Machine learning in neurosurgery [Neurosurgery Wiki]

This page is read only. You can view the source, but not change it. Ask your administrator if you think this is wrong.
====== Machine learning in neurosurgery ======

{{rss>https://pubmed.ncbi.nlm.nih.gov/rss/search/1RSu50XbTiN15Mbi14hM7qK3zh3HwMAfbe6z6B1fgm5L8zm5Ln/?limit=15&utm_campaign=pubmed-2&fc=20230926075428}}
----
Machine learning (ML) in neurosurgery is an evolving field with significant potential to improve diagnosis, treatment planning, intraoperative decision-making, and patient outcomes. Here are some key areas where ML is making an impact:

===== Preoperative Diagnosis and Risk Stratification =====

   - **Image Analysis & Classification**: ML algorithms, particularly deep learning models like convolutional neural networks (CNNs), are improving the detection of brain tumors, aneurysms, and spinal pathologies in MRI and CT scans.
   - **Radiomics & Predictive Modeling**: ML can extract quantitative features from medical images to predict tumor malignancy, recurrence risk, and response to treatment.
   - **Natural Language Processing (NLP)**: AI can analyze electronic health records (EHRs) to identify risk factors and classify neurosurgical conditions.

===== Surgical Planning & Navigation =====

   - **3D Reconstruction & Augmented Reality**: ML assists in preoperative planning by generating 3D models from imaging data, helping surgeons visualize complex anatomical structures.
   - **Robotics & AI-assisted Surgery**: Machine learning enhances precision in robotic-assisted neurosurgery (e.g., the ROSA and Mazor X systems for stereotactic procedures).
   - **Trajectory Optimization**: Algorithms optimize surgical pathways, reducing the risk of damage to critical structures.

===== Intraoperative Applications =====

   - **AI-guided Microsurgery**: ML assists in real-time tissue differentiation (e.g., distinguishing between tumor and normal brain tissue).
   - **Neurophysiological Monitoring**: AI enhances intraoperative electrophysiological monitoring, helping to predict neurological outcomes.
   - **Error Detection & Prevention**: Machine learning models analyze surgeon movements and alert in case of deviations from optimal technique.

===== Postoperative Care & Outcome Prediction =====

   - **Complication Prediction**: ML models predict post-surgical complications such as infections, hydrocephalus, or stroke based on patient data.
   - **Rehabilitation & Recovery Optimization**: AI-driven models personalize rehabilitation programs for patients recovering from neurosurgical procedures.

===== Research & Training =====

   - **Automated Literature Review**: AI tools summarize the latest research papers, accelerating the discovery of new neurosurgical techniques.
   - **Surgical Simulation**: ML-powered virtual reality (VR) simulators enhance surgical training by providing personalized feedback on performance.

### **Challenges & Future Directions**
   - **Data Quality & Bias**: ML models require high-quality, diverse datasets to generalize well.
   - **Regulatory & Ethical Issues**: AI-based decision-making must be transparent and validated for clinical use.
   - **Integration with Existing Systems**: Effective adoption requires integration with hospital information systems and surgical workflows.

Machine learning is revolutionizing neurosurgery, from preoperative planning to intraoperative assistance and postoperative care. As technology advances, AI-driven tools will continue to enhance surgical precision, improve patient outcomes, and transform the field of neurosurgery.

Would you like a deeper dive into a specific application or a practical example of how ML is currently used in neurosurgery?
----
A study aimed to summarize the current applications of ML in the analysis and assessment of neurosurgical skills. We conducted this systematic review in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. We searched the PubMed and Google Scholar databases for eligible studies published until November 15, 2022, and used the Medical Education Research Study Quality Instrument (MERSQI) to assess the quality of the included articles. Of the 261 studies identified, we included 17 in the final analysis. Studies were most commonly related to oncological, spinal, and vascular neurosurgery using microsurgical and endoscopic techniques. Machine learning-evaluated tasks included subpial brain tumor resection, anterior cervical discectomy and fusion, hemostasis of the lacerated internal carotid artery, brain vessel dissection and suturing, glove microsuturing, lumbar hemilaminectomy, and bone drilling. The data sources included files extracted from VR simulators and microscopic and endoscopic videos. The ML application was aimed at classifying participants into several expertise levels, analysis of differences between experts and novices, surgical instrument recognition, division of operation into phases, and prediction of blood loss. In two articles, ML models were compared with those of human experts. The machines outperformed humans in all tasks. The most popular algorithms used to classify surgeons by skill level were the support vector machine and k-nearest neighbors, and their accuracy exceeded 90%. The "you only look once" detector and RetinaNet usually solved the problem of detecting surgical instruments - their accuracy was approximately 70%. The experts differed by more confident contact with tissues, higher bimanual, smaller distance between the instrument tips, and relaxed and focused state of mind. The average MERSQI score was 13.9 (from 18). There is growing interest in the use of ML in neurosurgical training. Most studies have focused on the evaluation of microsurgical skills in oncological neurosurgery and on the use of virtual simulators; however, other subspecialties, skills, and simulators are being investigated. Machine learning models effectively solve different neurosurgical tasks related to skill classification, object detection, and outcome prediction. Properly trained ML models outperform human efficacy. Further research on ML application in neurosurgery is needed

Conclusion: The study found that ML is becoming increasingly important in neurosurgical training. While most studies focused on microsurgery for brain tumors, researchers are also looking into other types of surgeries and simulators. Machine Learning models are proving to be very effective in tasks related to neurosurgery skills, instrument recognition, and predicting outcomes. In fact, properly trained ML models performed better than humans.
((Titov O, Bykanov A, Pitskhelauri D. Neurosurgical skills analysis by machine learning models: systematic review. Neurosurg Rev. 2023 May 16;46(1):121. doi: 10.1007/s10143-023-02028-x. PMID: 37191734.)).
----


[[Machine learning]] applications have been reviewed in [[neurosurgery]]
((Senders JT, Arnaout O, Karhade AV, Dasenbrock HH, Gormley
WB, Broekman ML et al (2017) Natural and artificial intelligence
in neurosurgery: a systematic review. Neurosurgery 83(2):181–192))

see [[Machine learning for degenerative cervical myelopathy]].


----
Machine learning (ML) involves [[algorithm]]s learning patterns in large, complex datasets to predict and classify. Algorithms include [[neural network]]s (NN), [[logistic regression]] (LR), and [[support-vector machine]]s (SVM). ML may generate substantial improvements in neurosurgery. This systematic review assessed the current state of neurosurgical ML applications and the performance of algorithms applied. Our systematic search strategy yielded 6866 results, 70 of which met inclusion criteria. Performance statistics analyzed included area under the receiver operating characteristics curve (AUC), accuracy, sensitivity, and specificity. Natural language processing (NLP) was used to model topics across the corpus and to identify keywords within surgical subspecialties. ML applications were heterogeneous. The densest cluster of studies focused on preoperative evaluation, planning, and outcome prediction in spine surgery. The main algorithms applied were NN, LR, and SVM. Input and output features varied widely and were listed to facilitate future research. The accuracy (F(2,19) = 6.56, p < 0.01) and specificity (F(2,16) = 5.57, p < 0.01) of NN, LR, and SVM differed significantly. NN algorithms demonstrated significantly higher accuracy than LR. SVM demonstrated significantly higher specificity than LR. We found no significant difference between NN, LR, and SVM AUC and sensitivity. NLP topic modeling reached maximum coherence at seven topics, which were defined by modeling approach, surgery type, and pathology themes. Keywords captured research foci within surgical domains. ML technology accurately predicts outcomes and facilitates clinical decision-making in neurosurgery. NNs frequently outperformed other algorithms on supervised learning tasks. This study identified gaps in the literature and opportunities for future neurosurgical ML research
((Buchlak QD, Esmaili N, Leveque JC, Farrokhi F, Bennett C, Piccardi M, Sethi
RK. Machine learning applications to clinical decision support in neurosurgery:
an artificial intelligence augmented systematic review. Neurosurg Rev. 2019 Aug
17. doi: 10.1007/s10143-019-01163-8. [Epub ahead of print] Review. PubMed PMID:
31422572.
)).
----
A [[study]] implemented a supervised machine learning-based approach in modeling estimated symptom resolve time in high school athletes who incurred a concussion during sport activity.

They examined the efficacy of 10 classification algorithms using machine learning for prediction of symptom resolution time (within seven, fourteen, or twenty-eight days), with a dataset representing three years of concussions suffered by high school student-athletes in football (most concussion incidents) and other contact sports.

The most prevalent sport-related concussion reported symptom was headache (94.9%), followed by dizziness (74.3%) and difficulty concentrating (61.1%). For all three category thresholds of predicted symptom resolution time, single-factor ANOVAs revealed statistically significant performance differences across the ten classification models for all learners at a 95% confidence level (P=0.000). Naïve Bayes and Random Forest with either 100 or 500 trees were the top-performing learners with an area under the ROC curve performance ranging between 0.666 and 0.742 (0.0-1.0 scale).

Considering the limitations of these data specific to symptom presentation and resolve, supervised machine learning demonstrated efficacy, while warranting further exploration, in developing symptom-based prediction models for practical estimation of sport-related concussion recovery in enhancing clinical decision support
((Bergeron MF, Landset S, Maugans TA, Williams VB, Collins CL, Wasserman EB,
Khoshgoftaar TM. Machine Learning in Modeling High School Sport Concussion
Symptom Resolve. Med Sci Sports Exerc. 2019 Jan 25. doi:
10.1249/MSS.0000000000001903. [Epub ahead of print] PubMed PMID: 30694980.
)).



----
Current [[practice]] of [[neurosurgery]] depends on clinical practice [[guideline]]s and [[evidence based]] 
 [[research]] [[publication]]s that derive results using [[statistical]] methods. However, [[statistical analysis]] methods have some limitations such as the inability to analyze [[nonlinear]] [[variable]]s, requiring setting a level of [[significance]], being impractical for analyzing large amounts of [[data]] and the possibility of human bias. [[Machine learning]] is an emerging method for analyzing massive amounts of complex [[data]] which relies on [[algorithm]]s that allow [[computer]]s to learn and make accurate [[prediction]]s. 

Machine learning has been increasingly implemented in medical [[research]] as well as neurosurgical publications. A systematical review aimed to assemble the current neurosurgical literature that machine learning has been utilized, and to inform neurosurgeons on this novel method of data analysis
((Celtikci E. A Systematic Review on Machine Learning in Neurosurgery: The
Future of Decision-Making in Patient Care. Turk Neurosurg. 2018;28(2):167-173.
doi: 10.5137/1019-5149.JTN.20059-17.1. Review. PubMed PMID: 28481395.
))

ML is increasingly tested in neurosurgical applications and even demonstrated to emulate the performance of clinical [[expert]]s
((Mariak Z, Swiercz M, Krejza J, Lewko J, Lyson T. Intracranial pressure processing with artificial neural networks: classification of signal properties. Acta Neurochir (Wien) . 2000;142(4):407-411; discussion 411-402.))
((Nucci CG, De Bonis P, Mangiola A et al.   Intracranial pressure wave morphological classification: automated analysis and clinical validation. Acta Neurochir (Wien) . 2016;158(3):581-588; discussion 588.))
((Sieben G, Praet M, Roels H, Otte G, Boullart L, Calliauw L. The development of a decision support system for the pathological diagnosis of human cerebral tumours based on a neural network classifier. Acta Neurochir (Wien) . 1994;129(3-4):193-197.))
((Mathew B, Norris D, Mackintosh I, Waddell G. Artificial intelligence in the prediction of operative findings in low back surgery. Brit J Neurosurg . 1989;3(2):161-170.))
((Arle JE, Perrine K, Devinsky O, Doyle WK. Neural network analysis of preoperative variables and outcome in epilepsy surgery. J Neurosurg . 1999;90(6):998-1004.))
((Gazit T, Andelman F, Glikmann-Johnston Y et al.   Probabilistic machine learning for the evaluation of presurgical language dominance. J Neurosurg . 2016;125(2):1-13.))
((Shi HY, Hwang SL, Lee KT, Lin CL. In-hospital mortality after traumatic brain injury surgery: a nationwide population-based comparison of mortality predictors used in artificial neural network and logistic regression models. J Neurosurg . 2013;118(4):746-752.))
((Azimi P, Mohammadi HR. Predicting endoscopic third ventriculostomy success in childhood hydrocephalus: an artificial neural network analysis. J Neurosurg Pediatr . 2014;13(4):426-432.))
((Azimi P, Mohammadi H. Prediction of successful ETV outcome in childhood hydrocephalus: an artificial neural networks analysis. J Neurosurg . 2015;122(6):426-432.))
((Chang K, Zhang B, Guo X et al.   Multimodal imaging patterns predict survival in recurrent glioblastoma patients treated with bevacizumab. Neuro-oncology . 2016;18(12):1680-1687.))
((Jones TL, Byrnes TJ, Yang G, Howe FA, Bell BA, Barrick TR. Brain tumor classification using the diffusion tensor image segmentation (D-SEG) technique. Neuro-oncology . 2015;17(3):466-476.))
((Macyszyn L, Akbari H, Pisapia JM et al.   Imaging patterns predict patient survival and molecular subtype in glioblastoma via machine learning techniques. Neuro-oncology . 2016;18(3):417-425.))
((Teplyuk NM, Mollenhauer B, Gabriely G et al.   MicroRNAs in cerebrospinal fluid identify glioblastoma and metastatic brain cancers and reflect disease activity. Neuro-oncology . 2012;14(6):689-700.))
((Zhang B, Chang K, Ramkissoon S et al.   Multimodal MRI features predict isocitrate dehydrogenase genotype in high-grade gliomas. Neuro-oncology . 2017;19(1):109-117.))
((Fouke SJ, Weinberger K, Kelsey M et al.   A machine-learning-based classifier for predicting a multi-parametric probability map of active tumor extent within glioblastoma multiforme. Neuro-oncology . 2012;14:vi124-vi125.))
((Kim LM, Commean P, Boyd A et al.   Predicting the location and probability of viable tumor within glioblastoma multiforme with multiparametric magnetic resonance imaging. Neuro-oncology . 2012;14:vi120-vi128.))
((Orphanidou-Vlachou E, Vlachos N, Davies N, Arvanitis T, Grundy R, Peet A. Texture analysis of T1-and t2-weighted magnetic resonance images to discriminate posterior fossa tumors in children. Neuro-oncology . 2014;16:i123-i126.))
((Rayfield C, Swanson K. Predicting the response to treatment in GBM: Machine learning on clinical images. Neuro-oncology . 2015;17:v167.))
((Akbari H, Macyszyn L, Da X et al.   Imaging surrogates of infiltration obtained via multiparametric imaging pattern analysis predict subsequent location of recurrence of glioblastoma. Neurosurgery . 2016;78(4):572-580.))
((Mitchell TJ, Hacker CD, Breshears JD et al.   A novel data-driven approach to preoperative mapping of functional cortex using resting-state functional magnetic resonance imaging. Neurosurgery . 2013;73(6):969-982; discussion 982-963.))
((Oermann EK, Kress MA, Collins BT et al.   Predicting survival in patients with brain metastases treated with radiosurgery using artificial neural networks. Neurosurgery . 2013;72(6):944-951; discussion 952.))
((Taghva A. An automated navigation system for deep brain stimulator placement using hidden Markov models. Neurosurgery . 2010;66(3 Suppl Operative):108-117; discussion 117.))
((Dumont TM, Rughani AI, Tranmer BI. Prediction of symptomatic cerebral vasospasm after aneurysmal subarachnoid hemorrhage with an artificial neural network: feasibility and comparison with logistic regression models. World Neurosurg . 2011;75(1):57-63; discussion 25-58.)).
----
Automated analysis of radiological data for diagnosis, segmentation, or outcome prediction could, be one of the first ML applications that finds its way to actual clinical practice
((Obermeyer Z, Emanuel EJ. Predicting the future - big data, machine learning, and clinical medicine. N Engl J Med . 2016;375(13):1216-1219.)).
----
Current [[outcome prediction]] are largely based on and limited by [[regression]] methods. Utilization of [[machine learning]] (ML) methods that can handle multiple diverse inputs could strengthen predictive abilities and improve patient outcomes. [[Inpatient]] [[length of stay]] (LOS) is one such outcome that serves as a surrogate for patient [[disease]] severity and resource utilization.

To develop a novel method to systematically rank, select, and combine ML algorithms to build a model that predicts LOS following [[craniotomy]] for [[brain tumor]].

A training [[dataset]] of 41 222 patients who underwent craniotomy for brain tumor was created from the [[National Inpatient Sample]]. Twenty-nine ML algorithms were trained on 26 preoperative variables to predict LOS. Trained algorithms were ranked by calculating the root mean square logarithmic error (RMSLE) and top performing algorithms combined to form an ensemble. The ensemble was externally validated using a dataset of 4592 patients from the [[National Surgical Quality Improvement Program]]. Additional analyses identified variables that most strongly influence the ensemble model predictions.

The ensemble model predicted LOS with RMSLE of .555 (95% confidence interval, .553-.557) on internal validation and .631 on external validation. Nonelective surgery, preoperative pneumonia, sodium abnormality, or weight loss, and non-White race were the strongest predictors of increased LOS.

An ML ensemble model predicts LOS with good performance on internal and external validation, and yields clinical insights that may potentially improve patient outcomes. This systematic ML method can be applied to a broad range of clinical problems to improve patient care
((Muhlestein WE, Akagi DS, Davies JM, Chambless LB. Predicting Inpatient Length 
of Stay After Brain Tumor Surgery: Developing [[Machine Learning]] Ensembles to
Improve Predictive Performance. Neurosurgery. 2018 Aug 3. doi:
10.1093/neuros/nyy343. [Epub ahead of print] PubMed PMID: 30113665.
)).
----

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.

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.

Senders et al., 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
((Senders JT, Arnaout O, Karhade AV, Dasenbrock HH, Gormley WB, Broekman ML,
Smith TR. Natural and Artificial Intelligence in Neurosurgery: A Systematic
Review. Neurosurgery. 2018 Aug 1;83(2):181-192. doi: 10.1093/neuros/nyx384.
PubMed PMID: 28945910.
)).
----


Lazaridis et al., and others, have developed [[predictive model]]s based on [[machine learning]] from continuous time series of [[intracranial pressure]] and [[partial pressure of brain tissue oxygen]]. These models provide accurate predictions of physiologic crises events in a timely fashion, offering the opportunity for an earlier application of targeted interventions.They review the rationale for prediction, discuss available predictive models with examples, and offer suggestions for their future prospective testing in conjunction with preventive clinical [[algorithm]]s
((Lazaridis C, Rusin CG, Robertson CS. Secondary Brain Injury: Predicting and
Preventing Insults. Neuropharmacology. 2018 Jun 6. pii: S0028-3908(18)30279-X.
doi: 10.1016/j.neuropharm.2018.06.005. [Epub ahead of print] Review. PubMed PMID:
29885419.
)).
----

Machine [[learning]] (ML) is a domain of [[artificial intelligence]] that allows computer [[algorithm]]s to learn from experience without being explicitly programmed.

To summarize neurosurgical applications of ML where it has been compared to clinical expertise, here referred to as "natural intelligence."

Two important and rapidly developing scientific movements—data [[reproducibility]] and [[machine learning]]—are central to a recent Neuron paper by Chung et al
((Chung JE, Magland JF, Barnett AH, Tolosa VM, Tooker AC, Lee KY, Shah KG, Felix
SH, Frank LM, Greengard LF. A Fully Automated Approach to Spike Sorting. Neuron. 
2017 Sep 13;95(6):1381-1394.e6. doi: 10.1016/j.neuron.2017.08.030. PubMed PMID:
28910621; PubMed Central PMCID: PMC5743236.
))
----


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.

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.


They 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
((Senders JT, Arnaout O, Karhade AV, Dasenbrock HH, Gormley WB, Broekman ML,
Smith TR. Natural and Artificial Intelligence in Neurosurgery: A Systematic
Review. Neurosurgery. 2018 Aug 1;83(2):181-192. doi: 10.1093/neuros/nyx384.
PubMed PMID: 28945910.
)).
----
Yepes-Calderon et al. presented a [[segmentation]] strategy based on an algorithm that uses four features extracted from the medical images to create a statistical estimator capable of determining ventricular volume. When compared with manual segmentations, the correlation was 94% and holds promise for even better accuracy by incorporating the unlimited data available. The volume of any segmentable structure can be accurately determined utilizing the [[machine learning]] strategy presented and runs fully automatically within the PACS
((Yepes-Calderon F, Nelson MD, McComb JG. Automatically measuring brain
ventricular volume within PACS using artificial intelligence. PLoS One. 2018 Mar 
15;13(3):e0193152. doi: 10.1371/journal.pone.0193152. eCollection 2018. PubMed
PMID: 29543817.
)).

===== Machine learning for Aneurysmal subarachnoid hemorrhage outcome prediction =====
[[Machine learning for Aneurysmal subarachnoid hemorrhage outcome prediction]]

===== References =====