0.76.METHODS: The IBM MarketScan Commercial Claims and Encounters Database and Medicare Supplement were queried from 2007-2016 to identify adult patients who underwent an ACDF procedure (N=176,816). Traditional machine learning algorithms, logistic regression, support vector machines, were compared with deep neural networks to predict: 90-day post-operative complications, 90-day readmission, and 2-year reoperation. We further generated random deep learning model architectures and trained them on the 90-day complication task to approximate an upper bound. Lastly, using deep learning, we investigated the importance of each input variable for the prediction of 90-day post-operative complications in ACDF.RESULTS: For the prediction of 90-day complication, 90-day readmission, and 2-year reoperation, the deep neural network-based models achieved area under the curve (AUC) of 0.832, 0.713, and 0.671. Logistic regression achieved AUCs of 0.820, 0.712, and 0.671. SVM approaches were significantly lower. The upper bound of deep learning performance was approximated as 0.832. Myelopathy, age, HIV, previous myocardial infarctions, obesity, and documentary weakness were found to be the strongest variable to predict 90-day post-operative complications.CONCLUSIONS: The deep neural network may be used to predict complications for clinical applications after multi-center validation. The results suggest limited added knowledge exists in interactions between the input variables used for this task. Future work should identify novel variables to increase predictive power.
View details for DOI 10.1097/BRS.0000000000004481
View details for PubMedID 36149852
-
Utilization Trends, Cost, and Payments for Adult Spinal Deformity Surgery in Commercial and Medicare-Insured Populations.
Neurosurgery
2022
More
Abstract
BACKGROUND: Previous studies have characterized utilization rates and cost of adult spinal deformity (ASD) surgery, but the differences between these factors in commercially insured and Medicare populations are not well studied.OBJECTIVE: To identify predictors of increased payments for ASD surgery in commercially insured and Medicare populations.METHODS: We identified adult patients who underwent fusion for ASD, 2007 to 2015, in 20% Medicare inpatient file (n = 21614) and MarketScan commercial insurance database (n = 38789). Patient age, sex, race, insurance type, geographical region, Charlson Comorbidity Index, and length of stay were collected. Outcomes included predictors of increased payments, surgical utilization rates, total cost (calculated using Medicare charges and hospital-specific charge-to-cost ratios), and total Medicare and commercial payments for ASD.RESULTS: Rates of fusion increased from 9.0 to 8.4 per 10000 in 2007 to 20.7 and 18.2 per 10000 in 2015 in commercial and Medicare populations, respectively. The Medicare median total charges increased from
-
Chronic opioid use prior to ACDF surgery is associated with inferior post-operative outcomes: a propensity-matched study of 17,443 chronic opioid users.
World neurosurgery
2022
More
Abstract
STUDY DESIGN: Retrospective cohort OBJECTIVE: Candidates for anterior cervical discectomy and fusion (ACDF) have a higher rate of opioid use than does the public, but studies on pre-operative opioid use have not been conducted. We aimed to understand how pre-operative opioid use affects post-ACDF outcomes.METHODS: The MarketScan Database was queried from 2007-2015 to identify adult patients who underwent an ACDF. Patients were classified into separate cohorts based on the number of separate opioid prescriptions in the year before their ACDF. 90-day post-operative complications, post-operative readmission, re-operation, and total inpatient costs were compared between opioid strata. Propensity-score matching (PSM) matched patient cohorts across observed comorbidities.RESULTS: Of 81,671 ACDF patients, 31,312 (38.3%) were non-users, 30,302 (37.1%) were mild users, and 20,057 (24.6%) were chronic users. Chronic opioid users had a higher comorbidity burden, on average, than patients with less frequent opioid use (p<0.001). Chronic opioid users had higher rates of post-operative complications (9.1%) than mild opioid users (6.0%) and non-users (5.3%) (p<0.001), and higher rates of readmission and reoperation. After balancing opioid non-users vs. chronic opioid users along demographic, pre-operative comorbidity, and operative characteristics, post-operative complications remained elevated for chronic opioid users relative to opioid non-users (8.6% vs. 5.7%; p<0.001), as did rates of readmission and reoperation.CONCLUSIONS: Chronic opioid users had more comorbidities than opioid non-users and mild opioid users, longer hospitalizations, and higher rates of post-operative complication, readmission, and reoperation. After balancing patients across covariates, the outcome differences persisted, suggesting a durable association between pre-operative opioid use and negative post-operative outcomes.
View details for DOI 10.1016/j.wneu.2022.07.002
View details for PubMedID 35809840
-
Health Care Resource Utilization in Management of Opioid-Naive Patients With Newly Diagnosed Neck Pain.
JAMA network open
2022; 5 (7): e2222062
More
Abstract
Importance: Research has uncovered heterogeneity and inefficiencies in the management of idiopathic low back pain, but few studies have examined longitudinal care patterns following newly diagnosed neck pain.Objective: To understand health care utilization in patients with new-onset idiopathic neck pain.Design, Setting, and Participants: This cross-sectional study used nationally sourced longitudinal data from the IBM Watson Health MarketScan claims database (2007-2016). Participants included adult patients with newly diagnosed neck pain, no recent opioid use, and at least 1 year of continuous postdiagnosis follow-up. Exclusion criteria included prior or concomitant diagnosis of traumatic cervical disc dislocation, vertebral fractures, myelopathy, and/or cancer. Only patients with at least 1 year of prediagnosis lookback were included. Data analysis was performed from January 2021 to January 2022.Main Outcomes and Measures: The primary outcome of interest was 1-year postdiagnosis health care expenditures, including costs, opioid use, and health care service utilization. Early services were those received within 30 days of diagnosis. Multivariable regression models and regression-adjusted statistics were used.Results: In total, 679 030 patients (310 665 men [45.6%]) met the inclusion criteria, of whom 7858 (1.2%) underwent surgery within 1 year of diagnosis. The mean (SD) age was 44.62 (14.87) years among nonsurgical patients and 49.69 (9.53) years among surgical patients. Adjusting for demographics and comorbidities, 1-year regression-adjusted health care costs were
Our Team
The Stanford Neurosurgical Ariticial Intelligence and Machine Learning Laboratory is led by Dr. Anand Veeravagu, an Assistant Professor of Neurosurgery and Assistant Professor of Orthopedic Surgery, by courtesy, and Director of Minimally Invasive NeuroSpine Surgery at Stanford. Our laboratory team includes neurosurgery residents, clinical instructors, and medical students.
We're Hiring!
Stanford Neurosurgical Artificial Intelligence and Machine Learning Laboratory is inviting applications for a 1-year post-doctoral research position. It offers an excellent opportunity for academic advancement and exposure to clinical neurosurgery. Responsibilities include clinical research productivity, database management, analytics, writing and study coordination. Highly motivated individuals with a medical degree, background analytics and prior neurosurgery experience welcomed.
To apply, please contact us by email: neurobigdata@stanford.edu
Global Neurosurgery
Each year, members of our team travel to Uganda to help develop sustainable neurosurgical care for their country. We train neurosurgeons, residents, and care for patients with complex spinal disorders in order to facilitate the transfer of knowledge. We have cared for children and adults, many of whom have no other option for treatment. Though we complete many surgeries on each mission trip, our focus is to develop neurosurgical capabilities in each of the countries we visit, such that we leave behind a sustainable network of care.