Faculty Bibliography

Background/UNASSIGNED:Differentiating glioblastoma, brain metastasis, and central nervous system lymphoma (CNSL) on conventional magnetic resonance imaging (MRI) can present a diagnostic dilemma due to the potential for overlapping imaging features. We investigate whether machine learning evaluation of multimodal MRI can reliably differentiate these entities. Methods/UNASSIGNED:Preoperative brain MRI including diffusion weighted imaging (DWI), dynamic contrast enhanced (DCE), and dynamic susceptibility contrast (DSC) perfusion in patients with glioblastoma, lymphoma, or metastasis were retrospectively reviewed. Perfusion maps (rCBV, rCBF), permeability maps (K-trans, Kep, Vp, Ve), ADC, T1C+ and T2/FLAIR images were coregistered and two separate volumes of interest (VOIs) were obtained from the enhancing tumor and non-enhancing T2 hyperintense (NET2) regions. The tumor volumes obtained from these VOIs were utilized for supervised training of support vector classifier (SVC) and multilayer perceptron (MLP) models. Validation of the trained models was performed on unlabeled cases using the leave-one-subject-out method. Head-to-head and multiclass models were created. Accuracies of the multiclass models were compared against two human interpreters reviewing conventional and diffusion-weighted MR images. Results/UNASSIGNED:Twenty-six patients enrolled with histopathologically-proven glioblastoma (n=9), metastasis (n=9), and CNS lymphoma (n=8) were included. The trained multiclass ML models discriminated the three pathologic classes with a maximum accuracy of 69.2% accuracy (18 out of 26; kappa 0.540, P=0.01) using an MLP trained with the VpNET2 tumor volumes. Human readers achieved 65.4% (17 out of 26) and 80.8% (21 out of 26) accuracies, respectively. Using the MLP VpNET2 model as a computer-aided diagnosis (CADx) for cases in which the human reviewers disagreed with each other on the diagnosis resulted in correct diagnoses in 5 (19.2%) additional cases. Conclusions/UNASSIGNED:Our trained multiclass MLP using VpNET2 can differentiate glioblastoma, brain metastasis, and CNS lymphoma with modest diagnostic accuracy and provides approximately 19% increase in diagnostic yield when added to routine human interpretation.

Background/UNASSIGNED:Errors in grammar, spelling, and usage in radiology reports are common. To automatically detect inappropriate insertions, deletions, and substitutions of words in radiology reports, we proposed using a neural sequence-to-sequence (seq2seq) model. Methods/UNASSIGNED:Head CT and chest radiograph reports from Mount Sinai Hospital (MSH) (n=61,722 and 818,978, respectively), Mount Sinai Queens (MSQ) (n=30,145 and 194,309, respectively) and MIMIC-III (n=32,259 and 54,685) were converted into sentences. Insertions, substitutions, and deletions of words were randomly introduced. Seq2seq models were trained using corrupted sentences as input to predict original uncorrupted sentences. Three models were trained using head CTs from MSH, chest radiographs from MSH, and head CTs from all three collections. Model performance was assessed across different sites and modalities. A sample of original, uncorrupted sentences were manually reviewed for any error in syntax, usage, or spelling to estimate real-world proofreading performance of the algorithm. Results/UNASSIGNED:Seq2seq detected 90.3% and 88.2% of corrupted sentences with 97.7% and 98.8% specificity in same-site, same-modality test sets for head CTs and chest radiographs, respectively. Manual review of original, uncorrupted same-site same-modality head CT sentences demonstrated seq2seq positive predictive value (PPV) 0.393 (157/400; 95% CI, 0.346-0.441) and negative predictive value (NPV) 0.986 (789/800; 95% CI, 0.976-0.992) for detecting sentences containing real-world errors, with estimated sensitivity of 0.389 (95% CI, 0.267-0.542) and specificity 0.986 (95% CI, 0.985-0.987) over n=86,211 uncorrupted training examples. Conclusions/UNASSIGNED:Seq2seq models can be highly effective at detecting erroneous insertions, deletions, and substitutions of words in radiology reports. To achieve high performance, these models require site- and modality-specific training examples. Incorporating additional targeted training data could further improve performance in detecting real-world errors in reports.

With 15 drugs currently approved for the treatment of metastatic renal cell carcinoma (mRCC) and even more combination regimens with immunotherapy on the horizon, there remains a distinct lack of molecular biomarkers for therapeutic efficacy. Our study reports on real-world clinical outcomes of mRCC patients from a tertiary academic medical center treated with empirically selected standard-of-care therapy. We utilized the Stanford Renal Cell Carcinoma Database (RCCD) to report on various outcome measures, including overall survival (OS) and the median number of lines of targeted therapies received from the time of metastatic diagnosis. We found that most metastatic patients did not survive long enough to attempt even half of the available targeted therapies. We also noted that patients who failed to receive a clinical benefit within the first two lines of therapy could still go on to experience clinical benefit in later lines of therapy. The term, "clinical benefit" was assigned to a line of therapy if a patient remained on drug treatment for three months or longer. Moreover, patients with clinical benefit in at least one line of therapy experienced significantly longer OS compared to those who did not have clinical benefit in at least one line of therapy. Developing biomarkers that identify patients who will receive clinical benefit in individual lines of therapy is one potential strategy for achieving rational drug sequencing in mRCC.

Deciphering the massive volume of complex electronic data that has been compiled by hospital systems over the past decades has the potential to revolutionize modern medicine, as well as present significant challenges. Deep learning is uniquely suited to address these challenges, and recent advances in techniques and hardware have poised the field of medical machine learning for transformational growth. The clinical neurosciences are particularly well positioned to benefit from these advances given the subtle presentation of symptoms typical of neurologic disease. Here we review the various domains in which deep learning algorithms have already provided impetus for change-areas such as medical image analysis for the improved diagnosis of Alzheimer's disease and the early detection of acute neurologic events; medical image segmentation for quantitative evaluation of neuroanatomy and vasculature; connectome mapping for the diagnosis of Alzheimer's, autism spectrum disorder, and attention deficit hyperactivity disorder; and mining of microscopic electroencephalogram signals and granular genetic signatures. We additionally note important challenges in the integration of deep learning tools in the clinical setting and discuss the barriers to tackling the challenges that currently exist.

BACKGROUND:The incidence of meningioma has increased drastically recently, particularly in older adults. Surgical intervention has the potential to reduce neurologic symptoms and achieve favorable, long-term outcomes. There is considerable variability in the literature examining the relationship between age and outcomes after meningioma surgery. The objective of this study was to identify the relationship between age and postoperative complications after craniotomy for resection of meningioma. METHODS:The American College of Surgeons National Surgical Quality Improvement Program database was used to identify patients undergoing craniotomy for meningioma resection between 2005 and 2012. Multivariate analysis was used to identify associations between age and postoperative complications. RESULTS:Age >80 years is an independent risk factor for any complication (odds ratio [OR], 2.374; 95% confidence interval [CI], 1.3-4.4; P = 0.015), death within 30 days of surgery (OR, 15.7; 95% CI, 3.0-81.0; P < 0.001), and length of stay >5 days (OR, 3.2; 95% CI, 1.8-5.6; P < 0.001). CONCLUSIONS:Advanced age, particularly >80 years, is an independent predictor of morbidity and mortality in patients undergoing craniotomy for resection of meningioma. As such, it should be considered in preoperative optimization and risk stratification.

STUDY DESIGN:Retrospective analysis of prospectively collected data. OBJECTIVE:The aim of this study was to determine the ability of abnormal coagulation profile to predict adverse events following posterior cervical laminectomy and fusion (PCLF). SUMMARY OF BACKGROUND DATA:PCLF is an increasingly common procedure used to treat a variety of traumatic and degenerative spinal conditions. Abnormal coagulation profile is associated with postoperative adverse events, including blood transfusion. There is a paucity of literature that specifically addresses the relationship between coagulation profile and complications following PCLF. METHODS:ACS-NSQIP was utilized to identify patients undergoing PCLF between 2006 and 2013. A total of 3546 patients met inclusion criteria. Multivariate analysis was utilized to identify associations between abnormal coagulation profile and postoperative complications. RESULTS:Membership in the low-platelet cohort was an independent risk factor for myocardial infarction (Odds Ratio (OR) = 5.4 [1.0, 29.1], P = 0.049) and bleeding transfusion (OR = 2.0 [1.2, 3.4], P = 0.011). Membership in the high international normalized ratio group was an independent risk factor for pneumonia (OR = 6.3 [2.5, 16.1], P < 0.001), ventilation >48 hours (OR = 6.5 [2.3, 18.4], P < 0.001), organ space surgical site infection (OR = 11.1 [2.1, 57.3], P = 0.004), urinary tract infection (OR = 3.0 [1.2, 8.0], P = 0.024), bleeding transfusion (OR = 6.0 [3.4, 10.7], P < 0.001), sepsis (OR = 5.1 [1.6, 16.4], P = 0.006), and septic shock (OR = 6.8 [1.7, 27.4], P = 0.007). Membership in the bleeding disorders cohort was an independent predictor of unplanned intubation (OR = 3.2 [1.1, 9.5], P = 0.041), pneumonia (OR = 2.9 [1.2, 7.2], P = 0.023), ventilation >48 hours (OR = 4.8 [1.9, 12.4], P = 0.001), cerebrovascular accident/stroke with neurological deficit (OR = 24.8 [2.9, 210.6], P = 0.003), bleeding transfusion (OR = 2.1 [1.1, 4.1], P = 0.032), reoperation (OR = 3.6 [1.4, 9.3], P = 0.008), and sepsis (OR = 3.4 [1.1, 10.4], P = 0.031). CONCLUSION:This is the first large study to document abnormal coagulation profile as an independent predictor of outcomes following PCLF. Abnormal coagulation profile represents a predictor of complications that can be medically mitigated, and is therefore a valuable parameter to assess preoperatively. Coagulation profile should continue to play a role in targeting patients for risk stratification, preoperative optimization, and quality improvement initiatives. LEVEL OF EVIDENCE:3.

Stroke is a leading cause of long-term disability, and outcome is directly related to timely intervention. Not all patients benefit from rapid intervention, however. Thus a significant amount of attention has been paid to using neuroimaging to assess potential benefit by identifying areas of ischemia that have not yet experienced cellular death. The perfusion-diffusion mismatch, is used as a simple metric for potential benefit with timely intervention, yet penumbral patterns provide an inaccurate predictor of clinical outcome. Machine learning research in the form of deep learning (artificial intelligence) techniques using deep neural networks (DNNs) excel at working with complex inputs. The key areas where deep learning may be imminently applied to stroke management are image segmentation, automated featurization (radiomics), and multimodal prognostication. The application of convolutional neural networks, the family of DNN architectures designed to work with images, to stroke imaging data is a perfect match between a mature deep learning technique and a data type that is naturally suited to benefit from deep learning's strengths. These powerful tools have opened up exciting opportunities for data-driven stroke management for acute intervention and for guiding prognosis. Deep learning techniques are useful for the speed and power of results they can deliver and will become an increasingly standard tool in the modern stroke specialist's arsenal for delivering personalized medicine to patients with ischemic stroke.

STUDY DESIGN:A cross-sectional database study. OBJECTIVE:The aim of this study was to train and validate machine learning models to identify risk factors for complications following posterior lumbar spine fusion. SUMMARY OF BACKGROUND DATA:Machine learning models such as artificial neural networks (ANNs) are valuable tools for analyzing and interpreting large and complex datasets. ANNs have yet to be used for risk factor analysis in orthopedic surgery. METHODS:The American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) database was queried for patients who underwent posterior lumbar spine fusion. This query returned 22,629 patients, 70% of whom were used to train our models, and 30% were used to evaluate the models. The predictive variables used included sex, age, ethnicity, diabetes, smoking, steroid use, coagulopathy, functional status, American Society for Anesthesiology (ASA) class ≥3, body mass index (BMI), pulmonary comorbidities, and cardiac comorbidities. The models were used to predict cardiac complications, wound complications, venous thromboembolism (VTE), and mortality. Using ASA class as a benchmark for prediction, area under receiver operating curves (AUC) was used to determine the accuracy of our machine learning models. RESULTS:On the basis of AUC values, ANN and LR both outperformed ASA class for predicting all four types of complications. ANN was the most accurate for predicting cardiac complications, and LR was most accurate for predicting wound complications, VTE, and mortality, though ANN and LR had comparable AUC values for predicting all types of complications. ANN had greater sensitivity than LR for detecting wound complications and mortality. CONCLUSION:Machine learning in the form of logistic regression and ANNs were more accurate than benchmark ASA scores for identifying risk factors of developing complications following posterior lumbar spine fusion, suggesting they are potentially great tools for risk factor analysis in spine surgery. LEVEL OF EVIDENCE:3.