Aditya V. Karhade

Thirty-day readmission and reoperation after surgery for spinal tumors: a National Surgical Quality Improvement Program analysis

OBJECTIVE The goal of this study was to use a large national registry to evaluate the 30-day cumulative incidence and predictors of adverse events, readmissions, and reoperations after surgery for primary and secondary spinal tumors. METHODS Data from adult patients who underwent surgery for spinal tumors (2011-2014) were extracted from the prospective National Surgical Quality Improvement Program (NSQIP) registry. Multivariable logistic regression was used to evaluate predictors of reoperation, readmission, and major complications (death, neurological, cardiopulmonary, venous thromboembolism [VTE], surgical site infection [SSI], and sepsis). Variables screened included patient age, sex, tumor location, American Society of Anesthesiologists (ASA) physical classification, preoperative functional status, comorbidities, preoperative laboratory values, case urgency, and operative time. Additional variables that were evaluated when analyzing readmission included complications during the surgical hospitalization, hospital length of stay (LOS), and discharge disposition. RESULTS Among the 2207 patients evaluated, 51.4% had extradural tumors, 36.4% had intradural extramedullary tumors, and 12.3% had intramedullary tumors. By spinal level, 20.7% were cervical lesions, 47.4% were thoracic lesions, 29.1% were lumbar lesions, and 2.8% were sacral lesions. Readmission occurred in 10.2% of patients at a median of 18 days (interquartile range [IQR] 12-23 days); the most common reasons for readmission were SSIs (23.7%), systemic infections (17.8%), VTE (12.7%), and CNS complications (11.9%). Predictors of readmission were comorbidities (dyspnea, hypertension, and anemia), disseminated cancer, preoperative steroid use, and an extended hospitalization. Reoperation occurred in 5.3% of patients at a median of 13 days (IQR 8-20 days) postoperatively and was associated with preoperative steroid use and ASA Class 4-5 designation. Major complications occurred in 14.4% of patients: the most common complications and their median time to occurrence were VTE (4.5%) at 9 days (IQR 4-19 days) postoperatively, SSIs (3.6%) at 18 days (IQR 14-25 days), and sepsis (2.9%) at 13 days (IQR 7-21 days). Predictors of major complications included dependent functional status, emergency case status, male sex, comorbidities (dyspnea, bleeding disorders, preoperative systemic inflammatory response syndrome, preoperative leukocytosis), and ASA Class 3-5 designation (p < 0.05). The median hospital LOS was 5 days (IQR 3-9 days), the 30-day mortality rate was 3.3%, and the median time to death was 20 days (IQR 12.5-26 days). CONCLUSIONS In this NSQIP analysis, 10.2% of patients undergoing surgery for spinal tumors were readmitted within 30 days, 5.3% underwent a reoperation, and 14.4% experienced a major complication. The most common complications were SSIs, systemic infections, and VTE, which often occurred late (after discharge from the surgical hospitalization). Patients were primarily readmitted for new complications that developed following discharge rather than exacerbation of complications from the surgical hospital stay. The strongest predictors of adverse events were comorbidities, preoperative steroid use, and higher ASA classification. These models can be used by surgeons to risk-stratify patients preoperatively and identify those who may benefit from increased surveillance following hospital discharge.

Agents for fluorescence-guided glioma surgery: a systematic review of preclinical and clinical results

BackgroundFluorescence-guided surgery (FGS) is a technique used to enhance visualization of tumor margins in order to increase the extent of tumor resection in glioma surgery. In this paper, we systematically review all clinically tested fluorescent agents for application in FGS for glioma and all preclinically tested agents with the potential for FGS for glioma.MethodsWe searched the PubMed and Embase databases for all potentially relevant studies through March 2016. We assessed fluorescent agents by the following outcomes: rate of gross total resection (GTR), overall and progression-free survival, sensitivity and specificity in discriminating tumor and healthy brain tissue, tumor-to-normal ratio of fluorescent signal, and incidence of adverse events.ResultsThe search strategy resulted in 2155 articles that were screened by titles and abstracts. After full-text screening, 105 articles fulfilled the inclusion criteria evaluating the following fluorescent agents: 5-aminolevulinic acid (5-ALA) (44 studies, including three randomized control trials), fluorescein (11), indocyanine green (five), hypericin (two), 5-aminofluorescein-human serum albumin (one), endogenous fluorophores (nine) and fluorescent agents in a pre-clinical testing phase (30). Three meta-analyses were also identified.Conclusions5-ALA is the only fluorescent agent that has been tested in a randomized controlled trial and results in an improvement of GTR and progression-free survival in high-grade gliomas. Observational cohort studies and case series suggest similar outcomes for FGS using fluorescein. Molecular targeting agents (e.g., fluorophore/nanoparticle labeled with anti-EGFR antibodies) are still in the pre-clinical phase, but offer promising results and may be valuable future alternatives.

Natural and Artificial Intelligence in Neurosurgery: A Systematic Review

BACKGROUND Machine learning (ML) is a domain of artificial intelligence that allows computer algorithms to learn from experience without being explicitly programmed. OBJECTIVE To summarize neurosurgical applications of ML where it has been compared to clinical expertise, here referred to as “natural intelligence.” METHODS A systematic search was performed in the PubMed and Embase databases as of August 2016 to review all studies comparing the performance of various ML approaches with that of clinical experts in neurosurgical literature. RESULTS Twenty‐three studies were identified that used ML algorithms for diagnosis, presurgical planning, or outcome prediction in neurosurgical patients. Compared to clinical experts, ML models demonstrated a median absolute improvement in accuracy and area under the receiver operating curve of 13% (interquartile range 4‐21%) and 0.14 (interquartile range 0.07‐0.21), respectively. In 29 (58%) of the 50 outcome measures for which a P‐value was provided or calculated, ML models outperformed clinical experts (P < .05). In 18 of 50 (36%), no difference was seen between ML and expert performance (P > .05), while in 3 of 50 (6%) clinical experts outperformed ML models (P < .05). All 4 studies that compared clinicians assisted by ML models vs clinicians alone demonstrated a better performance in the first group. CONCLUSION We conclude that ML models have the potential to augment the decision‐making capacity of clinicians in neurosurgical applications; however, significant hurdles remain associated with creating, validating, and deploying ML models in the clinical setting. Shifting from the preconceptions of a human‐vs‐machine to a human‐and‐machine paradigm could be essential to overcome these hurdles.

National Databases for Neurosurgical Outcomes Research: Options, Strengths, and Limitations

BACKGROUND Quality improvement, value-based care delivery, and personalized patient care depend on robust clinical, financial, and demographic data streams of neurosurgical outcomes. The neurosurgical literature lacks a comprehensive review of large national databases. OBJECTIVE To assess the strengths and limitations of various resources for outcomes research in neurosurgery. METHODS A review of the literature was conducted to identify surgical outcomes studies using national data sets. The databases were assessed for the availability of patient demographics and clinical variables, longitudinal follow-up of patients, strengths, and limitations. RESULTS The number of unique patients contained within each data set ranged from thousands (Quality Outcomes Database [QOD]) to hundreds of millions (MarketScan). Databases with both clinical and financial data included PearlDiver, Premier Healthcare Database, Vizient Clinical Data Base and Resource Manager, and the National Inpatient Sample. Outcomes collected by databases included patient-reported outcomes (QOD); 30-day morbidity, readmissions, and reoperations (National Surgical Quality Improvement Program); and disease incidence and disease-specific survival (Surveillance, Epidemiology, and End Results-Medicare). The strengths of large databases included large numbers of rare pathologies and multi-institutional nationally representative sampling; the limitations of these databases included variable data veracity, variable data completeness, and missing disease-specific variables. CONCLUSION The improvement of existing large national databases and the establishment of new registries will be crucial to the future of neurosurgical outcomes research.

An introduction and overview of machine learning in neurosurgical care

BackgroundMachine learning (ML) is a branch of artificial intelligence that allows computers to learn from large complex datasets without being explicitly programmed. Although ML is already widely manifest in our daily lives in various forms, the considerable potential of ML has yet to find its way into mainstream medical research and day-to-day clinical care. The complex diagnostic and therapeutic modalities used in neurosurgery provide a vast amount of data that is ideally suited for ML models. This systematic review explores ML’s potential to assist and improve neurosurgical care.MethodA systematic literature search was performed in the PubMed and Embase databases to identify all potentially relevant studies up to January 1, 2017. All studies were included that evaluated ML models assisting neurosurgical treatment.ResultsOf the 6,402 citations identified, 221 studies were selected after subsequent title/abstract and full-text screening. In these studies, ML was used to assist surgical treatment of patients with epilepsy, brain tumors, spinal lesions, neurovascular pathology, Parkinson’s disease, traumatic brain injury, and hydrocephalus. Across multiple paradigms, ML was found to be a valuable tool for presurgical planning, intraoperative guidance, neurophysiological monitoring, and neurosurgical outcome prediction.ConclusionsML has started to find applications aimed at improving neurosurgical care by increasing the efficiency and precision of perioperative decision-making. A thorough validation of specific ML models is essential before implementation in clinical neurosurgical care. To bridge the gap between research and clinical care, practical and ethical issues should be considered parallel to the development of these techniques.

United States neurosurgery annual case type and complication trends between 2006 and 2013: An American College of Surgeons National Surgical Quality Improvement Program analysis

We aimed to identify trends in the neurosurgical practice environment in the United States from 2006 to 2013 using the American College of Surgeons-National Surgical Quality Improvement Program (NSQIP) database, and to determine the complication rate for spinal and cranial procedures and identify risk factors for post-operative complications across this time period. We performed a search of the American College of Surgeons-NSQIP database for all patients undergoing an operation with a surgeon whose primary specialty was neurological surgery from 2006 to 2013. Analysis of patient demographics and pre-operative co-morbidities was performed, and multivariate analysis was used to determine predictors of surgical complications. From 2006 to 2013, the percentage of spinal operations performed by neurosurgeons relative to cranial and peripheral nerve cases increased from 68.0% to 76.8% (p<0.001) according to the NSQIP database. The proportion of cranial cases during the same time period decreased from 29.7% to 21.6% (p<0.001). The overall 30-day complication rate among all 94,621 NSQIP reported patients undergoing operations with a neurosurgeon over this time period was 8.2% (5.6% for spinal operations, 16.1% for cranial operations). The overall rate decreased from 11.0% in 2006 to 7.5% in 2013 (p<0.001). Several predictors of post-operative complication were identified on multivariate analysis.

Development of Machine Learning Algorithms for Prediction of 5-Year Spinal Chordoma Survival

BACKGROUND Chordomas are locally invasive slow-growing tumors that are difficult to study because of the rarity of the tumors and the lack of significant volumes of patients with longitudinal follow-up. As such, there are currently no machine learning studies in the chordoma literature. The purpose of this study was to develop machine learning models for survival prediction and deploy them as open access web applications as a proof of concept for machine learning in rare nervous system lesions. METHODS The National Cancer Institutes Surveillance, Epidemiology, and End Results program database was used to identify adult patients diagnosed with spinal chordoma between 1995 and 2010. Four machine learning models were used to predict 5-year survival for spinal chordoma and assessed by discrimination, calibration, and overall performance. RESULTS The 5-year overall survival for 265 patients with spinal chordoma was 67.5%. Variables used for prediction were age at diagnosis, tumor size, tumor location, extent of tumor invasion, and extent of surgery. For 5-year survival prediction, the Bayes Point Machine achieved the best performance with a c statistic of 0.80, calibration slope of 1.01, calibration intercept of 0.03, and Brier score of 0.16. This model for 5-year mortality prediction was incorporated into an open access application and can be found online (https://sorg-apps.shinyapps.io/chordoma/). CONCLUSIONS This analysis of patients with spinal chordoma demonstrated that machine learning models can be developed for survival prediction in rare pathologies and have the potential to serve as the basis for creation of decision support tools in the future.

Improving Outcomes: Big Data and Predictive Analytics

Abstract The Computational Neurosurgery Outcomes Center at Brigham and Womens Hospital focuses on new knowledge creation and value-based innovation with big data, predictive analytics, patient-reported outcomes, and national databases. Work with artificial intelligence, including machine learning for structured data such as online medical records and natural language processing for unstructured data such as clinic notes, allows for analysis of high-velocity, high-volume, and high-complexity neurosurgical data. Use of mobile-based actively collected patient reported outcomes has been paralleled by the development of digital phenotyping and passive data collection. The study of surgical outcomes has evolved tremendously since the work of Ernest Amory Codman and big data, predictive analytics, and passively collected patient reported outcomes are the next frontier for outcomes research.

Use of Intraoperative Ultrasound During Spinal Surgery

Study Design: Review and technical report. Objective: Intraoperative ultrasound has been used by spine surgeons since the early 1980s. Since that time, more advanced modes of intraoperative imaging and navigation have become widely available. Although the use of ultrasound during spine surgery has fallen out of favor, it remains the only true real-time imaging modality that allows surgeons to visualize soft tissue anatomy instantly and continuously while operating. It is our objective to demonstrate that for this reason, ultrasound is a useful adjunctive technique for spine surgeons, especially when approaching intradural lesions or when addressing pathology in the ventral spinal canal via a posterior approach. Methods: Using PubMed, the existing literature regarding the use of intraoperative ultrasound during spinal surgery was evaluated. Also, surgical case logs were reviewed to identify spinal operations during which intraoperative ultrasound was used. Illustrative cases were selected and reviewed in detail. Results: This article provides a brief review of the history of intraoperative ultrasound in spine surgery and describes certain surgical scenarios during which this technique might be useful. Several illustrative cases are provided from our own experience. Conclusions: Surgeons should consider the use of intraoperative ultrasound when approaching intradural lesions or when addressing pathology ventral to the thecal sac via a posterior approach.