Simulated Brain Tumor Research Articles

428 Artificial Intelligence Training Versus In-person Expert Training in Teaching Simulated Tumor Resection Skills - A Cross-Over Randomized Controlled Trial

INTRODUCTION: Neurosurgical simulations equipped with artificial intelligence systems provide an objective quantitative assessment of surgical technical skills and tailored intelligent feedback. These systems allow for repeated practice of complex skills such as subpial brain tumor resection. However, more work is needed to assess the utility of intelligent systems in teaching tumor resection skills, compared to traditional human instruction. METHODS: Twenty-five trainees who are currently enrolled in four Canadian medical schools participated in two training sessions, during which they completed a simulated subpial tumor resection five times. Participants were randomly assigned to two feedback groups: (1) real-time intelligent instruction, and (2) in-person human instruction. They were then assigned to the other feedback group in the second (cross-over) session. A composite-score was given by the intelligent system to assess technical skills during each task. RESULTS: Trainees who received real-time intelligent instruction significantly improved their composite-score in the first and second training sessions (p = .017, p = .005, respectively). Trainees who received in-person human instruction in the first training session had no statistically significant changes in their composite-score from the first to the last task repetition (p = .119). The composite-score decreased significantly in the second training session with in-person human instruction (p = .004). CONCLUSIONS: Real-time intelligent instruction provides an effective way of teaching simulated brain tumor resection skills. Artificial intelligence systems may be a useful addition to current surgical training curricula, providing objective, and tailored assessment and teaching of surgical technical skills in risk-free realistically simulated patient cases. Educators can benefit from these systems to guide their supervision, while trainees can use them for self-guided skill acquisition. Further research may assess the generalizability of these findings to other surgical procedures.

10048-CO-2 CAN NATURAL LANGUAGE PROCESSING AI BE AN APPROPRIATE “ADVISER” FOR BRAIN TUMOR PATIENTS? CLINICAL APPLICATION AND ISSUES ON CHATGPT

Abstract INTRODUCTION Brain tumors affect approximately 20,000 individuals annually in Japan, with some cases requiring urgent treatment. However, the medical knowledge of non-medical individuals (the general public) is often limited, resulting in delays in seeking treatment even when neurological symptoms are evident. Having appropriate advisors who can encourage medical consultations could potentially improve this situation. While Natural Language Processing AI (NLP-AI) is widely used, its full integration into healthcare remains a work in progress. ChatGPT, a prominent NLP-AI, possesses extensive knowledge and is accessible 24/7 online. We explored the potential of leveraging its advanced capabilities and accessibility in the field of brain tumor management. METHODS To evaluate its ability to guide hospital visits effectively, we simulated various symptoms associated with primary brain tumors and consulted ChatGPT(GPT3.5) as patients. We also assessed how ChatGPT handled unnecessary noise information (unrelated symptoms) during symptom consultations. RESULTS ChatGPT consistently recommended hospital visits for all simulated brain tumor patients and provided appropriate advice regarding the urgency and relevant medical department. Notably, even when patients expressed hesitation and refusal hospital visits, ChatGPT persistently persuaded them to visit the hospital despite the user(patient)'s intention. Despite the introduction of noise information, ChatGPT accurately identified and emphasized the importance of relevant symptoms, prompting appropriate medical attention. However, accuracy declined when excessive noise was present. Additionally, ChatGPT generated concise summaries of medical histories for presentation during hospital visits. DISCUSSION While AI is widely utilized in medical imaging, the comprehensive integration of conversational AI in healthcare is still in progress. ChatGPT, despite not being a healthcare professional, holds potential as an "advisor" with extensive medical knowledge, accessible for consultations at any time, to support brain tumor patients. However, challenges related to accountability and expanding knowledge capabilities remain to be addressed.

SP11.14 Evaluating the Efficacy and Cognitive Load of a Real-Time Intelligent Coaching System versus Human Expert Instruction in Surgical Technical Skills Training: A Randomized Controlled Study

Abstract Aims An artificial intelligence-driven tutoring system provides risk-free training on realistic simulated procedures and objective assessment to improve surgical technical skills. This study compared its teaching efficacy to human instructor training. Methods Ninety-eight medical students completed 5 simulated brain tumor resections, scored by the intelligent system. Participants were randomly sorted into 3 feedback groups: (1) no real-time feedback, (2) real-time intelligent instruction, (3) real-time human instruction. Each participant completed a base-line repetition without feedback. Group 1 received visual feedback based on expert benchmarks following each repetition. Group 2 was tutored by the intelligent system in real-time. After each repetition, the system provided the student’s error clips and correlating expert-level demonstrations. Group 3 was tutored by human instructors in real-time. After each repetition, instructors provided critiques and demonstrated correction techniques. Following the final repetition, each participant completed a questionnaire of Leppink’s Cognitive Load Index on a 5-point Likert scale. Performance scores were compared from the first to the last repetition of the task within and between-groups to assess learning. Results Compared with baseline, groups 2 and 3 significantly improved in performance by the third and second repetition, respectively (p<0.01 and p=0.01). Between-groups comparison revealed significantly higher scores and extrinsic load for group 2 than group 3 in the fifth repetition, respectively (p<0.01 and p<0.01). No between-group differences existed for intrinsic and germane load. Conclusion The intelligent system provided more efficient learning than human instruction. Such systems may enhance trainees’ learning and facilitate the shift towards standardized, competency-based surgical training.

COMPARING THE EFFICIENCY OF A REAL-TIME ARTIFICIAL INTELLIGENCE INSTRUCTOR TO HUMAN EXPERT INSTRUCTORS IN SIMULATED SURGICAL TECHNICAL SKILLS TRAINING– A RANDOMIZED CONTROLLED TRIAL

Abstract BTFC travel award recipient Artificial intelligence systems provide risk-free training on realistically simulated patient cases and objective assessment of surgical technical skills. This randomized controlled study compared a real-time intelligent tutoring system in technical skills learning with human expert instructor-mediated training. METHODS: Ninety-eight medical students performed six simulated brain tumor resections. Participants were randomly allocated into (1)no-real-time feedback, (2)real-time intelligent instruction, and (3)in- person human instruction. All students performed the first repetition without receiving feedback (baseline). Group-1 received visual feedback only after each procedure based on expert benchmarks. Group-2 was instructed by the intelligent system in real-time. After each task, the students were shown their error-video clips generated by this system alongside expert-level demonstrations on how to improve. Group-3 was instructed by human instructors during the tasks. After each task, instructors summarized the areas of improvement and demonstrated correction techniques. Participant performance was scored by the intelligent system and also by blinded experts using OSATS scores. The performance score was compared within groups and between groups to compare learning. RESULTS: Compared to baseline performance, Group-2 and Group-3 significantly improved in the performance score by the third and second repetition, respectively (p<0.01, p=0.01). The between-groups comparison demonstrated that Group-2 scored significantly higher than Group-3 in the fifth repetition (p<0.01). Group-2 achieved significantly higher OSATS scores than Group-1 in the sixth task. CONCLUSIONS: Artificial intelligence may facilitate trainee learning by providing equally or more efficient learning when compared to human instruction. These systems may aid in developing competency-based standardized curricula in surgical training.

Modeling fibrous soft tissue dissection with elastic-plastic deformation for simulation of brain tumor removal

Background and objective: Realistic modeling the dissection of brain tissue is of key importance for simulation of brain tumor removal in virtual neurosurgery systems. However, existing methods are unable to characterize inelastic behaviors of brain tissue, such as plastic deformation and dissection evolution, making it ineffective in simulating brain tumor removal procedures.Methods: In this paper, a model of fibrous soft tissue dissection for the simulation of brain tumor removal is proposed. A dissection variable of representative volume element is used to characterize the dissection state of the fibrous soft tissue. The evolution of dissection with elastic-plastic deformation under the effects of external loads is presented.Results: Simulation results show that the proposed model provides realistic, stable and intuitive results in the simulation of fracture in fibrous soft tissues. As the external load increases, the fibrous soft tissue begins to crack, with the cracks growing and multiplying until they eventually merge to form a fracture. The proposed model is incorporated into the simulation of brain tumor removal.Conclusions: The experimental results demonstrate the feasibility of modeling fibrous soft tissue dissection with elastic-plastic deformation. A relative high degree of realistic visual feedback is achieved.

Utilizing artificial intelligence and electroencephalography to assess expertise on a simulated neurosurgical task

Virtual reality surgical simulators have facilitated surgical education by providing a safe training environment. Electroencephalography (EEG) has been employed to assess neuroelectric activity during surgical performance. Machine learning (ML) has been applied to analyze EEG data split into frequency bands. Although EEG is widely used in fields requiring expert performance, it has yet been used to classify surgical expertise. Thus, the goals of this study were to (a) develop an ML model to accurately differentiate skilled and less-skilled performance using EEG data recorded during a simulated surgery, (b) explore the relative importance of each EEG bandwidth to expertise, and (c) analyze differences in EEG band powers between skilled and less-skilled individuals. We hypothesized that EEG recordings during a virtual reality surgery task would accurately predict the expertise level of the participant. Twenty-one participants performed three simulated brain tumor resection procedures on the NeuroVR™ platform (CAE Healthcare, Montreal, Canada) while EEG data was recorded. Participants were divided into 2 groups. The skilled group was composed of five neurosurgeons and five senior neurosurgical residents (PGY4-6), and the less-skilled group was composed of six junior residents (PGY1-3) and five medical students. A total of 13 metrics from EEG frequency bands and ratios (e.g., alpha, theta/beta ratio) were generated. Seven ML model types were trained using EEG activity to differentiate between skilled and less-skilled groups. The artificial neural network achieved the highest testing accuracy of 100% (AUROC = 1.0). Model interpretation via Shapley analysis identified low alpha (8–10 Hz) as the most important metric for classifying expertise. Skilled surgeons displayed higher (p = 0.044) low-alpha than the less-skilled group. Furthermore, skilled surgeons displayed significantly lower TBR (p = 0.048) and significantly higher beta (13–30 Hz, p = 0.049), beta 1 (15–18 Hz, p = 0.014), and beta 2 (19–22 Hz, p = 0.015), thus establishing these metrics as important markers of expertise. ACGME Core CompetenciesPractice-Based Learning and Improvement.

Neutron Radiation Dose Measurements in a Scanning Proton Therapy Room: Can Parents Remain Near Their Children During Treatment?

PurposeThis study aims to characterize the neutron radiation field inside a scanning proton therapy treatment room including the impact of different pediatric patient sizes.Materials and MethodsWorking Group 9 of the European Radiation Dosimetry Group (EURADOS) has performed a comprehensive measurement campaign to measure neutron ambient dose equivalent, H*(10), at eight different positions around 1-, 5-, and 10-year-old pediatric anthropomorphic phantoms irradiated with a simulated brain tumor treatment. Several active detector systems were used.ResultsThe neutron dose mapping within the gantry room showed that H*(10) values significantly decreased with distance and angular deviation with respect to the beam axis. A maximum value of about 19.5 µSv/Gy was measured along the beam axis at 1 m from the isocenter for a 10-year-old pediatric phantom at 270° gantry angle. A minimum value of 0.1 µSv/Gy was measured at a distance of 2.25 m perpendicular to the beam axis for a 1-year-old pediatric phantom at 140° gantry angle.The H*(10) dependence on the size of the pediatric patient was observed. At 270° gantry position, the measured neutron H*(10) values for the 10-year-old pediatric phantom were up to 20% higher than those measured for the 5-year-old and up to 410% higher than for the 1-year-old phantom, respectively.ConclusionsUsing active neutron detectors, secondary neutron mapping was performed to characterize the neutron field generated during proton therapy of pediatric patients. It is shown that the neutron ambient dose equivalent H*(10) significantly decreases with distance and angle with respect to the beam axis. It is reported that the total neutron exposure of a person staying at a position perpendicular to the beam axis at a distance greater than 2 m from the isocenter remains well below the dose limit of 1 mSv per year for the general public (recommended by the International Commission on Radiological Protection) during the entire treatment course with a target dose of up to 60 Gy. This comprehensive analysis is key for general neutron shielding issues, for example, the safe operation of anesthetic equipment. However, it also enables the evaluation of whether it is safe for parents to remain near their children during treatment to bring them comfort. Currently, radiation protection protocols prohibit the occupancy of the treatment room during beam delivery.

510 Artificial Intelligence Tutoring Compared with Expert Instruction in Neurosurgical Simulation Training: A Randomized Controlled Trial

INTRODUCTION: Resection of subpial tumors adjacent to critical brain areas is one of the most high-stakes surgical procedures that lacks opportunities for safe deliberate practice. Furthermore, performance-based assessment in neurosurgical apprenticeship is inefficient and vulnerable to subjectivity. Utilizing Artificial Intelligence (AI) to classify neurosurgical psychomotor expertise, we developed the first AI-powered tutor in neurosurgical simulation training, known as the Virtual Operative Assistant (VOA), to augment technical skills acquisition training by providing intelligent feedback. METHODS: A multi-institutional randomized controlled trial compared VOA’s automated audiovisual metric-based feedback vs remote verbal debriefing with expert instruction and no-feedback controls. Medical students performed six simulated subpial brain tumor resections: five practice attempts followed with feedback and one complex realistic attempt evaluated skill retention and transfer. A deep learning model, Intelligent Continuous Expertise Monitoring System (ICEMS Expertise Score), and blinded Objective Structured Assessment of Technical Skills (OSATS) evaluated performance. Participants reported emotions before, during and after training and completed a cognitive load questionnaire following training. RESULTS: Seventy medical students from four institutions were randomly assigned to VOA (n=23), Instructor (n=24), and Control (n=23) Groups. 350 practice attempts were assessed by ICEMS, and 70 realistic attempts were evaluated by ICEMS and OSATS. During practice, VOA training resulted in a significant improvement of participants’ Expertise Scores that was on average 0.66 (95% CI 0.55-0.77) and 0.65 (95% CI 0.54-0.77) points higher than Instructor and Control Groups (p<.001). Realistic attempt’s average Expertise Score was significantly higher in the VOA Group compared to Instructor and Control Groups (mean difference 0.53; 0.49, respectively, p<.001). VOA and Instructor Group’s realistic attempt OSATS ratings were not significantly different. OSATS ratings demonstrated that VOA feedback resulted in significantly higher Respect for Tissue and Economy of Movement compared to Control while expert instruction significantly improved Instrument Handling compared to Control. There was no significant between-groups difference in cognitive load, positive-, and negative-activating emotions. CONCLUSION: VOA’s quantitative automated benchmark feedback demonstrated superior performance outcome, improved skill transfer, with equivalent OSATS ratings and similar cognitive and affective responses compared to remote expert instruction.

MORPHOLOGICAL SPECTRUM OF INTRACRANIAL SPACE OCCUPYING LESION AT TERTIARY CARE HOSPITAL-A CLINICOPATHOLOGICAL STUDY

Introduction: The annual incidence of CNS tumors ranges from 10 to 17 per 1,00,000 persons for intracranial tumors, [1] the majority of these are primary tumors. An “Intra-cranial space occupying lesion” (ICSOL)is dened as a mass lesion in the cranial cavity with a diverse etiology like benign or malignant neoplasm, inammatory or parasitic lesion, hematoma, or arterio-venous malformation.[3] Many non-neoplastic CNS lesions can clinically & radiologically simulate brain tumors. In such cases, histopathological [5] examination (HPE) can be helpful in differentiating between neoplastic and non-neoplastic etiologies. Aims and Objectives: This studywas undertaken to analyse theincidence and frequency of intracranial space occupying lesions, age and sex wise distribution, associated clinical symptoms and histopathological spectrum. Intracranial space occupying lesion biopsi Materials and Methods: es of 108 cases received indepartment of pathology, B.J. Medical College, Ahmedabad, during the period of March 2021 to August 2021. All specimens were preserved in 10% formalin and allowed to x for 24 hours, parafn embedded sections of 5 microns were cut. The hematoxylin and eosin stained sections of the CNS lesions were studied. One Hundred and Eight cases of Intracranial Spac Result: e Occupying lesions were studied, of which 8 (7.4%) cases were non neoplastic with 6 (75%) being cystic lesions and 2 (25%) were cerebral abscess. The neoplastic lesions comprised of 100 (92.6%) cases, which included 99 (99%) primary and 1 (1%) metastatic lesions. The most frequent type of CNS tumor was astrocytoma and oligodendroglioma (23 cases, 23%) followed by Meningioma (18 cases, 18%), pituitary adenoma (17 cases, 17%), Ependymoma(13 cases, 13 %) and metastatic tumor (1 case, 1%). The surgical pathologist plays an importa Conclusion: nt role in accurate diagnosis of various Intra [3] Cranial Space Occupying lesions which will be of immense help for patient prognosis and treatment.

SP2.1.1Continuous Monitoring and Assessment of Surgical Technical Skills Using Deep Learning

Abstract Aims Excellent surgical technical skills are of paramount importance to perform surgical procedures, safely and efficiently. Virtual reality surgical simulators can both simulate real operations while providing standardized, risk-free surgical hands-on experience. The integration of AI (artificial intelligence) and virtual reality simulators provides opportunities to carry out comprehensive continuous assessments of surgical performance. We developed and tested a deep learning algorithm which can continuously monitor and assess surgical bimanual performance on virtual reality surgical simulators. Methods Fifty participants from four expertise levels (14 experts/neurosurgeons, 14 senior residents, 10 junior residents, 12 medical students) performed a simulated subpial tumor resection 5 times and a complex simulated brain tumor operation once on the NeuroVR platform. Participants were asked to remove the tumors completely while minimizing bleeding and damage to surrounding tissues employing a simulated ultrasonic aspirator and bipolar forceps. A deep neural network continually tracked the surgical performance utilizing 16 performance metrics generated every 0.2-seconds. Results The deep neural network was successfully trained using neurosurgeons and medical students’ data, learning the composites of expertise comparing high and lower skill levels. The trained algorithm was able to score the technical skills of individuals continuously at 0.2-second intervals. Statistically significant differences in average scores were identified between the 4 groups. Conclusions AI-powered surgical simulators provide continuous assessment of bimanual technical skills during surgery which may further define the composites necessary to train surgical expertise. To our knowledge, this is the first attempt in surgery to continuously assess surgical technical skills using deep learning.

O51: ARTIFICIAL INTELLIGENCE UTILIZING RECURRENT NEURAL NETWORKS TO CONTINUOUSLY MONITOR COMPOSITES OF SURGICAL EXPERTISE

Abstract Introduction Many surgical adverse events occur secondary to technical errors related to poor bimanual skills, fatigue and lack of the required expertise. We developed AI algorithms to continuously assess surgical bimanual technical performance during virtual reality simulated surgical tasks. To our knowledge, this is the first attempt in surgery to train AI algorithms to continuously monitor and evaluate bimanual skills comprehensively. Method Fifty individuals from four expertise levels (14 experts/neurosurgeons, 14 senior residents, 10 junior residents, 12 medical students) performed two virtual reality simulated surgical tasks with haptic feedback: a subpial tumor resection 5 times and a complex, realistically simulated brain tumor operation once. Each task required complete tumor removal while minimizing bleeding and damage to surrounding tissues using a simulated ultrasonic aspirator and a bipolar. A recurrent neural network continually tracked individual bimanual performance utilizing 16 performance metrics generated every 0.2 seconds. Result The recurrent neural network algorithm was successfully trained using neurosurgeons and medical students' data, learning the composites of expertise comparing high and lower skill levels. The trained algorithm outlined and monitored technical skills every 0.2 second continuously organizing performance of each surgical task into three levels: ‘excellent’, ‘average’ and ‘poor’. The percentage time spent on each level was calculated and significant differences found between all four groups for ‘excellent’ and ‘poor’ levels. Conclusion AI-powered surgical simulators provide an advanced assessment and training tool. AI's ability to continuous assess bimanual technical skills during surgery may further define the composites necessary to train surgical expertise. Abbrev AI: artificial intelligence Take-home message By advanced artificial intelligence algorithms surgeon's bi-manual technical skills can be assessed continuously, time periods of poor performance which increase the possibility of errors in performance can be identified.

Resolution Enhancement of UWB Time-Reversal Microwave Imaging in Dispersive Environments

Time Reversal (TR) techniques allow optimal refocusing of ultrawideband (UWB) electromagnetic waves in complex propagation environments, due to the invariance of the wave equations. Propagation of electromagnetic waves through dispersive or lossy media breaks this invariance. Consequently, such media degrade the refocusing ability of TR techniques. In this work, we propose a novel algorithm for the enhancement of the UWB TR microwave imaging resolution in dispersive environments. The presented algorithm takes into account the frequency-dependent complex permittivity of the propagation medium across the entire bandwidth of the UWB pulse. Using this complex permittivity, it models the medium-, time-, and frequency- dependent attenuation in the wavelet domain to create inverse filters which compensate for the effects of the attenuation. This is the first algorithm that constructs inverse filters in the wavelet domain using the complex permittivity of the dispersive propagation medium across the entire bandwidth, rather than at a center frequency, of the excitation pulse. We also introduce a smart scaling concept to minimize undesired noise amplification. While our proposed approach is agnostic to the application scenario and thus could be utilized in various disciplines, we apply it in a biomedical imaging scenario and achieve enhancement in the resolution of UWB microwave TR imaging of a simulated brain tumor inside the digital human phantom (DHP).

Development of a performance model for virtual reality tumor resections.

Previous work from the authors has shown that hand ergonomics plays an important role in surgical psychomotor performance during virtual reality brain tumor resections. In the current study they propose a hypothetical model that integrates the human and task factors at play during simulated brain tumor resections to better understand the hand ergonomics needed for optimal safety and efficiency. They hypothesize that 1) experts (neurosurgeons), compared to novices (residents and medical students), spend a greater proportion of their time in direct contact with critical tumor areas; 2) hand ergonomic conditions (most favorable to unfavorable) prompt participants to adapt in order to optimize tumor resection; and 3) hand ergonomic adaptation is acquired with increasing expertise. In an earlier study, experts (neurosurgeons) and novices (residents and medical students) were instructed to resect simulated brain tumors on the NeuroVR (formerly NeuroTouch) virtual reality neurosurgical simulation platform. For the present study, the simulated tumors were divided into four quadrants (Q1 to Q4) to assess hand ergonomics at various levels of difficulty. The spatial distribution of time expended, force applied, and tumor volume removed was analyzed for each participant group (total of 22 participants). Neurosurgeons spent a significantly greater percentage of their time in direct contact with critical tumor areas. Under the favorable hand ergonomic conditions of Q1 and Q3, neurosurgeons and senior residents spent significantly more time in Q1 than in Q3. Although forces applied in these quadrants were similar, neurosurgeons, having spent more time in Q1, removed significantly more tumor in Q1 than in Q3. In a comparison of the most favorable (Q2) to unfavorable (Q4) hand ergonomic conditions, neurosurgeons adapted the forces applied in each quadrant to resect similar tumor volumes. Differences between Q2 and Q4 were emphasized in measures of force applied per second, tumor volume removed per second, and tumor volume removed per unit of force applied. In contrast, the hand ergonomics of medical students did not vary across quadrants, indicating the existence of an "adaptive capacity" in neurosurgeons. The study results confirm the experts' (neurosurgeons) greater capacity to adapt their hand ergonomics during simulated neurosurgical tasks. The proposed hypothetical model integrates the study findings with various human and task factors that highlight the importance of learning in the acquisition of hand ergonomic adaptation.

A Comparison of Visual Rating Scales and Simulated Virtual Reality Metrics in Neurosurgical Training: A Generalizability Theory Study

Adequate assessment and feedback remains a cornerstone of psychomotor skills acquisition, particularly within neurosurgery where the consequence of adverse operative events is significant. However, a critical appraisal of the reliability of visual rating scales in neurosurgery is lacking. Therefore, we sought to design a study to compare visual rating scales with simulated metrics in a neurosurgical virtual reality task. Neurosurgical faculty rated anonymized participant video recordings of the removal of simulated brain tumors using a visual rating scale made up of seven composite elements. Scale reliability was evaluated using generalizability theory, and scale subcomponents were compared with simulated metrics using Pearson correlation analysis. Four staff neurosurgeons evaluated 16 medical student neurosurgery applicants. Overall scale reliability and internal consistency were 0.73 and 0.90, respectively. Reliability of 0.71 was achieved with two raters. Individual participants, raters, and scale items accounted for 27%, 11%, and 0.6% of the data variability. The hemostasis scale component related to the greatest number of simulated metrics, whereas respect for no-go zones and tissue was correlated with none. Metrics relating to instrument force and patient safety (brain volume removed and blood loss) were captured by the fewest number of rating scale components. To our knowledge, this is the first study comparing participant's ratings with simulated performance. Given rating scales capture less well instrument force, quantity of brain volume removed, and blood loss, we suggest adopting a hybrid educational approach using visual rating scales in an operative environment, supplemented by simulated sessions to uncover potentially problematic surgical technique.

P.079 Development of a performance model for virtual reality tumor resections

Background: This work proposes a hypothetical model that integrates human factors (e.g. inherent ability and acquired expertise) and task factors (e.g. pre-procedural data, visual and haptic information) to better understand the hand ergonomics adaptation needed for optimal safety and efficiency during simulated brain tumor resections. Methods: Hand ergonomics of neurosurgeons, residents and medical students were assessed during simulated brain tumors resection on the NeuroVR virtual reality neurosurgical simulation platform. Spatial distribution of time expended, force applied, and tumor volume removed, and other metrics were analyzed in each tumor quadrant (Q1 to Q4). Results: Significant differences were observed between the most favorable hand ergonomics condition (Q2) and the unfavorable hand ergonomics condition (Q4). Neurosurgeons applied more total force, more mean force, and removed less tumor per unit of force applied in Q4. However, total volume removed was not significant between the two quadrants indicating hand ergonomics adaptation in order to maximize tumor removal. In comparison, hand ergonomics of medical students remained unchanged in all quadrants, indicating a learning phenomenon. Conclusions: Neurosurgeons are more capable of adapting their hand ergonomics during simulated brain tumor resections. Our proposed hypothetical model integrates our findings with the literature and highlights the importance of experience in the acquisition of adaptive hand ergonomics.

Virtual Reality Tumor Resection: The Force Pyramid Approach.

The force pyramid is a novel visual representation allowing spatial delineation of instrument force application during surgical procedures. In this study, the force pyramid concept is employed to create and quantify dominant hand, nondominant hand, and bimanual force pyramids during resection of virtual reality brain tumors. To address 4 questions: Do ergonomics and handedness influence force pyramid structure? What are the differences between dominant and nondominant force pyramids? What is the spatial distribution of forces applied in specific tumor quadrants? What differentiates "expert" and "novice" groups regarding their force pyramids? Using a simulated aspirator in the dominant hand and a simulated sucker in the nondominant hand, 6 neurosurgeons and 14 residents resected 8 different tumors using the CAE NeuroVR virtual reality neurosurgical simulation platform (CAE Healthcare, Montréal, Québec and the National Research Council Canada, Boucherville, Québec). Position and force data were used to create force pyramids and quantify tumor quadrant force distribution. Force distribution quantification demonstrates the critical role that handedness and ergonomics play on psychomotor performance during simulated brain tumor resections. Neurosurgeons concentrate their dominant hand forces in a defined crescent in the lower right tumor quadrant. Nondominant force pyramids showed a central peak force application in all groups. Bimanual force pyramids outlined the combined impact of each hand. Distinct force pyramid patterns were seen when tumor stiffness, border complexity, and color were altered. Force pyramids allow delineation of specific tumor regions requiring greater psychomotor ability to resect. This information can focus and improve resident technical skills training.

Towards patient-specific modeling of brain tumor growth and formation of secondary nodes guided by DTI-MRI.

Previous studies to simulate brain tumor progression, often investigate either temporal changes in cancer cell density or the overall tissue-level growth of the tumor mass. Here, we developed a computational model to bridge these two approaches. The model incorporates the tumor biomechanical response at the tissue level and accounts for cellular events by modeling cancer cell proliferation, infiltration to surrounding tissues, and invasion to distant locations. Moreover, acquisition of high resolution human data from anatomical magnetic resonance imaging, diffusion tensor imaging and perfusion imaging was employed within the simulations towards a realistic and patient specific model. The model predicted the intratumoral mechanical stresses to range from 20 to 34 kPa, which caused an up to 4.5 mm displacement to the adjacent healthy tissue. Furthermore, the model predicted plausible cancer cell invasion patterns within the brain along the white matter fiber tracts. Finally, by varying the tumor vascular density and its invasive outer ring thickness, our model showed the potential of these parameters for guiding the timing (83–90 days) of cancer cell distant invasion as well as the number (0–2 sites) and location (temportal and/or parietal lobe) of the invasion sites.

Groupwise registration of MR brain images with tumors

A novel groupwise image registration framework is developed for registering MR brain images with tumors. Our method iteratively estimates a normal-appearance counterpart for each tumor image to be registered and constructs a directed graph (digraph) of normal-appearance images to guide the groupwise image registration. Particularly, our method maps each tumor image to its normal appearance counterpart by identifying and inpainting brain tumor regions with intensity information estimated using a low-rank plus sparse matrix decomposition based image representation technique. The estimated normal-appearance images are groupwisely registered to a group center image guided by a digraph of images so that the total length of ‘image registration paths’ to be the minimum, and then the original tumor images are warped to the group center image using the resulting deformation fields. We have evaluated our method based on both simulated and real MR brain tumor images. The registration results were evaluated with overlap measures of corresponding brain regions and average entropy of image intensity information, and Wilcoxon signed rank tests were adopted to compare different methods with respect to their regional overlap measures. Compared with a groupwise image registration method that is applied to normal-appearance images estimated using the traditional low-rank plus sparse matrix decomposition based image inpainting, our method achieved higher image registration accuracy with statistical significance (p = 7.02 × 10−9).

Automaticity of Force Application During Simulated Brain Tumor Resection: Testing the Fitts and Posner Model

The Fitts and Posner model of motor learning hypothesized that with deliberate practice, learners progress through stages to an autonomous phase of motor ability. To test this model, we assessed the automaticity of neurosurgeons, senior residents, and junior residents when operating on 2 identical tumors using the NeuroVR virtual reality simulation platform. Participants resected 9 identical simulated tumors on 2 occasions (total = 18 resections). These resections were separated by the removal of a variable number of tumors with different visual and haptic complexities to mirror neurosurgical practice. Consistency of force application was used as a metric to assess automaticity and was defined as applying forces 1 standard deviation above or below a specific mean force application. Amount and specific location of force application during second identical tumor resection was compared to that used for the initial tumor. This study was conducted at the McGill Neurosurgical Simulation Research and Training Center, Montreal Neurologic Institute and Hospital, Montreal, Canada. Nine neurosurgeons, 10 senior residents, and 8 junior residents. Neurosurgeons display statistically significant increased consistency of force application when compared to resident groups when results from all tumor resections were assessed. Assessing individual tumor types demonstrates significant differences between the neurosurgeon and resident groups when resecting hard stiffness similar-to-background (white) tumors and medium-stiffness tumors. No statistical difference in consistency of force application was found when junior and senior residents were compared. "Experts" display significantly more automaticity when operating on identical simulated tumors separated by a series of different tumors using the NeuroVR platform. These results support the Fitts and Posner model of motor learning and are consistent with the concept that automaticity improves after completing residency training. The potential educational application of our findings is outlined related to neurosurgical resident training.

没入型脳外科手術シミュレータの開発

By using a surgery simulator, young doctors can train in surgical techniques to get essential skills and knowledge of surgery. In addition, surgeons can make a surgical plan and verify it. We have developed a neurosurgery simulator that can simulate surgical procedures in craniotomy such as brain retraction to expose a deep brain lesion. The simulator used virtual reality technology such as immersive display and haptic device to give a sense of presence. In this study, we simulated brain retraction and brain tumor resection to confirm operation verification and realistic sensation of the simulator.