Machine learning in neurosurgery: a global survey

Understanding Physician's Perspectives on AI in Health Care: Protocol for a Sequential Multiple Assignment Randomized Vignette Study

BACKGROUND

As the availability and performance of artificial intelligence (AI)-based clinical decision support (CDS) systems improve, physicians and other care providers poised to be on the front lines will be increasingly tasked with using these tools in patient care and incorporating their outputs into clinical decision-making processes. Vignette studies provide a means to explore emerging hypotheses regarding how context-specific factors, such as clinical risk, the amount of information provided about the AI, and the AI result, may impact physician acceptance and use of AI-based CDS tools. To best anticipate how such factors influence the decision-making of frontline physicians in clinical scenarios involving AI decision-support tools, hypothesis-driven research is needed that enables scenario testing before the implementation and deployment of these tools.

OBJECTIVE

This study's objectives are to (1) design an original, web-based vignette-based survey that features hypothetical scenarios based on emerging or real-world applications of AI-based CDS systems that will vary systematically by features related to clinical risk, the amount of information provided about the AI, and the AI result; and (2) test and determine causal effects of specific factors on the judgments and perceptions salient to physicians' clinical decision-making.

METHODS

US-based physicians with specialties in family or internal medicine will be recruited through email and mail (target n=420). Through a web-based survey, participants will be randomized to a 3-part "sequential multiple assignment randomization trial (SMART) vignette" detailing a hypothetical clinical scenario involving an AI decision support tool. The SMART vignette design is similar to the SMART design but adapted to a survey design. Each respondent will be randomly assigned to 1 of the possible vignette variations of the factors we are testing at each stage, which include the level of clinical risk, the amount of information provided about the AI, and the certainty of the AI output. Respondents will be given questions regarding their hypothetical decision-making in response to the hypothetical scenarios.

RESULTS

The study is currently in progress and data collection is anticipated to be completed in 2024.

CONCLUSIONS

The web-based vignette study will provide information on how contextual factors such as clinical risk, the amount of information provided about an AI tool, and the AI result influence physicians' reactions to hypothetical scenarios that are based on emerging applications of AI in frontline health care settings. Our newly proposed "SMART vignette" design offers several benefits not afforded by the extensively used traditional vignette design, due to the 2 aforementioned features. These advantages are (1) increased validity of analyses targeted at understanding the impact of a factor on the decision outcome, given previous outcomes and other contextual factors; and (2) balanced sample sizes across groups. This study will generate a better understanding of physician decision-making within this context.

INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID)

DERR1-10.2196/54787.

Virtual Reality in Preoperative Planning of Complex Cranial Surgery

OBJECTIVE

Changing paradigms of neurosurgical training and limited operative exposure during the residency period have made it necessary to evaluate newer technologies for training. Virtual reality (VR) technology provides three-dimensional reconstruction of routine imaging, along with the ability to see as well as interact. The application of VR technology in operative planning, which is an important part of neurosurgical training, has been incompletely studied so far.

METHODS

Sixteen final-year residents, post-M.Ch. (magister chirurgiae) residents, and fellows were included as study participants. They were divided into 2 groups based on their seniority for further analysis. Five complex cranial cases were selected and a multiple-choice question-based test was prepared by the authors, with 5 questions for each of the cases. The pretest score was determined based on performance on the test after participants accessed routine preoperative imaging. The posttest score was calculated after use of the VR system (ImmersiveTouch VR System, ImmersiveTouch Inc.). Analysis was performed by the investigators, who were blinded to the identity of the participant. Subanalysis based on the type of case and type of question was performed. Feedback was obtained from each participant regarding VR use.

RESULTS

There was an overall improvement in scores from pretest to posttest, which was also noted in the analysis based on the participants' seniority. This improvement was noted to be more for the vascular cases (15.89%) compared with the tumor cases (7.84%). Participants also fared better in questions related to surgical anatomy and surgical approach, compared with questions based on the diagnosis. There was overall positive feedback from participants regarding VR use, and most participants wanted VR to become a routine part of operative planning.

CONCLUSIONS

Our study shows that there is improvement in understanding of surgical aspects after use of this VR system.

Multi-stakeholder preferences for the use of artificial intelligence in healthcare: A systematic review and thematic analysis

INTRODUCTION

Despite the proliferation of Artificial Intelligence (AI) technology over the last decade, clinician, patient, and public perceptions of its use in healthcare raise a number of ethical, legal and social questions. We systematically review the literature on attitudes towards the use of AI in healthcare from patients, the general public and health professionals' perspectives to understand these issues from multiple perspectives.

METHODOLOGY

A search for original research articles using qualitative, quantitative, and mixed methods published between 1 Jan 2001 to 24 Aug 2021 was conducted on six bibliographic databases. Data were extracted and classified into different themes representing views on: (i) knowledge and familiarity of AI, (ii) AI benefits, risks, and challenges, (iii) AI acceptability, (iv) AI development, (v) AI implementation, (vi) AI regulations, and (vii) Human - AI relationship.

RESULTS

The final search identified 7,490 different records of which 105 publications were selected based on predefined inclusion/exclusion criteria. While the majority of patients, the general public and health professionals generally had a positive attitude towards the use of AI in healthcare, all groups indicated some perceived risks and challenges. Commonly perceived risks included data privacy; reduced professional autonomy; algorithmic bias; healthcare inequities; and greater burnout to acquire AI-related skills. While patients had mixed opinions on whether healthcare workers suffer from job loss due to the use of AI, health professionals strongly indicated that AI would not be able to completely replace them in their professions. Both groups shared similar doubts about AI's ability to deliver empathic care. The need for AI validation, transparency, explainability, and patient and clinical involvement in the development of AI was emphasised. To help successfully implement AI in health care, most participants envisioned that an investment in training and education campaigns was necessary, especially for health professionals. Lack of familiarity, lack of trust, and regulatory uncertainties were identified as factors hindering AI implementation. Regarding AI regulations, key themes included data access and data privacy. While the general public and patients exhibited a willingness to share anonymised data for AI development, there remained concerns about sharing data with insurance or technology companies. One key domain under this theme was the question of who should be held accountable in the case of adverse events arising from using AI.

CONCLUSIONS

While overall positivity persists in attitudes and preferences toward AI use in healthcare, some prevalent problems require more attention. There is a need to go beyond addressing algorithm-related issues to look at the translation of legislation and guidelines into practice to ensure fairness, accountability, transparency, and ethics in AI.

Demonstrating the successful application of synthetic learning in spine surgery for training multi-center models with increased patient privacy

From real-time tumor classification to operative outcome prediction, applications of machine learning to neurosurgery are powerful. However, the translation of many of these applications are restricted by the lack of "big data" in neurosurgery. Important restrictions in patient privacy and sharing of imaging data reduce the diversity of the datasets used to train resulting models and therefore limit generalizability. Synthetic learning is a recent development in machine learning that generates synthetic data from real data and uses the synthetic data to train downstream models while preserving patient privacy. Such an approach has yet to be successfully demonstrated in the spine surgery domain. Spine radiographs were collected from the VinDR-SpineXR dataset, with 1470 labeled as abnormal and 2303 labeled as normal. A conditional generative adversarial network (GAN) was trained on the radiographs to generate a spine radiograph and normal/abnormal label. A modified conditional GAN (SpineGAN) was trained on the same task. A convolutional neural network (CNN) was trained using the real data to label abnormal radiographs. A CNN was trained to label abnormal radiographs using synthetic images from the GAN and in a separate experiment from SpineGAN. Using the real radiographs, an AUC of 0.856 was achieved in abnormality classification. Training on synthetic data generated by the standard GAN (AUC of 0.814) and synthetic data generated by our SpineGAN (AUC of 0.830) resulted in similar classifier performance. SpineGAN generated images with higher FID and lower precision scores, but with higher recall and increased performance when used for synthetic learning. The successful application of synthetic learning was demonstrated in the spine surgery domain for the classification of spine radiographs as abnormal or normal. A modified domain-relevant GAN is introduced for the generation of spine images, evidencing the importance of domain-relevant generation techniques in synthetic learning. Synthetic learning can allow neurosurgery to use larger and more diverse patient imaging sets to train more generalizable algorithms with greater patient privacy.

Artificial Intelligence in Neurosurgery: A State-of-the-Art Review from Past to Future

In recent years, there has been a significant surge in discussions surrounding artificial intelligence (AI), along with a corresponding increase in its practical applications in various facets of everyday life, including the medical industry. Notably, even in the highly specialized realm of neurosurgery, AI has been utilized for differential diagnosis, pre-operative evaluation, and improving surgical precision. Many of these applications have begun to mitigate risks of intraoperative and postoperative complications and post-operative care. This article aims to present an overview of the principal published papers on the significant themes of tumor, spine, epilepsy, and vascular issues, wherein AI has been applied to assess its potential applications within neurosurgery. The method involved identifying high-cited seminal papers using PubMed and Google Scholar, conducting a comprehensive review of various study types, and summarizing machine learning applications to enhance understanding among clinicians for future utilization. Recent studies demonstrate that machine learning (ML) holds significant potential in neuro-oncological care, spine surgery, epilepsy management, and other neurosurgical applications. ML techniques have proven effective in tumor identification, surgical outcomes prediction, seizure outcome prediction, aneurysm prediction, and more, highlighting its broad impact and potential in improving patient management and outcomes in neurosurgery. This review will encompass the current state of research, as well as predictions for the future of AI within neurosurgery.

Robot-assisted screw fixation in a cadaver utilizing magnetic resonance imaging-based synthetic computed tomography: toward radiation-free spine surgery. Illustrative case

BACKGROUND

Synthetic computed tomography (sCT) can be created from magnetic resonance imaging (MRI) utilizing newer software. sCT is yet to be explored as a possible alternative to routine CT (rCT). In this study, rCT scans and MRI-derived sCT scans were obtained on a cadaver. Morphometric analysis was performed comparing the 2 scans. The ExcelsiusGPS robot was used to place lumbosacral screws with both rCT and sCT images.

OBSERVATIONS

In total, 14 screws were placed. All screws were grade A on the Gertzbein-Robbins scale. The mean surface distance difference between rCT and sCT on a reconstructed software model was -0.02 ± 0.05 mm, the mean absolute surface distance was 0.24 ± 0.05 mm, and the mean absolute error of radiodensity was 92.88 ± 10.53 HU. The overall mean tip distance for the sCT versus rCT was 1.74 ± 1.1 versus 2.36 ± 1.6 mm (p = 0.24); mean tail distance for the sCT versus rCT was 1.93 ± 0.88 versus 2.81 ± 1.03 mm (p = 0.07); and mean angular deviation for the sCT versus rCT was 3.2° ± 2.05° versus 4.04°± 2.71° (p = 0.53).

LESSONS

MRI-based sCT yielded results comparable to those of rCT in both morphometric analysis and robot-assisted lumbosacral screw placement in a cadaver study.

Perceptions of Doctors in Saudi Arabia Toward Virtual Reality and Augmented Reality Applications in Healthcare

Background Several studies suggested that artificial intelligence (AI), including virtual reality (VR) and augmented reality (AR), may help improve visualization, diagnostic, and therapeutic abilities and reduce medical and surgical errors. These technologies have been revolutionary in Saudi Arabia. We aimed to elucidate physicians' perceptions toward these technologies. Methodology We carried out a cross-sectional electronic questionnaire-based study in November 2021. The study targeted doctors of different medical and surgical specialties in the western region of Saudi Arabia. Results In our study, 53.2% of the participants were 25-30 years old. Most participants were residents (53.6%) with career experiences <5 years. Only 32.3% had a good familiarity with AR and VR technologies. However, 64.5% agreed that AR and VR technologies had practical applications in the medical field. Moreover, 35% agreed that the diagnostic and therapeutic ability was superior to the clinical experience of a human doctor. About 41.4% agreed they would always use AR and VR technologies for future medical decisions. Conclusion Doctors are open to using AR and VR technologies in healthcare. Although most people are unfamiliar with these technologies, most agree that they positively impact healthcare.

Attitudes of Anesthesiologists toward Artificial Intelligence in Anesthesia: A Multicenter, Mixed Qualitative-Quantitative Study

Artificial intelligence (AI) is predicted to play an increasingly important role in perioperative medicine in the very near future. However, little is known about what anesthesiologists know and think about AI in this context. This is important because the successful introduction of new technologies depends on the understanding and cooperation of end users. We sought to investigate how much anesthesiologists know about AI and what they think about the introduction of AI-based technologies into the clinical setting. In order to better understand what anesthesiologists think of AI, we recruited 21 anesthesiologists from 2 university hospitals for face-to-face structured interviews. The interview transcripts were subdivided sentence-by-sentence into discrete statements, and statements were then grouped into key themes. Subsequently, a survey of closed questions based on these themes was sent to 70 anesthesiologists from 3 university hospitals for rating. In the interviews, the base level of knowledge of AI was good at 86 of 90 statements (96%), although awareness of the potential applications of AI in anesthesia was poor at only 7 of 42 statements (17%). Regarding the implementation of AI in anesthesia, statements were split roughly evenly between pros (46 of 105, 44%) and cons (59 of 105, 56%). Interviewees considered that AI could usefully be used in diverse tasks such as risk stratification, the prediction of vital sign changes, or as a treatment guide. The validity of these themes was probed in a follow-up survey of 70 anesthesiologists with a response rate of 70%, which confirmed an overall positive view of AI in this group. Anesthesiologists hold a range of opinions, both positive and negative, regarding the application of AI in their field of work. Survey-based studies do not always uncover the full breadth of nuance of opinion amongst clinicians. Engagement with specific concerns, both technical and ethical, will prove important as this technology moves from research to the clinic.

Artificial Intelligence in Neurosurgery: A Bibliometric Analysis

BACKGROUND

Artificial intelligence (AI) has the potential to augment clinicians' diagnostic and decision-making capabilities. It is well suited to identify patterns and correlations within data sets and may be applied to identify elements of importance in complex and data-laden areas such as patient selection, diagnostics, treatment, and outcome prediction. The development of modern neurosurgery has been dependent on major technological advances. In line with this, a growing interest is seen in the use of AI to assist in neurosurgical research and enhance neurosurgical practices.

METHODS

A bibliometric analysis of the 50 most-cited articles alluding to the use of AI in neurosurgery, from inception until July of 2022, was undertaken using the Web of Science database. Statistical analyses were performed on R.

RESULTS

The citation count ranged from 29 to 159 (mean: 51.9, standard deviation: 24.8), and the top-cited article was a 2018 systematic review published in World Neurosurgery. Most articles were published after 2015 (85%). The United States was the largest contributing country on the list with 22 articles. Four first and last authors, each, had 2 or more publications. Female first and last authorship was attributed to 18% and 0% of the articles, respectively.

CONCLUSIONS

This review highlights the most-impactful articles pertaining to AI in the field of neurosurgery. Although female authors were significantly underrepresented on the list, their work was at least as impactful as their male peers. Finally, the striking dominance of articles originating from the developed world raises concerns as to the future of AI in attending to the global health crisis.

Limits on using the clock drawing test as a measure to evaluate patients with neurological disorders

BACKGROUND

The Clock Drawing Test (CDT) is used as a quick-to-conduct test for the diagnosis of dementia and a screening tool for cognitive impairments in neurological disorders. However, the association between the pattern of CDT impairments and the location of brain lesions has been controversial. We examined whether there is an association between the CDT scores and the location of brain lesions using the two available scoring systems.

METHOD

One hundred five patients with brain lesions identified by CT scanning were recruited for this study. The Montreal Cognitive Assessment (MoCA) battery including the CDT were administered to all partcipants. To score the CDT, we used a qualitative scoring system devised by Rouleau et al. (1992). For the quantitative scoring system, we adapted the algorithm method used by Mendes-Santos et al. (2015) based on an earlier study by Sunderland et al. (1989). For analyses, a machine learning algorithm was used.

RESULTS

Remarkably, 30% of the patients were not detected by the CDT. Quantitative and qualitative errors were categorized into different clusters. The classification algorithm did not differentiate the patients with traumatic brain injury 'TBI' from non-TBI, or the laterality of the lesion. In addition, the classification accuracy for identifying patients with specific lobe lesions was low, except for the parietal lobe with an accuracy of 63%.

CONCLUSION

The CDT is not an accurate tool for detecting focal brain lesions. While the CDT still is beneficial for use with patients suspected of having a neurodegenerative disorder, it should be cautiously used with patients with focal neurological disorders.

A computer vision approach to identifying the manufacturer of posterior thoracolumbar instrumentation systems

OBJECTIVE

Knowledge of the manufacturer of the previously implanted pedicle screw systems prior to revision spinal surgery may facilitate faster and safer surgery. Often, this information is unavailable because patients are referred by other centers or because of missing information in the patients' records. Recently, machine learning and computer vision have gained wider use in clinical applications. The authors propose a computer vision approach to classify posterior thoracolumbar instrumentation systems.

METHODS

Lateral and anteroposterior (AP) radiographs obtained in patients undergoing posterior thoracolumbar pedicle screw implantation for any indication at the authors' institution (2015-2021) were obtained. DICOM images were cropped to include both the pedicle screws and rods. Images were labeled with the manufacturer according to the operative record. Multiple feature detection methods were tested (SURF, MESR, and Minimum Eigenvalues); however, the bag-of-visual-words technique with KAZE feature detection was ultimately used to construct a computer vision support vector machine (SVM) classifier for lateral, AP, and fused lateral and AP images. Accuracy was tested using an 80%/20% training/testing pseudorandom split over 100 iterations. Using a reader study, the authors compared the model performance with the current practice of surgeons and manufacturer representatives identifying spinal hardware by visual inspection.

RESULTS

Among the three image types, 355 lateral, 379 AP, and 338 fused radiographs were obtained. The five pedicle screw implants included in this study were the Globus Medical Creo, Medtronic Solera, NuVasive Reline, Stryker Xia, and DePuy Expedium. When the two most common manufacturers used at the authors' institution were binarily classified (Globus Medical and Medtronic), the accuracy rates for lateral, AP, and fused images were 93.15% ± 4.06%, 88.98% ± 4.08%, and 91.08% ± 5.30%, respectively. Classification accuracy decreased by approximately 10% with each additional manufacturer added. The multilevel five-way classification accuracy rates for lateral, AP, and fused images were 64.27% ± 5.13%, 60.95% ± 5.52%, and 65.90% ± 5.14%, respectively. In the reader study, the model performed five-way classification on 100 test images with 79% accuracy in 14 seconds, compared with an average of 44% accuracy in 20 minutes for two surgeons and three manufacturer representatives.

CONCLUSIONS

The authors developed a KAZE feature detector with an SVM classifier that successfully identified posterior thoracolumbar hardware at five-level classification. The model performed more accurately and efficiently than the method currently used in clinical practice. The relative computational simplicity of this model, from input to output, may facilitate future prospective studies in the clinical setting.

Neurosurgery inpatient outcome prediction for discharge planning with deep learning and transfer learning

INTRODUCTION

Deep learning may be able to assist with the prediction of neurosurgical inpatient outcomes. The aims of this study were to investigate deep learning and transfer learning in the prediction of several inpatient outcomes including timing of discharge and discharge destination.

METHOD

Data were collected on consecutive neurosurgical admissions from existing databases over a 15-month period. Following pre-processing artificial neural networks were applied to admission notes and ward round notes to predict four inpatient outcomes. Models were developed on the training dataset, before being tested on a hold-out test dataset and a validation dataset.

RESULTS

1341 individual admissions were included in the study. Using transfer learning and an artificial neural network an area under the receiver operator curve (AUC) of 0.81 and 0.80 on the derivation and validation datasets was able to be achieved for the prediction of discharge within the next 48 hours using daily ward round notes. This result is in comparison to an AUC of 0.71 and 0.68 using an artificial neural network without transfer learning for the same outcome. When the artificial neural network with transfer learning was applied to the other outcomes AUC of 0.72, 0.93 and 0.83 was achieved on the validation datasets for predicting discharge within the next 7 days, survival to discharge and discharge to home as a destination.

CONCLUSIONS

Deep learning may predict inpatient neurosurgery outcomes from free-text medical data. Recurrent predictions with ward round notes enable the use of information obtained throughout hospital admissions in these estimates.

Comparison of Deep Learning and Classical Machine Learning Algorithms to Predict Postoperative Outcomes for Anterior Cervical Discectomy and Fusion Procedures With State-of-the-art Performance

STUDY DESIGN

Retrospective cohort.

OBJECTIVE

Due to anterior cervical discectomy and fusion (ACDF) popularity, it is important to predict postoperative complications, unfavorable 90-day readmissions, and two-year reoperations to improve surgical decision-making, prognostication, and planning.

SUMMARY OF BACKGROUND DATA

Machine learning has been applied to predict postoperative complications for ACDF; however, studies were limited by sample size and model type. These studies achieved ≤0.70 area under the curve (AUC). Further approaches, not limited to ACDF, focused on specific complication types and resulted in AUC between 0.70 and 0.76.

MATERIALS AND METHODS

The IBM MarketScan Commercial Claims and Encounters Database and Medicare Supplement were queried from 2007 to 2016 to identify adult patients who underwent an ACDF procedure (N=176,816). Traditional machine learning algorithms, logistic regression, and support vector machines, were compared with deep neural networks to predict: 90-day postoperative complications, 90-day readmission, and two-year reoperation. We further generated random deep learning model architectures and trained them on the 90-day complication task to approximate an upper bound. Last, using deep learning, we investigated the importance of each input variable for the prediction of 90-day postoperative complications in ACDF.

RESULTS

For the prediction of 90-day complication, 90-day readmission, and two-year reoperation, the deep neural network-based models achieved AUC of 0.832, 0.713, and 0.671. Logistic regression achieved AUCs of 0.820, 0.712, and 0.671. Support vector machine approaches were significantly lower. The upper bound of deep learning performance was approximated as 0.832. Myelopathy, age, human immunodeficiency virus, previous myocardial infarctions, obesity, and documentary weakness were found to be the strongest variable to predict 90-day postoperative complications.

CONCLUSIONS

The deep neural network may be used to predict complications for clinical applications after multicenter 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.

Machine learning in clinical diagnosis, prognostication, and management of acute traumatic spinal cord injury (SCI): A systematic review

Background

Machine learning has been applied to improve diagnosis and prognostication of acute traumatic spinal cord injury. We investigate potential for clinical integration of machine learning in this patient population to navigate variability in injury and recovery.

Materials and methods

We performed a systematic review using PRISMA guidelines through PubMed database to identify studies that use machine learning algorithms for clinical application toward improvements in diagnosis, management, and predictive modeling.

Results

Of the 132 records identified, a total of 13 articles met inclusion criteria and were included in final analysis. Of the 13 articles, 5 focused on diagnostic accuracy and 8 were related to prognostication or management of traumatic spinal cord injury. Across studies, 1983 patients with spinal cord injury were evaluated with most classifying as ASIA C or D. Retrospective designs were used in 10 of 13 studies and 3 were prospective. Studies focused on MRI evaluation and segmentation for diagnostic accuracy and prognostication, investigation of mean arterial pressure in acute care and intraoperative settings, prediction of ambulatory and functional ability, chronic complication prevention, and psychological quality of life assessments. Decision tree, random forests (RF), support vector machines (SVM), hierarchical cluster tree analysis (HCTA), artificial neural networks (ANN), convolutional neural networks (CNN) machine learning subtypes were used.

Conclusions

Machine learning represents a platform technology with clinical application in traumatic spinal cord injury diagnosis, prognostication, management, rehabilitation, and risk prevention of chronic complications and mental illness. SVM models showed improved accuracy when compared to other ML subtypes surveyed. Inherent variability across patients with SCI offers unique opportunity for ML and personalized medicine to drive desired outcomes and assess risks in this patient population.

Current Attitudes Toward Neuroanatomy: A Comparative Cross-Sectional Survey of Neurosurgeons from the United Kingdom and Worldwide

OBJECTIVE

How attitudes toward neuroanatomy and preferences of studying resources vary among neurosurgeons is unknown. The impact of the coronavirus disease 2019 (COVID-19) pandemic on anatomy learning habits is also yet to be elucidated. In this study, we explore these objectives, to guide the development of future neurosurgeon-tailored anatomy education and resources.

METHODS

This was a 2-stage, cross-sectional study design comprising a local pilot survey followed by a structured 17-item questionnaire, distributed to both neurosurgical trainees and consultants. Grade and nationality differences in sentiment agreement were statistically compared.

RESULTS

A total of 365 responses were received from 32 countries (overall response rate, 23.2%). Neuroanatomy is highly regarded among most neurosurgeons and takes a central role in their professional identity. Yet, 69% of neurosurgeons wanted to spend more time learning. Common study prompts included perceived operative complexity, lack of familiarity and teaching. Financial barriers and motivation were obstacles limiting neuroanatomy learning, more so among trainee neurosurgeons, with personal commitment barriers significantly varying with geographic location. Surgical relevance, accessibility, and image quality were important factors when selecting anatomy resources, with cost and up-to-datedness being important for juniors. The COVID-19 pandemic saw a shift toward virtual resources, particularly affecting United Kingdom-based trainees.

CONCLUSIONS

Although neuroanatomy is well regarded, barriers exist that impede further neuroanatomy learning. Neurosurgical training programs should tailor anatomy education according to the seniority and background of their residents. Furthermore, resources that are surgically relevant and accessible and are of high image quality are more likely to be better used.

Acceptance of clinical artificial intelligence among physicians and medical students: A systematic review with cross-sectional survey

Background

Artificial intelligence (AI) needs to be accepted and understood by physicians and medical students, but few have systematically assessed their attitudes. We investigated clinical AI acceptance among physicians and medical students around the world to provide implementation guidance.

Materials and methods

We conducted a two-stage study, involving a foundational systematic review of physician and medical student acceptance of clinical AI. This enabled us to design a suitable web-based questionnaire which was then distributed among practitioners and trainees around the world.

Results

Sixty studies were included in this systematic review, and 758 respondents from 39 countries completed the online questionnaire. Five (62.50%) of eight studies reported 65% or higher awareness regarding the application of clinical AI. Although, only 10–30% had actually used AI and 26 (74.28%) of 35 studies suggested there was a lack of AI knowledge. Our questionnaire uncovered 38% awareness rate and 20% utility rate of clinical AI, although 53% lacked basic knowledge of clinical AI. Forty-five studies mentioned attitudes toward clinical AI, and over 60% from 38 (84.44%) studies were positive about AI, although they were also concerned about the potential for unpredictable, incorrect results. Seventy-seven percent were optimistic about the prospect of clinical AI. The support rate for the statement that AI could replace physicians ranged from 6 to 78% across 40 studies which mentioned this topic. Five studies recommended that efforts should be made to increase collaboration. Our questionnaire showed 68% disagreed that AI would become a surrogate physician, but believed it should assist in clinical decision-making. Participants with different identities, experience and from different countries hold similar but subtly different attitudes.

Conclusion

Most physicians and medical students appear aware of the increasing application of clinical AI, but lack practical experience and related knowledge. Overall, participants have positive but reserved attitudes about AI. In spite of the mixed opinions around clinical AI becoming a surrogate physician, there was a consensus that collaborations between the two should be strengthened. Further education should be conducted to alleviate anxieties associated with change and adopting new technologies.

Automated prediction of the Thoracolumbar Injury Classification and Severity Score from CT using a novel deep learning algorithm

OBJECTIVE

Damage to the thoracolumbar spine can confer significant morbidity and mortality. The Thoracolumbar Injury Classification and Severity Score (TLICS) is used to categorize injuries and determine patients at risk of spinal instability for whom surgical intervention is warranted. However, calculating this score can constitute a bottleneck in triaging and treating patients, as it relies on multiple imaging studies and a neurological examination. Therefore, the authors sought to develop and validate a deep learning model that can automatically categorize vertebral morphology and determine posterior ligamentous complex (PLC) integrity, two critical features of TLICS, using only CT scans.

METHODS

All patients who underwent neurosurgical consultation for traumatic spine injury or degenerative pathology resulting in spine injury at a single tertiary center from January 2018 to December 2019 were retrospectively evaluated for inclusion. The morphology of injury and integrity of the PLC were categorized on CT scans. A state-of-the-art object detection region-based convolutional neural network (R-CNN), Faster R-CNN, was leveraged to predict both vertebral locations and the corresponding TLICS. The network was trained with patient CT scans, manually labeled vertebral bounding boxes, TLICS morphology, and PLC annotations, thus allowing the model to output the location of vertebrae, categorize their morphology, and determine the status of PLC integrity.

RESULTS

A total of 111 patients were included (mean ± SD age 62 ± 20 years) with a total of 129 separate injury classifications. Vertebral localization and PLC integrity classification achieved Dice scores of 0.92 and 0.88, respectively. Binary classification between noninjured and injured morphological scores demonstrated 95.1% accuracy. TLICS morphology accuracy, the true positive rate, and positive injury mismatch classification rate were 86.3%, 76.2%, and 22.7%, respectively. Classification accuracy between no injury and suspected PLC injury was 86.8%, while true positive, false negative, and false positive rates were 90.0%, 10.0%, and 21.8%, respectively.

CONCLUSIONS

In this study, the authors demonstrate a novel deep learning method to automatically predict injury morphology and PLC disruption with high accuracy. This model may streamline and improve diagnostic decision support for patients with thoracolumbar spinal trauma.

Hemodynamic Imaging in Cerebral Diffuse Glioma—Part B: Molecular Correlates, Treatment Effect Monitoring, Prognosis, and Future Directions

Gliomas, and glioblastoma in particular, exhibit an extensive intra- and inter-tumoral molecular heterogeneity which represents complex biological features correlating to the efficacy of treatment response and survival. From a neuroimaging point of view, these specific molecular and histopathological features may be used to yield imaging biomarkers as surrogates for distinct tumor genotypes and phenotypes. The development of comprehensive glioma imaging markers has potential for improved glioma characterization that would assist in the clinical work-up of preoperative treatment planning and treatment effect monitoring. In particular, the differentiation of tumor recurrence or true progression from pseudoprogression, pseudoresponse, and radiation-induced necrosis can still not reliably be made through standard neuroimaging only. Given the abundant vascular and hemodynamic alterations present in diffuse glioma, advanced hemodynamic imaging approaches constitute an attractive area of clinical imaging development. In this context, the inclusion of objective measurable glioma imaging features may have the potential to enhance the individualized care of diffuse glioma patients, better informing of standard-of-care treatment efficacy and of novel therapies, such as the immunotherapies that are currently increasingly investigated. In Part B of this two-review series, we assess the available evidence pertaining to hemodynamic imaging for molecular feature prediction, in particular focusing on isocitrate dehydrogenase (IDH) mutation status, MGMT promoter methylation, 1p19q codeletion, and EGFR alterations. The results for the differentiation of tumor progression/recurrence from treatment effects have also been the focus of active research and are presented together with the prognostic correlations identified by advanced hemodynamic imaging studies. Finally, the state-of-the-art concepts and advancements of hemodynamic imaging modalities are reviewed together with the advantages derived from the implementation of radiomics and machine learning analyses pipelines.

Prediction of Shunt Responsiveness in Suspected Patients With Normal Pressure Hydrocephalus Using the Lumbar Infusion Test: A Machine Learning Approach

BACKGROUND

Machine learning (ML) approaches can significantly improve the classical Rout-based evaluation of the lumbar infusion test (LIT) and the clinical management of the normal pressure hydrocephalus.

OBJECTIVE

To develop a ML model that accurately identifies patients as candidates for permanent cerebral spinal fluid shunt implantation using only intracranial pressure and electrocardiogram signals recorded throughout LIT.

METHODS

This was a single-center cohort study of prospectively collected data of 96 patients who underwent LIT and 5-day external lumbar cerebral spinal fluid drainage (external lumbar drainage) as a reference diagnostic method. A set of selected 48 intracranial pressure/electrocardiogram complex signal waveform features describing nonlinear behavior, wavelet transform spectral signatures, or recurrent map patterns were calculated for each patient. After applying a leave-one-out cross-validation training-testing split of the data set, we trained and evaluated the performance of various state-of-the-art ML algorithms.

RESULTS

The highest performing ML algorithm was the eXtreme Gradient Boosting. This model showed a good calibration and discrimination on the testing data, with an area under the receiver operating characteristic curve of 0.891 (accuracy: 82.3%, sensitivity: 86.1%, and specificity: 73.9%) obtained for 8 selected features. Our ML model clearly outperforms the classical Rout-based manual classification commonly used in clinical practice with an accuracy of 62.5%.

CONCLUSION

This study successfully used the ML approach to predict the outcome of a 5-day external lumbar drainage and hence which patients are likely to benefit from permanent shunt implantation. Our automated ML model thus enhances the diagnostic utility of LIT in management.

Feature Importance Analysis for Postural Deformity Detection System Using Explainable Predictive Modeling Technique

This study aimed to analyze feature importance by applying explainable artificial intelligence (XAI) to postural deformity parameters extracted from a computer vision-based posture analysis system (CVPAS). Overall, 140 participants were screened for CVPAS and enrolled. The main data analyzed were shoulder height difference (SHD), wrist height difference (WHD), and pelvic height difference (PHD) extracted using a CVPAS. Standing X-ray imaging and radiographic assessments were performed. Predictive modeling was implemented with XGBoost, random forest regressor, and logistic regression using XAI techniques for global and local feature analyses. Correlation analysis was performed between radiographic assessment and AI evaluation for PHD, SHD, and Cobb angle. Main global features affecting scoliosis were analyzed in the order of importance for PHD (0.18) and ankle height difference (0.06) in predictive modeling. Outstanding local features were PHD, WHD, and KHD that predominantly contributed to the increase in the probability of scoliosis, and the prediction probability of scoliosis was 94%. When the PHD was >3 mm, the probability of scoliosis increased sharply to 85.3%. The paired t-test result for AI and radiographic assessments showed that the SHD, Cobb angle, and scoliosis probability were significant (p < 0.05). Feature importance analysis using XAI to postural deformity parameters extracted from a CVPAS is a useful clinical decision support system for the early detection of posture deformities. PHD was a major parameter for both global and local analyses, and 3 mm was a threshold for significantly increasing the probability of local interpretation of each participant and the prediction of postural deformation, which leads to the prediction of participant-specific scoliosis.

Deep Learning for Outcome Prediction in Neurosurgery: A Systematic Review of Design, Reporting, and Reproducibility

Deep learning (DL) is a powerful machine learning technique that has increasingly been used to predict surgical outcomes. However, the large quantity of data required and lack of model interpretability represent substantial barriers to the validity and reproducibility of DL models. The objective of this study was to systematically review the characteristics of DL studies involving neurosurgical outcome prediction and to assess their bias and reporting quality. Literature search using the PubMed, Scopus, and Embase databases identified 1949 records of which 35 studies were included. Of these, 32 (91%) developed and validated a DL model while 3 (9%) validated a pre-existing model. The most commonly represented subspecialty areas were oncology (16 of 35, 46%), spine (8 of 35, 23%), and vascular (6 of 35, 17%). Risk of bias was low in 18 studies (51%), unclear in 5 (14%), and high in 12 (34%), most commonly because of data quality deficiencies. Adherence to transparent reporting of a multivariable prediction model for individual prognosis or diagnosis reporting standards was low, with a median of 12 transparent reporting of a multivariable prediction model for individual prognosis or diagnosis items (39%) per study not reported. Model transparency was severely limited because code was provided in only 3 studies (9%) and final models in 2 (6%). With the exception of public databases, no study data sets were readily available. No studies described DL models as ready for clinical use. The use of DL for neurosurgical outcome prediction remains nascent. Lack of appropriate data sets poses a major concern for bias. Although studies have demonstrated promising results, greater transparency in model development and reporting is needed to facilitate reproducibility and validation.

Machine Learning in Neuro-Oncology, Epilepsy, Alzheimer's Disease, and Schizophrenia

Applications of machine learning (ML) in translational medicine include therapeutic drug creation, diagnostic development, surgical planning, outcome prediction, and intraoperative assistance. Opportunities in the neurosciences are rich given advancement in our understanding of the brain, expanding indications for intervention, and diagnostic challenges often characterized by multiple clinical and environmental factors. We present a review of ML in neuro-oncology, epilepsy, Alzheimer's disease, and schizophrenia to highlight recent progression in these field, optimizing machine learning capabilities in their current forms. Supervised learning models appear to be the most commonly incorporated algorithm models for machine learning across the reviewed neuroscience disciplines with primary aim of diagnosis. Accuracy ranges are high from 63% to 99% across all algorithms investigated. Machine learning contributions to neurosurgery, neurology, psychiatry, and the clinical and basic science neurosciences may enhance current medical best practices while also broadening our understanding of dynamic neural networks and the brain.

A Brief History of Machine Learning in Neurosurgery

The history of machine learning in neurosurgery spans three decades and continues to develop at a rapid pace. The earliest applications of machine learning within neurosurgery were first published in the 1990s as researchers began developing artificial neural networks to analyze structured datasets and supervised tasks. By the turn of the millennium, machine learning had evolved beyond proof-of-concept; algorithms had success detecting tumors in unstructured clinical imaging, and unsupervised learning showed promise for tumor segmentation. Throughout the 2000s, the role of machine learning in neurosurgery was further refined. Well-trained models began to consistently best expert clinicians at brain tumor diagnosis. Additionally, the digitization of the healthcare industry provided ample data for analysis, both structured and unstructured. By the 2010s, the use of machine learning within neurosurgery had exploded. The rapid deployment of an exciting new toolset also led to the growing realization that it may offer marginal benefit at best over conventional logistical regression models for analyzing tabular datasets. Additionally, the widespread adoption of machine learning in neurosurgical clinical practice continues to lag until additional validation can ensure generalizability. Many exciting contemporary applications nonetheless continue to demonstrate the unprecedented potential of machine learning to revolutionize neurosurgery when applied to appropriate clinical challenges.

Machine Learning in Pituitary Surgery

Machine learning applications in neurosurgery are increasingly reported for diverse tasks such as faster and more accurate preoperative diagnosis, enhanced lesion characterization, as well as surgical outcome, complications and healthcare cost prediction. Even though the pertinent literature in pituitary surgery is less extensive with respect to other neurosurgical diseases, past research attempted to answer clinically relevant questions to better assist surgeons and clinicians. In the present chapter we review reported ML applications in pituitary surgery including differential diagnosis, preoperative lesion characterization (immunohistochemistry, cavernous sinus invasion, tumor consistency), surgical outcome and complication predictions (gross total resection, tumor recurrence, and endocrinological remission, cerebrospinal fluid leak, postoperative hyponatremia). Moreover, we briefly discuss from a practical standpoint the current barriers to clinical translation of machine learning research. On the topic of pituitary surgery, published reports can be considered mostly preliminary, requiring larger training populations and strong external validation. Thoughtful selection of clinically relevant outcomes of interest and transversal application of model development pipeline-together with accurate methodological planning and multicenter collaborations-have the potential to overcome current limitations and ultimately provide additional tools for more informed patient management.

Machine Intelligence in Clinical Neuroscience: Taming the Unchained Prometheus

The democratization of machine learning (ML) through availability of open-source learning libraries, the availability of datasets in the "big data" era, increasing computing power even on mobile devices, and online training resources have both led to an explosion in applications and publications of ML in the clinical neurosciences, but has also enabled a dangerous amount of flawed analyses and cardinal methodological errors committed by benevolent authors. While powerful ML methods are nowadays available to almost anyone and can be applied after just few minutes of familiarizing oneself with these methods, that does not imply that one has mastered these techniques. This textbook for clinicians aims to demystify ML by illustrating its methodological foundations, as well as some specific applications throughout clinical neuroscience, and its limitations. While our mind can recognize, abstract, and deal with the many uncertainties in clinical practice, algorithms cannot. Algorithms must remain tools of our own mind, tools that we should be able to master, control, and apply to our advantage in an adjunctive manner. Our hope is that this book inspires and instructs physician-scientists to continue to develop the seeds that have been planted for machine intelligence in clinical neuroscience, not forgetting their inherent limitations.

Machine Learning Algorithms in Neuroimaging: An Overview

Machine learning (ML) and artificial intelligence (AI) applications in the field of neuroimaging have been on the rise in recent years, and their clinical adoption is increasing worldwide. Deep learning (DL) is a field of ML that can be defined as a set of algorithms enabling a computer to be fed with raw data and progressively discover-through multiple layers of representation-more complex and abstract patterns in large data sets. The combination of ML and radiomics, namely the extraction of features from medical images, has proven valuable, too: Radiomic information can be used for enhanced image characterization and prognosis or outcome prediction. This chapter summarizes the basic concepts underlying ML application for neuroimaging and discusses technical aspects of the most promising algorithms, with a specific focus on Convolutional Neural Networks (CNNs) and Generative Adversarial Networks (GANs), in order to provide the readership with the fundamental theoretical tools to better understand ML in neuroimaging. Applications are highlighted from a practical standpoint in the last section of the chapter, including: image reconstruction and restoration, image synthesis and super-resolution, registration, segmentation, classification, and outcome prediction.

Wide range of applications for machine-learning prediction models in orthopedic surgical outcome: a systematic review

Background and purpose - Advancements in software and hardware have enabled the rise of clinical prediction models based on machine learning (ML) in orthopedic surgery. Given their growing popularity and their likely implementation in clinical practice we evaluated which outcomes these new models have focused on and what methodologies are being employed.Material and methods - We performed a systematic search in PubMed, Embase, and Cochrane Library for studies published up to June 18, 2020. Studies reporting on non-ML prediction models or non-orthopedic outcomes were excluded. After screening 7,138 studies, 59 studies reporting on 77 prediction models were included. We extracted data regarding outcome, study design, and reported performance metrics.Results - Of the 77 identified ML prediction models the most commonly reported outcome domain was medical management (17/77). Spinal surgery was the most commonly involved orthopedic subspecialty (28/77). The most frequently employed algorithm was neural networks (42/77). Median size of datasets was 5,507 (IQR 635-26,364). The median area under the curve (AUC) was 0.80 (IQR 0.73-0.86). Calibration was reported for 26 of the models and 14 provided decision-curve analysis.Interpretation - ML prediction models have been developed for a wide variety of topics in orthopedics. Topics regarding medical management were the most commonly studied. Heterogeneity between studies is based on study size, algorithm, and time-point of outcome. Calibration and decision-curve analysis were generally poorly reported.

Patterns of prophylactic anticonvulsant use in spontaneous intracerebral and subarachnoid hemorrhage: results of a practitioner survey

BACKGROUND

The use of prophylactic anti-seizure medications (ASMs) in the management of patients with spontaneous intracerebral hemorrhage (sICH) and aneurysmal subarachnoid hemorrhage (aSAH) is controversial.

OBJECTIVE

The purpose of this survey was to better characterize the current state of prophylactic ASM use in sICH and aSAH in North America.

METHODS

US and Canadian neurosurgeons, neurologists, and interventional neuroradiologists with an interest in or expertise in the management of neurovascular disease were surveyed using an electronic survey tool.

RESULTS

Seven hundred ninety-four survey requests were sent; responses were received from 103 (13%). The majority of respondents were neurosurgeons (84%). Thirty-eight percent of respondents self-identified as vascular neurosurgeons and 10% self-identified as neurocritical care specialists. Seventy-two percent were in academic practice. When asked their preference for ASM prophylaxis (aSAH, sICH, or both), the most common response was to use prophylaxis in both aSAH and sICH (43, 45%). Twenty-one (22%) did not use routine prophylaxis, while 22 (23%) used prophylaxis only in aSAH and 9 (9%) only in sICH. The majority of practitioners (35, 67%) who answered that they used ASM prophylaxis in sICH, used ASMs selectively. For aSAH, the vast majority (53, 82%) used prophylaxis for all patients. Respondents felt that they were more likely to use ASMs for sICH patients if the sICH was in a cortical location, supratentorial location, or was related to a structural abnormality (e.g., tumor, arteriovenous malformation) Levetiracetam (Keppra) was the most commonly used ASM (73, 99%). When asked whether the statement "Current AHA/ASA Guidelines recommend against the use of prophylactic anticonvulsants in spontaneous ICH" was true or false, 78 (83%) responded correctly that the recommendation is true. Only 24 respondents answered the question as to whether they would be willing to randomize sICH and/or aSAH patients to management with or without ASM prophylaxis. Of these, 13 (54%) said they would be willing to randomize sICH patients, while only 6 (25%) were willing to randomize aSAH patients. There were no statistically significant differences in responses to survey questions when analyzed by practice type (academic versus non-academic) or physician specialty (critical care versus non-critical care, or vascular neurology/neurosurgery versus other).

CONCLUSION

The use of ASMs for seizure prophylaxis after sICH and aSAH remains widespread despite the lack of any specific evidence-based guideline to support the practice. A large-scale randomized controlled trial is needed to add clarity to the practice of prophylactic ASM use in patients with spontaneous intracranial hemorrhage.

Educating the Next Generation of Global Neurosurgeons: Competencies, Skills, and Resources for Medical Students Interested in Global Neurosurgery

BACKGROUND

Global neurosurgery operates at the intersection of neurosurgery and public health. Although most global neurosurgery initiatives have targeted neurosurgeons and trainees, medical students represent the future of global neurosurgery.

METHODS

A narrative review of the literature regarding research methodology, education, economics, health policy, health advocacy, relevant to global neurosurgery was conducted.

RESULTS

We summarize pearls that all medical students interested in global neurosurgery should know.

DISCUSSION

To become effective agents of change within global neurosurgery, medical students must master competencies of motivation, organization, collaborativeness, dependability, flexibility, resilience, creative problem-solving, ethical thinking, cultural humility, and global awareness and gain knowledge and skills regarding research, education, policy making, and advocacy. Discussions with neurosurgeons and trainees, neurosurgery interest groups, conferences, university global neurosurgery initiatives, and student organizations represent opportunities for learning and becoming involved in global neurosurgery.

Assessment Practice of Patient-Centered Outcomes in Surgical Neuro-Oncology: Survey-Based Recommendations for Clinical Routine

The psycho-oncological burden related to the diagnosis of an intracranial tumor is often accompanied by neurocognitive deficits and changes in character, overall affecting health-related quality of life (HRQoL) and activities of daily living. Regular administration of adequate screening tools is crucial to ensure a timely detection of needs for support and/or specific interventions. Although efforts have been made to assure the quality of neuro-oncological care, clinical assessment practice of patient-reported outcomes (PROs) remains overall heterogeneous, calling for a concise recommendation tailored to neuro-oncological patients. Therefore, this survey, promoted by the German Society of Neurosurgery, was conducted to evaluate the status quo of health care resources and PRO/neurocognition assessment practices throughout departments of surgical neuro-oncology in Germany. 72/127 (57%) of registered departments participated in the study, including 83% of all university hospital units. A second aim was to shed light on the impact of quality assurance strategies (i.e., department certification as part of an integrative neuro-oncology cancer center; CNOC) on the assessment practice, controlled for interacting structural factors, i.e., university hospital status (UH) and caseload. Despite an overall good to excellent availability of relevant health care structures (psycho-oncologist: 90%, palliative care unit: 97%, neuropsychology: 75%), a small majority of departments practice patient-centered screenings (psycho-oncological burden: 64%, HRQoL: 76%, neurocognition: 58%), however, much less frequently outside the framework of clinical trials. In this context, CNOC affiliation, representing a specific health care quality assurance process, was associated with significantly stronger PRO assessment practices regarding psycho-oncological burden, independent of UH status (common odds ratio=5.0, p=0.03). Nevertheless, PRO/neurocognitive assessment practice was not consistent even across CNOC. The overall most commonly used PRO/neurocognitive assessment tools were the Distress Thermometer (for psycho-oncological burden; 64%), the EORTC QLQ-C30 combined with the EORTC QLQ-BN20 (for HRQoL; 52%) and the Mini-Mental Status Test (for neurocognition; 67%), followed by the Montreal Cognitive Assessment (MoCA; 33%). Accordingly, for routine clinical screening, the authors recommend the Distress Thermometer and the EORTC QLQ-C30 and QLQ-BN20, complemented by the MoCA as a comparatively sensitive yet basic neurocognitive test. This recommendation is intended to encourage more regular, adequate, and standardized routine assessments in neuro-oncological practice.

Machine learning for sperm selection

Infertility rates and the number of couples seeking fertility care have increased worldwide over the past few decades. Over 2.5 million cycles of assisted reproductive technologies are being performed globally every year, but the success rate has remained at ~33%. Machine learning, an automated method of data analysis based on patterns and inference, is increasingly being deployed within the health-care sector to improve diagnostics and therapeutics. This technique is already aiding embryo selection in some fertility clinics, and has also been applied in research laboratories to improve sperm analysis and selection. Tremendous opportunities exist for machine learning to advance male fertility treatments. The fundamental challenge of sperm selection - selecting the most promising candidate from 10⁸ gametes - presents a challenge that is uniquely well-suited to the high-throughput capabilities of machine learning algorithms paired with modern data processing capabilities.

Easily Created Prediction Model Using Automated Artificial Intelligence Framework (Prediction One, Sony Network Communications Inc., Tokyo, Japan) for Subarachnoid Hemorrhage Outcomes Treated by Coiling and Delayed Cerebral Ischemia

Introduction Reliable prediction models of subarachnoid hemorrhage (SAH) outcomes and delayed cerebral ischemia (DCI) are needed to decide the treatment strategy. Automated artificial intelligence (AutoAI) is attractive, but there are few reports on AutoAI-based models for SAH functional outcomes and DCI. We herein made models using an AutoAI framework, Prediction One (Sony Network Communications Inc., Tokyo, Japan), and compared it to other previous statistical prediction scores. Methods We used an open dataset of 298 SAH patients, who were with non-severe neurological grade and treated by coiling. Modified Rankin Scale 0-3 at six months was defined as a favorable functional outcome and DCI occurrence as another outcome. We randomly divided them into a 248-patient training dataset and a 50-patient test dataset. Prediction One made the model using training dataset with 5-fold cross-validation. We evaluated the model using the test dataset and compared the area under the curves (AUCs) of the created models. Those of the modified SAFIRE score and the Fisher computed tomography (CT) scale to predict the outcomes. Results The AUCs of the AutoAI-based models for functional outcome in the training and test dataset were 0.994 and 0.801, and those for the DCI occurrence were 0.969 and 0.650. AUCs for functional outcome calculated using modified SAFIRE score were 0.844 and 0.892. Those for the DCI occurrence calculated using the Fisher CT scale were 0.577 and 0.544. Conclusions We easily and quickly made AutoAI-based prediction models. The models' AUCs were not inferior to the previous prediction models despite the easiness.

Postsurgical functional outcome prediction model using deep learning framework (Prediction One, Sony Network Communications Inc.) for hypertensive intracerebral hemorrhage

BACKGROUND

Reliable prediction models of intracerebral hemorrhage (ICH) outcomes are needed for decision-making of the treatment. Statistically making such prediction models needs a large number of samples and time-consuming statistical analysis. Deep learning (DL), one of the artificial intelligence, is attractive, but there were no reports on DL-based functional outcome prediction models for ICH outcomes after surgery. We herein made a functional outcome prediction model using DLframework, Prediction One (Sony Network Communications Inc., Tokyo, Japan), and compared it to original ICH score, ICH Grading Scale, and FUNC score.

METHODS

We used 140 consecutive hypertensive ICH patients' data in our hospital between 2012 and 2019. All patients were surgically treated. Modified Rankin Scale 0-3 at 6 months was defined as a favorable outcome. We randomly divided them into 100 patients training dataset and 40 patients validation dataset. Prediction One made the prediction model using the training dataset with 5-fold cross-validation. We calculated area under the curves (AUCs) regarding the outcome using the DL-based model, ICH score, ICH Grading Scale, and FUNC score. The AUCs were compared.

RESULTS

The model made by Prediction One using 64 variables had AUC of 0.997 in the training dataset and that of 0.884 in the validation dataset. These AUCs were superior to those derived from ICH score, ICH Grading Scale, and FUNC score.

CONCLUSION

We easily and quickly made prediction models using Prediction One, even with a small single-center dataset. The accuracy of the DL-based model was superior to those of previous statistically calculated models.

Preliminary development of a deep learning-based automated primary headache diagnosis model using Japanese natural language processing of medical questionnaire

Background

Primary headaches are widespread and costly public health problems. However, there are insufficient medical resources for their treatment in Japan due to two reasons. First, the numbers of headache specialists and clinics remain insufficient. Second, neurologists and neurosurgeons mainly treat headaches in Japan. However, they mainly work as general stroke neurologists, so they cannot focus on primary headache treatment. To solve these problems, we preliminarily developed a deep learning (DL)-based automated diagnosis model from patients' Japanese unstructured sentences in the medical questionnaire using a DL framework. We hypothesized that the model would reduce the time and burden on both doctors and patients and improve their quality of life.

Methods

We retrospectively investigated our primary headache database and developed a diagnosis model using the DL framework (Prediction One, Sony Network Communications Inc., Japan). We used age, sex, date, and embedding layer made by the medical questionnaire's natural language processing (NLP).

Results

Eight hundred and forty-eight primary headache patients (495 women and 353 men) are included. The median (interquartile range) age was 59 (40-74). Migraine accounted for 46%, tension-type headache for 47%, trigeminal autonomic cephalalgias for 5%, and other primary headache disorders for 2%. The accuracy, mean precision, mean recall, and mean F value of the developed diagnosis model were 0.7759, 0.8537, 0.6086, and 0.6353, which were satisfactory.

Conclusion

The DL-based diagnosis model for primary headaches using the raw medical questionnaire's Japanese NLP would be useful in performing efficient medical practice after ruling out the secondary headaches.

Easily created prediction model using deep learning software (Prediction One, Sony Network Communications Inc.) for subarachnoid hemorrhage outcomes from small dataset at admission

BACKGROUND

Reliable prediction models of subarachnoid hemorrhage (SAH) outcomes are needed for decision-making of the treatment. SAFIRE score using only four variables is a good prediction scoring system. However, making such prediction models needs a large number of samples and time-consuming statistical analysis. Deep learning (DL), one of the artificial intelligence, is attractive, but there were no reports on prediction models for SAH outcomes using DL. We herein made a prediction model using DL software, Prediction One (Sony Network Communications Inc., Tokyo, Japan) and compared it to SAFIRE score.

METHODS

We used 153 consecutive aneurysmal SAH patients data in our hospital between 2012 and 2019. Modified Rankin Scale (mRS) 0-3 at 6 months was defined as a favorable outcome. We randomly divided them into 102 patients training dataset and 51 patients external validation dataset. Prediction one made the prediction model using the training dataset with internal cross-validation. We used both the created model and SAFIRE score to predict the outcomes using the external validation set. The areas under the curve (AUCs) were compared.

RESULTS

The model made by Prediction One using 28 variables had AUC of 0.848, and its AUC for the validation dataset was 0.953 (95%CI 0.900-1.000). AUCs calculated using SAFIRE score were 0.875 for the training dataset and 0.960 for the validation dataset, respectively.

CONCLUSION

We easily and quickly made prediction models using Prediction One, even with a small single-center dataset. The accuracy of the model was not so inferior to those of previous statistically calculated prediction models.