RESUMEN
OBJECTIVE: Develop a pioneer surgical anonymization algorithm for reliable and accurate real-time removal of out-of-body images validated across various robotic platforms. BACKGROUND: The use of surgical video data has become a common practice in enhancing research and training. Video sharing requires complete anonymization, which, in the case of endoscopic surgery, entails the removal of all nonsurgical video frames where the endoscope can record the patient or operating room staff. To date, no openly available algorithmic solution for surgical anonymization offers reliable real-time anonymization for video streaming, which is also robotic-platform and procedure-independent. METHODS: A data set of 63 surgical videos of 6 procedures performed on four robotic systems was annotated for out-of-body sequences. The resulting 496.828 images were used to develop a deep learning algorithm that automatically detected out-of-body frames. Our solution was subsequently benchmarked against existing anonymization methods. In addition, we offer a postprocessing step to enhance the performance and test a low-cost setup for real-time anonymization during live surgery streaming. RESULTS: Framewise anonymization yielded a receiver operating characteristic area under the curve score of 99.46% on unseen procedures, increasing to 99.89% after postprocessing. Our Robotic Anonymization Network outperforms previous state-of-the-art algorithms, even on unseen procedural types, despite the fact that alternative solutions are explicitly trained using these procedures. CONCLUSIONS: Our deep learning model, Robotic Anonymization Network, offers reliable, accurate, and safe real-time anonymization during complex and lengthy surgical procedures regardless of the robotic platform. The model can be used in real time for surgical live streaming and is openly available.
Asunto(s)
Algoritmos , Procedimientos Quirúrgicos Robotizados , Humanos , Anonimización de la Información , Grabación en Video , Aprendizaje ProfundoRESUMEN
BACKGROUND: The ergonomic advantages and potential challenges that robotic surgery poses to the well-being of surgeons are mainly unexplored. The most recent surgical robot introduced on the European market is the Hugo™ RAS System by Medtronic. This study aims to evaluate the ergonomic benefits of the Hugo™ RAS System, which is available in our training laboratory, CeMIT (Center for Medical Innovation and Technology Cologne). METHODS AND PROCEDURES: Using the previously established Cologne Ergonomic Measurement Setup for Robotic Surgery (CEMRobSurg), we measured three parameters related to ergonomic posture from subjects with different levels of surgical expertise (laypeople, medical students, surgical residents, and expert robotic surgeons). The heart rate was measured continuously using a polar band. The noise level was measured while using the Hugo™ RAS System, and automated photographs using our locally developed methodology were captured of the participant every 2 s to assess body posture. The ergonomic measurements were conducted while the subject performed the same standardized robotic training exercises (Peg Board, Rope Walk, and Ring Walk). RESULTS: A total of 53 participants were enrolled in this study. The average noise level during all measurements was 54.87 dB. The highest stress level was measured in surgical residents with a sympathetic nervous system index (SNS index) of 1.15 (min - 1.43, max 3.56). The lowest stress level was measured in robotic experts with an SNS index of 0.23 (min - 0.18, max 0.91). We observed a risk-prone positioning of the neck and elbow in medical students (mean 39.6° and 129.48°, respectively). Robotic experts showed a risk positioning in the knee and hip region (mean 107.89° and 90.31°, respectively). CONCLUSION: This is the first study to analyze and objectify the ergonomic posture of medical students, surgical trainees, surgeons, and laypeople using the open console, modular Hugo™ RAS System. Our findings offer recommendations for operating surgeons and allow for a comparative analysis between the different robotic systems. Further evaluations in real-time operative scenarios will follow.
RESUMEN
Previous studies have shown that surgical residents can safely perform a variation of complex abdominal surgeries when provided with adequate training, proper case selection, and appropriate supervision. Their outcomes are equivalent when compared to experienced board-certified surgeons. Our previously published training curriculum for robotic assisted minimally invasive esophagectomy already demonstrated a possible reduction in time to reach proficiency. However, esophagectomy is a technically challenging procedure and comes with high morbidity rates of up to 60%, making it difficult to provide opportunities to train surgical residents. We aimed to investigate if a surgical resident could safely perform complex esophageal surgery when a structured modular teaching curriculum is applied. A structured teaching program based on our previously published modular step-up approach was applied by two experienced board-certified esophageal surgeons. Our IRB-approved (Institutional Review Board) database was searched to identify all Ivor-Lewis esophagectomies performed by the selected surgical resident from August 2019 to July 2021. The cumulative sum method was used to analyze the learning curve of the surgical resident. Outcomes of patients operated by the resident were then compared to our overall cohort of open, hybrid, and robotic Ivor-Lewis esophagectomies from May 2016 to May 2020. The total cohort included 567 patients, of which 65 were operated by the surgical resident and 502 patients were operated by experienced esophageal cancer surgeons as the control group. For baseline characteristics, a significant difference for BMI (Body mass index) was observed, which was lower in the resident's group (25.5 kg/m2 vs. 26.8 kg/m2 (P = 0.046). A significant difference of American Society of Anesthesiologists- and Eastern Cooperative Oncology Group-scores was seen, and a subgroup analysis including all patients with American Society of Anesthesiologists I and Eastern Cooperative Oncology Group 0 was performed revealing no significant differences. Postoperative complications did not differ between groups. The anastomotic leak rate was 13.8% in the resident's cohort and 12% in the control cohort (P = 0.660). Major complications (Clavien-Dindo ≥ IIIb) occurred in 16.9% of patients in both groups. Oncological outcome, defined by harvested lymph nodes (35 vs. 32.33, P = 0.096), proportion of lymph node compliant performed operations (86.2% vs. 88.4%, P = 0.590), and R0-resection rate (96.9% vs. 96%, P = 0.766), was not compromised when esophagectomies were performed by the resident. The resident completed the learning curves after 39 cases for the total operating time, 38 cases for the thoracic operating time, 26 cases for the number of harvested lymph nodes, 29 cases for anastomotic leak rate, and finally 58 cases for the comprehensive complication index. For postoperative complications, no significant difference was seen between patients operated in the resident group versus the control group, with a third of patients being discharged with a textbook outcome in both cohorts. Furthermore, no difference in oncological quality of the resection was found, emphasizing safety and feasibility of our training program. A structured modular step-up for training a surgical resident to perform complex esophageal cancer surgery can successfully maintain patient safety and outcomes.
Asunto(s)
Competencia Clínica , Neoplasias Esofágicas , Esofagectomía , Internado y Residencia , Procedimientos Quirúrgicos Robotizados , Humanos , Internado y Residencia/métodos , Neoplasias Esofágicas/cirugía , Esofagectomía/educación , Esofagectomía/métodos , Esofagectomía/efectos adversos , Femenino , Masculino , Persona de Mediana Edad , Anciano , Procedimientos Quirúrgicos Robotizados/educación , Procedimientos Quirúrgicos Robotizados/efectos adversos , Curva de Aprendizaje , Tutoría/métodos , Curriculum , Hospitales de Alto Volumen , Estudios RetrospectivosRESUMEN
BACKGROUND: Robotic-assisted minimally invasive esophagectomy (RAMIE) was first introduced in 2003 and has since then shown to significantly improve the postoperative course. Previous studies have shown that a structured training pathway based on proficiency-based progression using individual skill levels as measures of reach of competence can enhance surgical performance. The aim of this study was to evaluate and help understand our pathway to reach surgical expert levels using a proficiency-based approach introducing RAMIE at our German high-volume center. METHODS: All patients undergoing RAMIE performed by two experienced surgeons for esophageal cancer since the introduction of the robotic technique in 2017 was included in this analysis. Intraoperative outcomes and postoperative outcomes were included in the analysis. The cumulative sum method was used to analyze how many cases are needed to reach expert levels for different performance characteristics and skill sets during robotic-assisted minimally invasive esophagectomy. RESULTS: From 06/2017 to 03/2022, a total of 154 patients underwent RAMIE at our facility and were included in the analysis. An advancement in performance level was observed for total operating time after 70 cases and for thoracic operative time after 79 cases. Lymph node yield showed an increase up until case 60 in the CUSUM analysis. Length of hospital stay stabilized after case 55. The CCI score inflection point was at case 55 in both CUSUM and regression analyses. Anastomotic leak rate stabilized at case 38 and showed another inflection point after case 83. CONCLUSION: Our data and analysis showed the progression from proficient to expert performance levels during the implementation of RAMIE at a European high-volume center. Further analysis of surgeons, especially with a different training status has yet to reveal if the caseloads found in this study are universally applicable. However, skill acquisition and respective measures of such are diverse and as a great range of number of cases was observed, we believe that the learning curve and ascent in performance levels cannot be defined by one parameter alone.
Asunto(s)
Boehmeria , Neoplasias Esofágicas , Procedimientos Quirúrgicos Robotizados , Humanos , Esofagectomía/métodos , Procedimientos Quirúrgicos Robotizados/métodos , Neoplasias Esofágicas/cirugía , Neoplasias Esofágicas/patología , Ganglios Linfáticos/patología , Procedimientos Quirúrgicos Mínimamente Invasivos/métodos , Resultado del Tratamiento , Complicaciones Posoperatorias/epidemiología , Complicaciones Posoperatorias/etiología , Complicaciones Posoperatorias/patología , Estudios RetrospectivosRESUMEN
BACKGROUND: Surgery generates a vast amount of data from each procedure. Particularly video data provides significant value for surgical research, clinical outcome assessment, quality control, and education. The data lifecycle is influenced by various factors, including data structure, acquisition, storage, and sharing; data use and exploration, and finally data governance, which encompasses all ethical and legal regulations associated with the data. There is a universal need among stakeholders in surgical data science to establish standardized frameworks that address all aspects of this lifecycle to ensure data quality and purpose. METHODS: Working groups were formed, among 48 representatives from academia and industry, including clinicians, computer scientists and industry representatives. These working groups focused on: Data Use, Data Structure, Data Exploration, and Data Governance. After working group and panel discussions, a modified Delphi process was conducted. RESULTS: The resulting Delphi consensus provides conceptualized and structured recommendations for each domain related to surgical video data. We identified the key stakeholders within the data lifecycle and formulated comprehensive, easily understandable, and widely applicable guidelines for data utilization. Standardization of data structure should encompass format and quality, data sources, documentation, metadata, and account for biases within the data. To foster scientific data exploration, datasets should reflect diversity and remain adaptable to future applications. Data governance must be transparent to all stakeholders, addressing legal and ethical considerations surrounding the data. CONCLUSION: This consensus presents essential recommendations around the generation of standardized and diverse surgical video databanks, accounting for multiple stakeholders involved in data generation and use throughout its lifecycle. Following the SAGES annotation framework, we lay the foundation for standardization of data use, structure, and exploration. A detailed exploration of requirements for adequate data governance will follow.
Asunto(s)
Inteligencia Artificial , Mejoramiento de la Calidad , Humanos , Consenso , Recolección de DatosRESUMEN
BACKGROUND: Artificial intelligence (AI) holds tremendous potential to reduce surgical risks and improve surgical assessment. Machine learning, a subfield of AI, can be used to analyze surgical video and imaging data. Manual annotations provide veracity about the desired target features. Yet, methodological annotation explorations are limited to date. Here, we provide an exploratory analysis of the requirements and methods of instrument annotation in a multi-institutional team from two specialized AI centers and compile our lessons learned. METHODS: We developed a bottom-up approach for team annotation of robotic instruments in robot-assisted partial nephrectomy (RAPN), which was subsequently validated in robot-assisted minimally invasive esophagectomy (RAMIE). Furthermore, instrument annotation methods were evaluated for their use in Machine Learning algorithms. Overall, we evaluated the efficiency and transferability of the proposed team approach and quantified performance metrics (e.g., time per frame required for each annotation modality) between RAPN and RAMIE. RESULTS: We found a 0.05 Hz image sampling frequency to be adequate for instrument annotation. The bottom-up approach in annotation training and management resulted in accurate annotations and demonstrated efficiency in annotating large datasets. The proposed annotation methodology was transferrable between both RAPN and RAMIE. The average annotation time for RAPN pixel annotation ranged from 4.49 to 12.6 min per image; for vector annotation, we denote 2.92 min per image. Similar annotation times were found for RAMIE. Lastly, we elaborate on common pitfalls encountered throughout the annotation process. CONCLUSIONS: We propose a successful bottom-up approach for annotator team composition, applicable to any surgical annotation project. Our results set the foundation to start AI projects for instrument detection, segmentation, and pose estimation. Due to the immense annotation burden resulting from spatial instrumental annotation, further analysis into sampling frequency and annotation detail needs to be conducted.
Asunto(s)
Laparoscopía , Procedimientos Quirúrgicos Robotizados , Robótica , Humanos , Procedimientos Quirúrgicos Robotizados/métodos , Inteligencia Artificial , Nefrectomía/métodosRESUMEN
BACKGROUND: Caroli Disease (CD) and Caroli Syndrome (CS) are rare disorders presenting with dilation of the intrahepatic bile ducts. CD/CS are associated with cholangiocarcinoma (CCA). However, the true incidence of CCA is still unclear, although it may serve as an indication for surgery. In this paper, we analyzed (I) the incidence of CCA in German centers, (II) reviewed our single center population together with its clinical presentation and (III) performed a thorough literature review. METHODS: 17 large HPB-centers across Germany were contacted and their patients after surgical treatment due to CD/CS with histopathology were included. Medline search for all studies published in English or German literature was performed. Patients who underwent surgery at our department between 2012 and 2020 due to CD or CS were analyzed. RESULTS: In the multicenter study, 79 patients suffered from CD and 119 patients from CS, with a total number of 198 patients. In 14 patients, CCA was found (Overall: 7,1%; CD: 6,3%, CS 7,6%). Between 2012 and 2020, 1661 liver resections were performed at our department. 14 patients underwent surgery due to CD or CS. Histological examination showed synchronous cholangiocarcinoma in one patient. The literature review revealed a CCA-rate of 7,3% in large series, whereas in case reports a rate of 6,8% was found. CONCLUSION: There is risk of malignant transformation and patients with CD might also benefit from resection due to improvement of symptoms. Therefore, resection is strongly advised. As certain patients with CS require transplantation, treatment should not be guided by the relatively low rate of CCA but by the concomitant diseases that come along with hepatic failure.
Asunto(s)
Neoplasias de los Conductos Biliares , Enfermedad de Caroli , Colangiocarcinoma , Neoplasias de los Conductos Biliares/diagnóstico , Neoplasias de los Conductos Biliares/epidemiología , Neoplasias de los Conductos Biliares/cirugía , Conductos Biliares Intrahepáticos/patología , Enfermedad de Caroli/complicaciones , Enfermedad de Caroli/epidemiología , Enfermedad de Caroli/cirugía , Colangiocarcinoma/diagnóstico , Colangiocarcinoma/epidemiología , Colangiocarcinoma/cirugía , Hepatectomía/efectos adversos , HumanosRESUMEN
Analysis of relations between objects and comprehension of abstract concepts in the surgical video is important in AI-augmented surgery. However, building models that integrate our knowledge and understanding of surgery remains a challenging endeavor. In this paper, we propose a novel way to integrate conceptual knowledge into temporal analysis tasks using temporal concept graph networks. In the proposed networks, a knowledge graph is incorporated into the temporal video analysis of surgical notions, learning the meaning of concepts and relations as they apply to the data. We demonstrate results in surgical video data for tasks such as verification of the critical view of safety, estimation of the Parkland grading scale as well as recognizing instrument-action-tissue triplets. The results show that our method improves the recognition and detection of complex benchmarks as well as enables other analytic applications of interest.
Asunto(s)
Redes Neurales de la Computación , Procedimientos Quirúrgicos Operativos , Grabación en VideoRESUMEN
BACKGROUND: Postesophagectomy diaphragmatic prolapse (PDP) is a major complication after esophagectomy with significant mortality and morbidity. However, in the current literature, treatment and outcomes are not evaluated for patients undergoing an Ivor Lewis Robot-assisted minimally invasive esophagectomy (IL-RAMIE). The aim of this study is to evaluate the incidence of PDP after IL-RAMIE. Moreover, the study aims to determine whether using a minimally invasive approach in the management of PDP after an IL-RAMIE procedure is safe and feasible. MATERIALS AND METHODS: This study includes all patients who received an IL-RAMIE at our high-volume center (>200 esophagectomies/year) between April 2017 and December 2022 and developed PDP. The analysis focuses on time to prolapse, symptoms, treatment, surgical method, and recurrence rates of these patients. RESULTS: A total of 185 patients underwent an IL-RAMIE at our hospital. Eleven patients (5.9%) developed PDP. Patients presented with PDP after a medium time of 241 days with symptoms like reflux, nausea, vomiting, and pain. One-third of these patients did not suffer from any symptoms. In all cases, a CT scan was performed in which the colon transversum always presented as the herniated organ. In one patient, prolapse of the small intestine, pancreas, and greater omentum also occurred. A total of 91% of these patients received a revisional surgery in a minimally invasive manner with a mean hospital stay of 12 days. In four patients, PDP recurred (36%) after 13, 114, 119 and 237 days, respectively. CONCLUSION: This study shows that a minimally invasive approach in repositioning PDP is a safe and effective option after IL-RAMIE.
RESUMEN
While the sentinel lymph node concept is routinely applied in other surgical fields, no established and valid modality for lymph node mapping for esophageal cancer surgery currently exists. Near-infrared light fluorescence (NIR) using indocyanine green (ICG) has been recently proven to be a safe technology for peritumoral injection and consecutive lymph node mapping in small surgical cohorts, mostly without the usage of robotic technology. The aim of this study was to identify the lymphatic drainage pattern of esophageal cancer during highly standardized RAMIE and to correlate the intraoperative images with the histopathological dissemination of lymphatic metastases. Patients with clinically advanced stage squamous cell carcinoma or adenocarcinoma of the esophagus undergoing a RAMIE at our Center of Excellence for Surgery of the Upper Gastrointestinal Tract were prospectively included in this study. Patients were admitted on the day prior to surgery, and an additional EGD with endoscopic injection of the ICG solution around the tumor was performed. Intraoperative imaging procedures were performed using the Stryker 1688 or the FIREFLY fluorescence imaging system, and resected lymph nodes were sent to pathology. A total of 20 patients were included in the study, and feasibility and safety for the application of NIR using ICG during RAMIE were shown. NIR imaging to detect lymph node metastases can be safely performed during RAMIE. Further analyses in our center will focus on pathological analyses of ICG-positive tissue and quantification using artificial intelligence tools with a correlation of long-term follow-up data.