Aim: Automated surgical skill assessment is poised to become an invaluable asset in surgical residency training. In our study, we aimed to create deep learning (DL) computer vision artificial intelligence (AI) models capable of automatically assessing trainee performance and determining proficiency on robotic suturing tasks.

Methods: Participants performed two robotic suturing tasks on a bench-top model created by our lab. Videos were recorded of each surgeon performing a backhand suturing task and a railroad suturing task at 30 frames per second (FPS) and downsampled to 15 FPS for the study. Each video was segmented into four sub-stitch phases: needle positioning, targeting, driving, and withdrawal. Each sub-stitch was annotated with a binary technical score (ideal or non-ideal), reflecting the operator’s skill while performing the suturing action. For DL analysis, 16-frame overlapping clips were sampled from the videos with a stride of 1. To extract the features useful for classification, two pretrained Video Swin Transformer models were fine-tuned using these clips: one to classify the sub-stitch phase and another to predict the technical score. The model outputs were then combined and used to train a Random Forest Classifier to predict the surgeon's proficiency level.

Results: A total of 102 videos from 27 surgeons were evaluated using 3-fold cross-validation, 51 videos for the backhand suturing task and 51 videos for the railroad suturing task. Performance was assessed on sub-stitch classification accuracy, technical score accuracy, and surgeon proficiency prediction. The clip-based Video Swin Transformer models achieved an average classification accuracy of 70.23% for sub-stitch classification and 68.4% for technical score prediction on the test folds. Combining the model outputs, the Random Forest Classifier achieved an average accuracy of 66.7% in predicting surgeon proficiency.

Conclusion: This study shows the feasibility of creating a DL-based automatic assessment tool for robotic-assisted surgery. Using machine learning models, we predicted the proficiency level of a surgeon with 66.7% accuracy. Our dry lab model proposes a standardized training and assessment tool for suturing tasks using computer vision.

Artificial intelligence (AI), machine learning (ML), and image guidance are increasingly being used to support surgeons in preoperative, intraoperative, and postoperative decision making and optimized patient care. Surgery is the cornerstone of curative treatment in pancreatic diseases, and a large amount of perioperative data are becoming available with the widespread application of minimally invasive surgical techniques. AI is showing promise in the prediction of malignancy and resectability from preoperative images. A further clinical focus is the prediction of postoperative complications, especially pancreatic fistula, and several AI algorithms now outperform conventional fistula risk scores. Future research will be directed toward refinement of intraoperative decision support systems, individualization of surgical training, and improvement of pre- and postoperative oncologic risk stratification to personalize the sequence of surgery and chemotherapy. This review summarizes recent developments in AI and image guidance for pancreatic surgery.

Aim: The study examines the evolving role of podcasts as educational tools for hepatopancreaticobiliary (HPB) surgeons, emphasizing their flexibility, accessibility, and potential for fostering global learning communities.

Methods: We review the rise of podcasts, focusing on general and HPB surgery podcasts, including their content, reach, and engagement metrics. It compares data from various podcast platforms and explores the integration of artificial intelligence (AI)-powered translation and dubbing technologies to address language barriers in medical education.

Results: Podcasts are becoming an increasingly valuable resource for surgeons, particularly those in HPB surgery. Specialized podcasts such as “Behind the Knife” offer high engagement with focused content, while general surgery podcasts provide broader yet less concentrated education. AI technologies can enhance global accessibility by overcoming language barriers, enabling non-English-speaking surgeons to benefit from these educational resources. AI-driven voice synthesis tools like Rask.AI, ElevenLabs, and Lovo.ai enhance the listening experience by creating natural-sounding, real-time dubbed voiceovers in multiple languages, with Lovo.ai specifically leveraging machine learning to ensure accurate and contextually appropriate translations for specialized fields like surgery.

Conclusion: Podcasts are poised to become central to HPB surgical education, offering flexibility and global reach. Addressing content fragmentation and language barriers through AI and centralizing resources could enhance their effectiveness, making high-quality medical education more inclusive and accessible worldwide.

Aim: Traditional methods of evaluating surgical performance, such as self-assessment and peer review, are limited by bias and inconsistency. Recent advances in artificial intelligence (AI) have introduced novel tools for objective evaluation of surgical techniques. This study reports the implementation of an AI-powered video management system across four surgical centers and its impact on the documentation, analysis, and standardization of minimally invasive surgeries (MIS).

Methods: A retrospective analysis was conducted of all MIS procedures performed at four centers within Assuta Medical Centers between July 2023 and June 2024. The AI system (Theator^TM, Inc.) is integrated with endoscopic cameras to automatically document, store, and analyze surgeries in real time, focusing on key intraoperative steps. Performance metrics, including the achievement of key surgical steps, were recorded. Rates of surgeon engagement in self-assessment and postoperative reviews were also evaluated.

Results: A total of 11,080 MIS procedures were performed, with 96.7% (10,725) documented by the AI system. The most frequently performed procedures were laparoscopic inguinal hernia repair (36.6%), gastric bypass (22.7%), and cholecystectomy (19.9%). The Critical View of Safety (CVS) was achieved in 60.6% of cholecystectomies, with inter-center variability ranging from 14% to 70%. Surgeon self-assessment was conducted in 22.2% of documented cases.

Conclusion: The implementation of an AI-powered video management system facilitated comprehensive surgical documentation and analysis, supporting both the standardization of surgical key steps and surgeon self-assessment. This system holds promise for improving surgical performance and safety through enhanced feedback and data-driven practice improvements.

Lower limb amputation (LLA) secondary to trauma, oncologic, diabetic, and vascular disease represents a significant patient challenge in terms of restoring function to pre-injury levels. This can be secondary to wear and use of a prosthetic limb, as well as limitations in range of motion or chronic pain. This study aimed to review and discuss the available, and potentially soon-to-be-available, roles of artificial intelligence (AI) in extremity amputation care. Specifically, we discuss the current state of AI technology in LLA prevention, management, peripheral nerve injury treatment, and lower limb prosthesis design, as well as highlighting current advancements and the direction of these linked fields.

Aim: We aim to develop a machine learning (ML)-driven predictive algorithm for patient-reported dysphagia after instrumented cervical fusion. Additionally, we aim to identify features for the prediction of dysphagia and to develop a web-based risk calculator for outcome prediction.

Methods: We identified consecutive adults who underwent instrumented cervical fusion at a single institution between 2013-2020. We developed regression-based and ML-based prognostic models and assessed model performance using discrimination and calibration. Additionally, we identified patient features driving the performance of the most effective model.

Results: Nine hundred and forty-seven patients were included in this study. There were 62 cases of dysphagia. The gradient boosting model was well-calibrated and demonstrated the highest discrimination of all tested models. The most important features for model performance included: anterior approach, deformity, revision procedure, bipolar disorder, diabetes mellitus, depression/anxiety, male sex, and myelopathy.

Conclusion: We report a ML-driven model that accurately predicts patient-reported dysphagia after instrumented cervical fusion. Prediction of dysphagia risk may inform preoperative counseling and appropriate risk stratification. Furthermore, this model may identify modifiable risk factors that may be addressed preoperatively to reduce the risk of dysphagia after cervical fusion.

Traumatic brain injury can cause intracranial hemorrhages, which, if not diagnosed and treated early, may lead to fatal outcomes due to excessive bleeding inside the cranium. In the United States, stroke is the fifth-leading cause of death, and approximately 10% of strokes result from intracranial hemorrhages. Identifying the presence, location, and type of hemorrhage is a critical step in treating emergency room patients. To determine the position and size of hemorrhages, X-ray computed tomography (CT) scans are commonly employed. While radiologists are highly skilled in analyzing CT scan images, the process is time-consuming. This study deals with intracranial hemorrhage detection using a deep-learning approach. First, we introduce convolutional neural networks (CNNs), a type of neural network designed for image-based datasets. Subsequently, we discuss various hyperparameter optimization techniques to enhance CNN training efficiency. As CNN training can be computationally expensive and time-intensive in many instances, we address this challenge by leveraging transfer learning with pre-trained models. We explore different transfer learning architectures, including VGGNet, AlexNet, EfficientNetB2, ResNet, MobileNet, and InceptionNet. The proposed framework for intracranial hemorrhage detection is implemented using transfer learning on the Radiological Society of North America Intracranial Hemorrhage Detection dataset, which is publicly available on Kaggle. Specifically, VGGNet is employed within this framework using powerful deep-learning libraries such as TensorFlow. This methodology can also be generalized to classify CT scan images in other biomedical domains.

Artificial intelligence (AI) is poised to revolutionize surgical care by leveraging the vast and complex “data lake” of healthcare information. This perspective piece outlines how AI may harness structured and unstructured data to improve patient outcomes. Advances in deep learning and foundational models have enabled the development of predictive analytics, automated clinical documentation, personalized patient chatbots, remote monitoring, and enhanced medical imaging. Examples include the ACS NSQIP risk calculator, Sepsis ImmunoScore, startups in ambient transcription, and cutting-edge AI applications in intraoperative imaging and real-time diagnostics. However, the adoption of AI in healthcare requires overcoming challenges, including data privacy, bias, integration into clinical workflows, interoperability, cost, ethical concerns, and regulatory hurdles. As AI technologies evolve, collaboration between surgeons and scientists will be critical to ensure ethical, patient-centered designs. This manuscript calls for surgeons to lead AI applications role in surgery, bridging technology with meaningful use cases to positively align with clinical practice.

From novel reconstructive operations to the creation of negative pressure wound therapy, plastic surgery is defined by a rich history of surgical and technological innovation. One of the latest technologies changing the face of medicine is artificial intelligence (AI), with its increasing popularity embodied by the meteoric rise in AI-related publications over the last decade. Abdominal wall reconstruction (AWR) is a discipline within plastic surgery that has taken an interest in AI technology, incorporating it into research to better understand hernia outcomes, interpret preoperative data, and improve patient-specific care and education. This review aims to explore the current breadth of AI use within AWR to give readers a better understanding of where the field currently stands and inspire ideas for where it may go in the future.

Artificial intelligence (AI) is profoundly impacting most, if not all, scientific and medical disciplines. In abdominal wall surgery (AWS), which includes common procedures such as hernia repair, abdominal wall reconstruction, and separation, AI models trained on surgical data have immense potential to enhance clinical practice and patient outcomes. The benefits include better procedure planning, standardization, interventional guidance, awareness of critical structures, complication prevention, quality assurance, and patient monitoring. Moreover, AI may significantly transform surgical education by enhancing training, skill assessment, and feedback mechanisms, leading to better-prepared surgeons. This review article highlights the latest developments in AI and AWS, focusing on key emerging applications and why embracing AI model prediction uncertainty is essential to translating these research efforts to clinical practice.

Traditional imaging techniques are limited by their preoperative nature, limited image resolution, and the need for radiologist interpretation. Multiple advanced imaging technologies have been developed, which may enhance surgical precision and patient outcomes.

Near-infrared fluorescence (NIRF) imaging, particularly with indocyanine green (ICG) dye, enables perfusion assessments and may help prevent anastomotic leaks. Additionally, NIRF can augment the identification of tumour growth patterns and lymphatic networks, thereby improving resection margin accuracy. Combining NIRF with radioisotope tracers allows for deep tissue navigation with high-precision dissection. In advanced disease, radioisotope scans may also enable prompt identification and excision of distally affected lymph nodes. Hyperspectral imaging (HSI) provides molecular-level information without the need for harmful contrast agents. HSI tissue vascularisation data may help shorten procedure times and reduce perioperative morbidity. Furthermore, when combined with neural networks, the technology can improve tumour detection and tissue differentiation.

Extended reality (XR) has multiple applications within surgical imaging. Augmented reality (AR) allows for intraoperative image overlays, thereby improving surgical navigation. Additionally, virtual reality (VR) may help users to visualise three-dimensional anatomical reconstructions, with applications in surgical training and patient consent. Artificial intelligence (AI) systems offer enhanced perioperative information to surgeons, such as the prediction of both disease progression and patient response to treatment. These benefits are compounded when paired with imaging modalities, such as HSI and XR.

Overall, advanced imaging technologies offer an exciting future for surgical practice and improved patient outcomes. Further work, including standardised protocols and ethical frameworks, is required.

Artificial intelligence (AI) is reshaping healthcare, particularly within the realm of spinal surgery, enhancing diagnostics, treatment, and patient management. AI is not only enhancing the technical aspects of spinal surgery but also revolutionizing patient care through personalized management, setting a new standard within the field. This computational renaissance has received increasing attention from providers and regulatory bodies to ensure novel technologies are being safely and effectively used. This review explores contemporary uses of AI in adult spinal deformity (ASD) surgery and the extent of their validation. Given the increasing complexity of ASD surgery and the expanding capabilities of AI, this review is essential to synthesize current applications, evaluate methodological strengths and limitations, and highlight future research opportunities in this evolving field.

Liver transplantation (LT) is the definitive treatment for end-stage liver disease and certain liver cancers. This involves complex decision making across the transplant continuum. Artificial intelligence (AI), with its ability to analyze high-dimensional data and derive meaningful patterns, shows promise as a transformative tool to address these challenges. In this narrative review, we searched PubMed from January 2021 to October 2024 using keywords such as “artificial intelligence”, “machine learning”, “deep learning”, and “liver transplantation”. Only full-text, English-language studies on adult populations (with minimum sample sizes deemed appropriate by each study’s design) were included, with a total of 65 articles. These publications examined AI applications in pre-transplant risk assessment (9), donor liver assessment (11), transplant oncology (11), graft survival prediction (7), overall survival prediction (11), immunosuppression management (4), and post-transplant risk prediction (12). Tree-based methods showed high accuracy in predictive tasks, while deep learning excelled in medical imaging analysis. Despite these advancements, only 6% of studies addressed algorithmic fairness, and 41% of neural network implementations lacked interpretability methods. Key challenges included data harmonization, multicenter validation, and integration with existing clinical workflows. Despite these limitations, AI continues to show promise for optimizing critical steps along the LT continuum. As the field progresses, the focus must remain on using AI to expand access and optimize care, ensuring it supports rather than restricts transplant opportunities.