RESUMEN
The convergence of digital pathology and artificial intelligence could assist histopathology image analysis by providing tools for rapid, automated morphological analysis. This systematic review explores the use of artificial intelligence for histopathological image analysis of digitised central nervous system (CNS) tumour slides. Comprehensive searches were conducted across EMBASE, Medline and the Cochrane Library up to June 2023 using relevant keywords. Sixty-eight suitable studies were identified and qualitatively analysed. The risk of bias was evaluated using the Prediction model Risk of Bias Assessment Tool (PROBAST) criteria. All the studies were retrospective and preclinical. Gliomas were the most frequently analysed tumour type. The majority of studies used convolutional neural networks or support vector machines, and the most common goal of the model was for tumour classification and/or grading from haematoxylin and eosin-stained slides. The majority of studies were conducted when legacy World Health Organisation (WHO) classifications were in place, which at the time relied predominantly on histological (morphological) features but have since been superseded by molecular advances. Overall, there was a high risk of bias in all studies analysed. Persistent issues included inadequate transparency in reporting the number of patients and/or images within the model development and testing cohorts, absence of external validation, and insufficient recognition of batch effects in multi-institutional datasets. Based on these findings, we outline practical recommendations for future work including a framework for clinical implementation, in particular, better informing the artificial intelligence community of the needs of the neuropathologist.
Asunto(s)
Inteligencia Artificial , Neoplasias del Sistema Nervioso Central , Humanos , Neoplasias del Sistema Nervioso Central/patología , Procesamiento de Imagen Asistido por Computador/métodosRESUMEN
PURPOSE: Accessible patient information sources are vital in educating patients about the benefits and risks of spinal surgery, which is crucial for obtaining informed consent. We aim to assess the effectiveness of a natural language processing (NLP) pipeline in recognizing surgical procedures from clinic letters and linking this with educational resources. METHODS: Retrospective examination of letters from patients seeking surgery for degenerative spinal disease at a single neurosurgical center. We utilized MedCAT, a named entity recognition and linking NLP, integrated into the electronic health record (EHR), which extracts concepts and links them to systematized nomenclature of medicine-clinical terms (SNOMED-CT). Investigators reviewed clinic letters, identifying words or phrases that described or identified operations and recording the SNOMED-CT terms as ground truth. This was compared to SNOMED-CT terms identified by the model, untrained on our dataset. A pipeline linking clinic letters to patient-specific educational resources was established, and precision, recall, and F1 scores were calculated. RESULTS: Across 199 letters the model identified 582 surgical procedures, and the overall pipeline after adding rules a total of 784 procedures (precision = 0.94, recall = 0.86, F1 = 0.91). Across 187 letters with identified SNOMED-CT terms the integrated pipeline linking education resources directly to the EHR was successful in 157 (78%) patients (precision = 0.99, recall = 0.87, F1 = 0.92). CONCLUSIONS: NLP accurately identifies surgical procedures in pre-operative clinic letters within an untrained subspecialty. Performance varies among letter authors and depends on the language used by clinicians. The identified procedures can be linked to patient education resources, potentially improving patients' understanding of surgical procedures.
Asunto(s)
Procesamiento de Lenguaje Natural , Educación del Paciente como Asunto , Humanos , Educación del Paciente como Asunto/métodos , Estudios Retrospectivos , Registros Electrónicos de Salud , Systematized Nomenclature of MedicineRESUMEN
OBJECTIVE: Spinal cerebrospinal fluid (CSF) leaks are common, and their management is heterogeneous. For high-flow leaks, numerous studies advocate for primary dural repair and CSF diversion. The LiquoGuard7® allows automated and precise pressure and volume control, and calculation of patient-specific CSF production rate (prCSF), which is hypothesized to be increased in the context of durotomies and CSF leaks. METHODS: This single-centre illustrative case series included patients undergoing complex spinal surgery where: 1) a high flow intra-operative and/or post-operative CSF leak was expected and 2) lumbar CSF drainage was performed using a LiquoGuard7®. CSF diversion was tailored to prCSF for each patient, combined with layered spinal wound closure. RESULTS: Three patients were included, with a variety of pathologies: T7/T8 disc prolapse, T8-T9 meningioma, and T4-T5 metastatic spinal cord compression. The first two patients underwent CSF diversion to prevent post-op CSF leak, whilst the third required this in response to post-op CSF leak. CSF hyperproduction was evident in all cases (mean >/=140ml/hr). With patient-specific CSF diversion regimes, no cases required further intervention for CSF fistulae repair (including for pleural CSF effusion), wound breakdown or infection. CONCLUSIONS: Patient-specific cerebrospinal fluid drainage may be a useful tool in the management of high-flow intra-operative and post-operative CSF leaks during complex spinal surgery. These systems may reduce post-operative CSF leakage from the wound or into adjacent body cavities. Further larger studies are needed to evaluate the comparative benefits and cost-effectiveness of this approach.
RESUMEN
OBJECTIVE: A scoping review of the literature was conducted to identify intraoperative artificial intelligence (AI) applications for robotic surgery under development and categorize them by (1) purpose of the applications, (2) level of autonomy, (3) stage of development, and (4) type of measured outcome. BACKGROUND: In robotic surgery, AI-based applications have the potential to disrupt a field so far based on a master-slave paradigm. However, there is no available overview about this technology's current stage of development and level of autonomy. METHODS: MEDLINE and EMBASE were searched between January 1, 2010 and May 21, 2022. Abstract screening, full-text review, and data extraction were performed independently by 2 reviewers. The level of autonomy was defined according to the Yang and colleagues' classification and stage of development according to the Idea, Development, Evaluation, Assessment, and Long-term follow-up framework. RESULTS: One hundred twenty-nine studies were included in the review. Ninety-seven studies (75%) described applications providing Robot Assistance (autonomy level 1), 30 studies (23%) application enabling Task Autonomy (autonomy level 2), and 2 studies (2%) application achieving Conditional autonomy (autonomy level 3). All studies were at Idea, Development, Evaluation, Assessment, and Long-term follow-up stage 0 and no clinical investigations on humans were found. One hundred sixteen (90%) conducted in silico or ex vivo experiments on inorganic material, 9 (7%) ex vivo experiments on organic material, and 4 (3%) performed in vivo experiments in porcine models. CONCLUSIONS: Clinical evaluation of intraoperative AI applications for robotic surgery is still in its infancy and most applications have a low level of autonomy. With increasing levels of autonomy, the evaluation focus seems to shift from AI-specific metrics to process outcomes, although common standards are needed to allow comparison between systems.
Asunto(s)
Procedimientos Quirúrgicos Robotizados , Humanos , Animales , Porcinos , Inteligencia Artificial , BenchmarkingRESUMEN
INTRODUCTION: AO Spine RECODE-DCM was a multi-stakeholder priority setting partnership (PSP) to define the top ten research priorities for degenerative cervical myelopathy (DCM). Priorities were generated and iteratively refined using a series of surveys administered to surgeons, other healthcare professionals (oHCP) and people with DCM (PwDCM). The aim of this work was to utilise word clouds to enable the perspectives of people with the condition to be heard earlier in the PSP process than is traditionally the case. The objective was to evaluate the added value of word clouds in the process of defining research uncertainties in National Institute for Health Research (NIHR) James Lind Alliance (JLA) Priority Setting Partnerships. METHODS: Patient-generated word clouds were created for the four survey subsections of the AO Spine RECODE-DCM PSP: diagnosis, treatment, long-term management and other issues. These were then evaluated as a nested methodological study. Word-clouds were created and iteratively refined by an online support group of people with DCM, before being curated by the RECODE-DCM management committee and expert healthcare professional representatives. The final word clouds were embedded within the surveys administered at random to 50% of participants. DCM research uncertainties suggested by participants were compared pre- and post-word cloud presentation. RESULTS: A total of 215 (50.9%) participants were randomised to the word cloud stream, including 118 (55%) spinal surgeons, 52 (24%) PwDCM and 45 (21%) oHCP. Participants submitted 434 additional uncertainties after word cloud review: word count was lower and more uniform across each survey subsections compared to pre-word cloud uncertainties. Twenty-three (32%) of the final 74 PSP summary questions did not have a post-word cloud contribution and no summary question was formed exclusively on post-word cloud uncertainties. There were differences in mapping of pre- and post-word cloud uncertainties to summary questions, with greater mapping of post-word cloud uncertainties to the number 1 research question priority: raising awareness. Five of the final summary questions were more likely to map to the research uncertainties suggested by participants after having reviewed the word clouds. CONCLUSIONS: Word clouds may increase the perspective of underrepresented stakeholders in the research question gathering stage of priority setting partnerships. This may help steer the process towards research questions that are of highest priority for people with the condition.
Asunto(s)
Investigación Biomédica , Prioridades en Salud , Humanos , Incertidumbre , Personal de Salud , Encuestas y CuestionariosRESUMEN
PURPOSE: Heterogeneous reporting in baseline variables in patients undergoing transsphenoidal resection of pituitary adenoma precludes meaningful meta-analysis. We therefore examined trends in reported baseline variables, and degree of heterogeneity of reported variables in 30 years of literature. METHODS: A systematic review of PubMed and Embase was conducted on studies that reported outcomes for transsphenoidal surgery for pituitary adenoma 1990-2021. The protocol was registered a priori and adhered to the PRISMA statement. Full-text studies in English with > 10 patients (prospective), > 500 patients (retrospective), or randomised trials were included. RESULTS: 178 studies were included, comprising 427,659 patients: 52 retrospective (29%); 118 prospective (66%); 9 randomised controlled trials (5%). The majority of studies were published in the last 10 years (71%) and originated from North America (38%). Most studies described patient demographics, such as age (165 studies, 93%) and sex (164 studies, 92%). Ethnicity (24%) and co-morbidities (25%) were less frequently reported. Clinical baseline variables included endocrine (60%), ophthalmic (34%), nasal (7%), and cognitive (5%). Preoperative radiological variables were described in 132 studies (74%). MRI alone was the most utilised imaging modality (67%). Further specific radiological baseline variables included: tumour diameter (52 studies, 39%); tumour volume (28 studies, 21%); cavernous sinus invasion (53 studies, 40%); Wilson Hardy grade (25 studies, 19%); Knosp grade (36 studies, 27%). CONCLUSIONS: There is heterogeneity in the reporting of baseline variables in patients undergoing transsphenoidal surgery for pituitary adenoma. This review supports the need to develop a common data element to facilitate meaningful comparative research, trial design, and reduce research inefficiency.
Asunto(s)
Adenoma , Neoplasias Hipofisarias , Humanos , Adenoma/cirugía , Adenoma/patología , Neoplasias Hipofisarias/cirugía , Neoplasias Hipofisarias/patología , Estudios Prospectivos , Estudios Retrospectivos , Resultado del TratamientoRESUMEN
PURPOSE: Postoperative cerebrospinal fluid rhinorrhoea (CSFR) remains a frequent complication of endonasal approaches to pituitary and skull base tumours. Watertight skull base reconstruction is important in preventing CSFR. We sought to systematically review the current literature of available skull base repair techniques. METHODS: Pubmed and Embase databases were searched for studies (2000-2020) that (a) reported on the endonasal resection of pituitary and skull base tumours, (b) focussed on skull base repair techniques and/or postoperative CSFR risk factors, and (c) included CSFR data. Roles, advantages and disadvantages of each repair method were detailed. Random-effects meta-analyses were performed where possible. RESULTS: 193 studies were included. Repair methods were categorised based on function and anatomical level. There was absolute heterogeneity in repair methods used, with no independent studies sharing the same repair protocol. Techniques most commonly used for low CSFR risk cases were fat grafts, fascia lata grafts and synthetic grafts. For cases with higher CSFR risk, multilayer regimes were utilized with vascularized flaps, gasket sealing and lumbar drains. Lumbar drain use for high CSFR risk cases was supported by a randomised study (Oxford CEBM: Grade B recommendation), but otherwise there was limited high-level evidence. Pooled CSFR incidence by approach was 3.7% (CI 3-4.5%) for transsphenoidal, 9% (CI 7.2-11.3%) for expanded endonasal, and 5.3% (CI 3.4-7%) for studies describing both. Further meaningful meta-analyses of repair methods were not performed due to significant repair protocol heterogeneity. CONCLUSIONS: Modern reconstructive protocols are heterogeneous and there is limited evidence to suggest the optimal repair technique after pituitary and skull base tumour resection. Further studies are needed to guide practice.
Asunto(s)
Rinorrea de Líquido Cefalorraquídeo , Neoplasias de la Base del Cráneo , Pérdida de Líquido Cefalorraquídeo/etiología , Endoscopía , Humanos , Complicaciones Posoperatorias , Estudios Retrospectivos , Base del Cráneo/cirugía , Neoplasias de la Base del Cráneo/cirugíaRESUMEN
PURPOSE: Surgical workflow analysis seeks to systematically break down operations into hierarchal components. It facilitates education, training, and understanding of surgical variations. There are known educational demands and variations in surgical practice in endoscopic transsphenoidal approaches to pituitary adenomas. Through an iterative consensus process, we generated a surgical workflow reflective of contemporary surgical practice. METHODS: A mixed-methods consensus process composed of a literature review and iterative Delphi surveys was carried out within the Pituitary Society. Each round of the survey was repeated until data saturation and > 90% consensus was reached. RESULTS: There was a 100% response rate and no attrition across both Delphi rounds. Eighteen international expert panel members participated. An extensive workflow of 4 phases (nasal, sphenoid, sellar and closure) and 40 steps, with associated technical errors and adverse events, were agreed upon by 100% of panel members across rounds. Both core and case-specific or surgeon-specific variations in operative steps were captured. CONCLUSIONS: Through an international expert panel consensus, a workflow for the performance of endoscopic transsphenoidal pituitary adenoma resection has been generated. This workflow captures a wide range of contemporary operative practice. The agreed "core" steps will serve as a foundation for education, training, assessment and technological development (e.g. models and simulators). The "optional" steps highlight areas of heterogeneity of practice that will benefit from further research (e.g. methods of skull base repair). Further adjustments could be made to increase applicability around the world.
Asunto(s)
Adenoma , Neoplasias Hipofisarias , Adenoma/cirugía , Endoscopía , Humanos , Neoplasias Hipofisarias/cirugía , Estudios Retrospectivos , Hueso Esfenoides , Resultado del Tratamiento , Flujo de TrabajoRESUMEN
INTRODUCTION: The gold-standard treatment for symptomatic anterior skull base meningiomas is surgical resection. The endoscope-assisted supraorbital "keyhole" approach (eSKA) is a promising technique for surgical resection of olfactory groove (OGM) and tuberculum sellae meningioma (TSM) but has yet to be compared with the microscopic transcranial (mTCA) and the expanded endoscopic endonasal approach (EEA) in the context of existing literature. METHODS: An updated study-level meta-analysis on surgical outcomes and complications of OGM and TSM operated with the eSKA, mTCA, and EEA was conducted using random-effect models. RESULTS: A total of 2285 articles were screened, yielding 96 studies (2191 TSM and 1510 OGM patients). In terms of effectiveness, gross total resection incidence was highest in mTCA (89.6% TSM, 91.1% OGM), followed by eSKA (85.2% TSM, 84.9% OGM) and EEA (83.9% TSM, 82.8% OGM). Additionally, the EEA group had the highest incidence of visual improvement (81.9% TSM, 54.6% OGM), followed by eSKA (65.9% TSM, 52.9% OGM) and mTCA (63.9% TSM, 45.7% OGM). However, in terms of safety, the EEA possessed the highest cerebrospinal fluid leak incidence (9.2% TSM, 14.5% OGM), compared with eSKA (2.1% TSM, 1.6% OGM) and mTCA (1.6% TSM, 6.5% OGM). Finally, mortality and intraoperative arterial injury were 1% or lower across all subgroups. CONCLUSIONS: In the context of diverse study populations, the eSKA appeared not to be associated with increased adverse outcomes when compared with mTCA and EEA and offered comparable effectiveness. Case-selection is paramount in establishing a role for the eSKA in anterior skull base tumours.
Asunto(s)
Neoplasias Meníngeas/cirugía , Meningioma/cirugía , Cirugía Endoscópica por Orificios Naturales/métodos , Neuroendoscopía/métodos , Neoplasias de la Base del Cráneo/cirugía , Pérdida de Líquido Cefalorraquídeo/epidemiología , Ojo , Femenino , Humanos , Masculino , Persona de Mediana Edad , Cirugía Endoscópica por Orificios Naturales/efectos adversos , Neuroendoscopía/efectos adversos , Complicaciones Posoperatorias/epidemiología , Lesiones del Sistema Vascular/epidemiologíaRESUMEN
OBJECTIVE: The extended endoscopic approach provides unimpaired visualization and direct access to ventral skull base pathology, but is associated with cerebrospinal fluid (CSF) leak in up to 25% of patients. To evaluate the impact of improved surgical techniques and devices to better repair skull base defects, we assessed published surgical outcomes of the extended endoscopic endonasal approach in the last two decades for a well-defined homogenous group of tuberculum sellae and olfactory groove meningioma patients. METHODS: Random-effects meta-analyses were performed for studies published between 2004 (first publications) and April 2020. We evaluated CSF leak as primary outcome. Secondary outcomes were gross total resection, improvement in visual outcomes in those presenting with a deficit, intraoperative arterial injury, and 30-day mortality. For the main analyses, publications were pragmatically grouped based on publication year in three categories: 2004-2010, 2011-2015, and 2016-2020. RESULTS: We included 29 studies describing 540 patients with tuberculum sellae and 115 with olfactory groove meningioma. The percentage patients with CSF leak dropped over time from 22% (95% CI: 6-43%) in studies published between 2004 and 2010, to 16% (95% CI: 11-23%) between 2011 and 2015, and 4% (95% CI: 1-9%) between 2016 and 2020. Outcomes of gross total resection, visual improvement, intraoperative arterial injury, and 30-day mortality remained stable over time CONCLUSIONS: We report a noticeable decrease in CSF leak over time, which might be attributed to the development and improvement of new closure techniques (e.g., Hadad-Bassagasteguy flap, and gasket seal), refined multilayer repair protocols, and lumbar drain usage.
Asunto(s)
Pérdida de Líquido Cefalorraquídeo/epidemiología , Neoplasias Meníngeas/cirugía , Meningioma/cirugía , Cirugía Endoscópica por Orificios Naturales/efectos adversos , Complicaciones Posoperatorias/epidemiología , Neoplasias de la Base del Cráneo/cirugía , Adulto , Anciano , Femenino , Humanos , Masculino , Persona de Mediana EdadRESUMEN
BACKGROUND: The endonasal transsphenoidal approach (TSA) has emerged as the preferred approach in order to treat pituitary adenoma and related sellar pathologies. The recently adopted expanded endonasal approach (EEA) has improved access to the ventral skull base whilst retaining the principles of minimally invasive surgery. Despite the advantages these approaches offer, cerebrospinal fluid (CSF) rhinorrhoea remains a common complication. There is currently a lack of comparative evidence to guide the best choice of skull base reconstruction, resulting in considerable heterogeneity of current practice. This study aims to determine: (1) the scope of the methods of skull base repair; and (2) the corresponding rates of postoperative CSF rhinorrhoea in contemporary neurosurgical practice in the UK and Ireland. METHODS: We will adopt a multicentre, prospective, observational cohort design. All neurosurgical units in the UK and Ireland performing the relevant surgeries (TSA and EEA) will be eligible to participate. Eligible cases will be prospectively recruited over 6 months with 6 months of postoperative follow-up. Data points collected will include: demographics, tumour characteristics, operative data), and postoperative outcomes. Primary outcomes include skull base repair technique and CSF rhinorrhoea (biochemically confirmed and/or requiring intervention) rates. Pooled data will be analysed using descriptive statistics. All skull base repair methods used and CSF leak rates for TSA and EEA will be compared against rates listed in the literature. ETHICS AND DISSEMINATION: Formal institutional ethical board review was not required owing to the nature of the study - this was confirmed with the Health Research Authority, UK. CONCLUSIONS: The need for this multicentre, prospective, observational study is highlighted by the relative paucity of literature and the resultant lack of consensus on the topic. It is hoped that the results will give insight into contemporary practice in the UK and Ireland and will inform future studies.
Asunto(s)
Rinorrea de Líquido Cefalorraquídeo , Pérdida de Líquido Cefalorraquídeo , Rinorrea de Líquido Cefalorraquídeo/epidemiología , Rinorrea de Líquido Cefalorraquídeo/etiología , Rinorrea de Líquido Cefalorraquídeo/cirugía , Estudios de Cohortes , Humanos , Complicaciones Posoperatorias , Estudios Prospectivos , Estudios Retrospectivos , Base del Cráneo/cirugíaRESUMEN
Traumatic brain injury leads to cellular damage which in turn results in the rapid release of damage-associated molecular patterns (DAMPs) that prompt resident cells to release cytokines and chemokines. These in turn rapidly recruit neutrophils, which assist in limiting the spread of injury and removing cellular debris. Microglia continuously survey the CNS (central nervous system) compartment and identify structural abnormalities in neurons contributing to the response. After some days, when neutrophil numbers start to decline, activated microglia and astrocytes assemble at the injury site-segregating injured tissue from healthy tissue and facilitating restorative processes. Monocytes infiltrate the injury site to produce chemokines that recruit astrocytes which successively extend their processes towards monocytes during the recovery phase. In this fashion, monocytes infiltration serves to help repair the injured brain. Neurons and astrocytes also moderate brain inflammation via downregulation of cytotoxic inflammation. Depending on the severity of the brain injury, T and B cells can also be recruited to the brain pathology sites at later time points.
Asunto(s)
Astrocitos/patología , Lesiones Traumáticas del Encéfalo/patología , Encéfalo/patología , Microglía/patología , Neuronas/patología , Animales , Astrocitos/metabolismo , Encéfalo/metabolismo , Lesiones Traumáticas del Encéfalo/metabolismo , Quimiocinas/metabolismo , Citocinas/metabolismo , Encefalitis/metabolismo , Encefalitis/patología , Humanos , Microglía/metabolismo , Neuronas/metabolismoRESUMEN
The Expanded Endoscopic Endonasal Approach, one of the best examples of endoscopic neurosurgery, allows access to the skull base through the natural orifice of the nostril. Current standard instruments lack articulation limiting operative access and surgeon dexterity, and thus, could benefit from robotic articulation. In this study, a handheld robotic system with a series of detachable end-effectors for this approach is presented. This system is comprised of interchangeable articulated 2/3 degrees-of-freedom 3 mm instruments that expand the operative workspace and enhance the surgeon's dexterity, an ergonomically designed handheld controller with a rotating joystick-body that can be placed at the position most comfortable for the user, and the accompanying control box. The robotic instruments were experimentally evaluated for their workspace, structural integrity, and force-delivery capabilities. The entire system was then tested in a pre-clinical context during a phantom feasibility test, followed up by a cadaveric pilot study by a cohort of surgeons of varied clinical experience. Results from this series of experiments suggested enhanced dexterity and adequate robustness that could be associated with feasibility in a clinical context, as well as improvement over current neurosurgical instruments.
RESUMEN
BACKGROUND: Artificial intelligence (AI) is expected to play a greater role in neurosurgery. There is a need for neurosurgeons capable of critically appraising AI literature to evaluate its implementation or communicate information to patients. However, there are a lack of courses delivered at a level appropriate for individuals to develop such skills. We assessed the impact of a two day (non-credit bearing) online digital literacy course on the ability of individuals to critically appraise AI literature in neurosurgery. METHODS: We performed a prospective, quasi-experimental non-randomised, controlled study with an intervention arm comprised of individuals enrolled in our two-day digital health literacy course and a waiting-list control arm used for comparison. We assessed participants' pre- and post-course knowledge, confidence and course acceptability using Qualtrics surveys designed for the purpose of this study. RESULTS: A total of 62 participants (33 participants, 29 waitlist controls), including neurosurgical trainees and both undergraduate and post-graduate students, attended the course and completed the pre-course survey. The two groups did not vary significantly in terms of age or demographics. Following the course, participants significantly improved in their knowledge of AI (mean difference=3.86, 95% CI=2.97-4.75, p-value<0.0001) and confidence in critically appraising literature using AI (p-value=0.002). Similar differences in knowledge (mean difference=3.15, 95% CI=1.82-4.47, p-value<0.0001) and confidence (p-value<0.0001) were found when compared to the control group. CONCLUSION: Bespoke courses delivered at an appropriate level can improve clinicians' understanding of the application of AI in neurosurgery, without the need for in-depth technical knowledge or programming skills.
RESUMEN
PURPOSE: Endoscopic pituitary surgery entails navigating through the nasal cavity and sphenoid sinus to access the sella using an endoscope. This procedure is intricate due to the proximity of crucial anatomical structures (e.g. carotid arteries and optic nerves) to pituitary tumours, and any unintended damage can lead to severe complications including blindness and death. Intraoperative guidance during this surgery could support improved localization of the critical structures leading to reducing the risk of complications. METHODS: A deep learning network PitSurgRT is proposed for real-time localization of critical structures in endoscopic pituitary surgery. The network uses high-resolution net (HRNet) as a backbone with a multi-head for jointly localizing critical anatomical structures while segmenting larger structures simultaneously. Moreover, the trained model is optimized and accelerated by using TensorRT. Finally, the model predictions are shown to neurosurgeons, to test their guidance capabilities. RESULTS: Compared with the state-of-the-art method, our model significantly reduces the mean error in landmark detection of the critical structures from 138.76 to 54.40 pixels in a 1280 × 720-pixel image. Furthermore, the semantic segmentation of the most critical structure, sella, is improved by 4.39% IoU. The inference speed of the accelerated model achieves 298 frames per second with floating-point-16 precision. In the study of 15 neurosurgeons, 88.67% of predictions are considered accurate enough for real-time guidance. CONCLUSION: The results from the quantitative evaluation, real-time acceleration, and neurosurgeon study demonstrate the proposed method is highly promising in providing real-time intraoperative guidance of the critical anatomical structures in endoscopic pituitary surgery.
Asunto(s)
Endoscopía , Neoplasias Hipofisarias , Humanos , Endoscopía/métodos , Neoplasias Hipofisarias/cirugía , Cirugía Asistida por Computador/métodos , Aprendizaje Profundo , Hipófisis/cirugía , Hipófisis/anatomía & histología , Hipófisis/diagnóstico por imagen , Seno Esfenoidal/cirugía , Seno Esfenoidal/anatomía & histología , Seno Esfenoidal/diagnóstico por imagenRESUMEN
BACKGROUND: Endoscopic pituitary adenoma surgery has a steep learning curve, with varying surgical techniques and outcomes across centers. In other surgeries, superior performance is linked with superior surgical outcomes. This study aimed to explore the prediction of patient-specific outcomes using surgical video analysis in pituitary surgery. METHODS: Endoscopic pituitary adenoma surgery videos from a single center were annotated by experts for surgical workflow (3 phases, 15 steps) and surgical skill (using modified Objective Structured Assessment of Technical Skills; mOSATS). Quantitative workflow metrics were calculated, including phase duration and step transitions. Poisson or logistic regression was used to assess the association of workflow metrics and mOSATS with common inpatient surgical outcomes. RESULTS: 100 videos from 100 patients were included. Nasal phase mean duration was 24mins and mean mOSATS was 21.2/30. Mean duration was 34mins and mean mOSATS was 20.9/30 for the sellar phase, and 11mins and 21.7/30 respectively for the closure phase. The most common adverse outcomes were new anterior pituitary hormone deficiency (n=26), dysnatremia (n=24) and cerebrospinal fluid (CSF) leak (n=5). Higher mOSATS for all three phases and shorter operation duration was associated with decreased length of stay (p=0.003 & p<0.001). Superior closure phase mOSATS were associated with reduced post-operative CSF leak (p<0.001), and superior sellar phase mOSATS were associated with reduced post-operative visual deterioration (p=0.041). CONCLUSION: Superior surgical skill and shorter surgical time were associated with superior surgical outcomes, at a generic and phase-specific level. Such video-based analysis has promise for integration into data-driven training and service improvement initiatives.
RESUMEN
BACKGROUND: Superior surgical skill improves surgical outcomes in endoscopic pituitary adenoma surgery. Video-based coaching programs, pioneered in professional sports, have shown promise in surgical training. In this study, we developed and assessed a video-based coaching program, using artificial intelligence (AI). METHODS: AI-assisted video-based surgical coaching was implemented over 6 months with the pituitary surgery team. The program consisted of 1) monthly random video analysis and review; and 2) quarterly 2-hour educational meetings discussing these videos and learning points. Each video was annotated for surgical phases and steps using AI, which improved video interactivity and allowed calculation of quantitative metrics. Primary outcomes were program feasibility, acceptability and appropriateness. Surgical performance (via modified Objective Structured Assessment of Technical Skills; mOSATS) and early surgical outcomes were recorded for every case during the 6-month coaching period, and a preceding 6-month control period. Beta and logistic regression were used to assess the change in mOSATS and surgical outcomes after the coaching program implementation. RESULTS: All participants highly rated the program's feasibility, acceptability and appropriateness. During the coaching program, 63 endoscopic pituitary adenoma cases were included, with 41 in the control group. Surgical performance across all operative phases improved during the coaching period (p<0.001), with a reduction in new post-operative anterior pituitary hormone deficit (p=0.01). CONCLUSION: We have developed a novel AI-assisted video surgical coaching program for endoscopic pituitary adenoma surgery - demonstrating its viability and impact on surgical performance. Early results also suggest improvement in patient outcomes. Future studies should be multicenter and longer term.
RESUMEN
The next generation of surgical robotics is poised to disrupt healthcare systems worldwide, requiring new frameworks for evaluation. However, evaluation during a surgical robot's development is challenging due to their complex evolving nature, potential for wider system disruption and integration with complementary technologies like artificial intelligence. Comparative clinical studies require attention to intervention context, learning curves and standardized outcomes. Long-term monitoring needs to transition toward collaborative, transparent and inclusive consortiums for real-world data collection. Here, the Idea, Development, Exploration, Assessment and Long-term monitoring (IDEAL) Robotics Colloquium proposes recommendations for evaluation during development, comparative study and clinical monitoring of surgical robots-providing practical recommendations for developers, clinicians, patients and healthcare systems. Multiple perspectives are considered, including economics, surgical training, human factors, ethics, patient perspectives and sustainability. Further work is needed on standardized metrics, health economic assessment models and global applicability of recommendations.
Asunto(s)
Inteligencia Artificial , Procedimientos Quirúrgicos Robotizados , Humanos , RobóticaRESUMEN
Operation notes are a crucial component of patient care. However, writing them manually is prone to human error, particularly in high pressured clinical environments. Automatic generation of operation notes from video recordings can alleviate some of the administrative burdens, improve accuracy, and provide additional information. To achieve this for endoscopic pituitary surgery, 27-steps were identified via expert consensus. Then, for the 97-videos recorded for this study, a timestamp of each step was annotated by an expert surgeon. To automatically determine whether a step is present in a video, a three-stage architecture was created. Firstly, for each step, a convolution neural network was used for binary image classification on each frame of a video. Secondly, for each step, the binary frame classifications were passed to a discriminator for binary video classification. Thirdly, for each video, the binary video classifications were passed to an accumulator for multi-label step classification. The architecture was trained on 77-videos, and tested on 20-videos, where a 0.80 weighted-F1 score was achieved. The classifications were inputted into a clinically based predefined template, and further enriched with additional video analytics. This work therefore demonstrates automatic generation of operative notes from surgical videos is feasible, and can assist surgeons during documentation.
RESUMEN
The vital physiological role of the pituitary gland, alongside its proximity to critical neurovascular structures, means that pituitary adenomas can cause significant morbidity or mortality. While enormous advancements have been made in the surgical care of pituitary adenomas, numerous challenges remain, such as treatment failure and recurrence. To meet these clinical challenges, there has been an enormous expansion of novel medical technologies (eg, endoscopy, advanced imaging, artificial intelligence). These innovations have the potential to benefit each step of the patient's journey, and ultimately, drive improved outcomes. Earlier and more accurate diagnosis addresses this in part. Analysis of novel patient data sets, such as automated facial analysis or natural language processing of medical records holds potential in achieving an earlier diagnosis. After diagnosis, treatment decision-making and planning will benefit from radiomics and multimodal machine learning models. Surgical safety and effectiveness will be transformed by smart simulation methods for trainees. Next-generation imaging techniques and augmented reality will enhance surgical planning and intraoperative navigation. Similarly, surgical abilities will be augmented by the future operative armamentarium, including advanced optical devices, smart instruments, and surgical robotics. Intraoperative support to surgical team members will benefit from a data science approach, utilizing machine learning analysis of operative videos to improve patient safety and orientate team members to a common workflow. Postoperatively, neural networks leveraging multimodal datasets will allow early detection of individuals at risk of complications and assist in the prediction of treatment failure, thus supporting patient-specific discharge and monitoring protocols. While these advancements in pituitary surgery hold promise to enhance the quality of care, clinicians must be the gatekeepers of the translation of such technologies, ensuring systematic assessment of risk and benefit prior to clinical implementation. In doing so, the synergy between these innovations can be leveraged to drive improved outcomes for patients of the future.