Infectious Mediastinitis: A Retrospective Cohort Study.

Background: This study aimed to characterize the demographics, microbiology, management and treatment outcomes of mediastinitis according to the origin of the infection. Methods: This retrospective observational study enrolled patients who had mediastinitis diagnosed according to the criteria defined by the Centers for Disease Control and Prevention and were treated in Strasbourg University Hospital, France, between 1 January 2010 and 31 December 2020. Results: We investigated 151 cases, including 63 cases of poststernotomy mediastinitis (PSM), 60 cases of mediastinitis due to esophageal perforation (MEP) and 17 cases of descending necrotizing mediastinitis (DNM). The mean patient age (standard deviation) was 63 (14.5) years, and 109 of 151 patients were male. Microbiological documentation varied according to the origin of the infection. When documented, PSM cases were mostly monomicrobial (36 of 53 cases [67.9%]) and involved staphylococci (36 of 53 [67.9%]), whereas MEP and DNM cases were mostly plurimicrobial (38 of 48 [79.2%] and 8 of 12 [66.7%], respectively) and involved digestive or oral flora microorganisms, respectively. The median duration of anti-infective treatment was 41 days (interquartile range, 21-56 days), and 122 of 151 patients (80.8%) benefited from early surgical management. The overall 1-year survival rate was estimated to be 64.8% (95% confidence interval, 56.6%-74.3%), but varied from 80.1% for DNM to 61.5% for MEP. Conclusions: Mediastinitis represents a rare yet deadly infection. The present cohort study exhibited the different patterns observed according to the origin of the infection. Greater insight and knowledge on these differences may help guide the management of these complex infections, especially with respect to empirical anti-infective treatments.

Challenges in multi-centric generalization: phase and step recognition in Roux-en-Y gastric bypass surgery.

PURPOSE: Most studies on surgical activity recognition utilizing artificial intelligence (AI) have focused mainly on recognizing one type of activity from small and mono-centric surgical video datasets. It remains speculative whether those models would generalize to other centers. METHODS: In this work, we introduce a large multi-centric multi-activity dataset consisting of 140 surgical videos (MultiBypass140) of laparoscopic Roux-en-Y gastric bypass (LRYGB) surgeries performed at two medical centers, i.e., the University Hospital of Strasbourg, France (StrasBypass70) and Inselspital, Bern University Hospital, Switzerland (BernBypass70). The dataset has been fully annotated with phases and steps by two board-certified surgeons. Furthermore, we assess the generalizability and benchmark different deep learning models for the task of phase and step recognition in 7 experimental studies: (1) Training and evaluation on BernBypass70; (2) Training and evaluation on StrasBypass70; (3) Training and evaluation on the joint MultiBypass140 dataset; (4) Training on BernBypass70, evaluation on StrasBypass70; (5) Training on StrasBypass70, evaluation on BernBypass70; Training on MultiBypass140, (6) evaluation on BernBypass70 and (7) evaluation on StrasBypass70. RESULTS: The model's performance is markedly influenced by the training data. The worst results were obtained in experiments (4) and (5) confirming the limited generalization capabilities of models trained on mono-centric data. The use of multi-centric training data, experiments (6) and (7), improves the generalization capabilities of the models, bringing them beyond the level of independent mono-centric training and validation (experiments (1) and (2)). CONCLUSION: MultiBypass140 shows considerable variation in surgical technique and workflow of LRYGB procedures between centers. Therefore, generalization experiments demonstrate a remarkable difference in model performance. These results highlight the importance of multi-centric datasets for AI model generalization to account for variance in surgical technique and workflows. The dataset and code are publicly available at https://github.com/CAMMA-public/MultiBypass140.

Prognostic impact of positive microscopic margins (R1 resection) in patients with GIST (gastrointestinal stromal tumours): Results of a multicenter European study.

BACKGROUND: Although several prognostic factors in GIST have been well studied such as tumour size, mitotic rate, or localization, the influence of microscopic margins or R1 resection remains controversial. The aim of this study was to evaluate the influence of R1 resection on the prognosis of GIST in a large multicentre retrospective series of patients. METHODS: From 2001 to 2013, 1413 patients who underwent surgery for any site of GIST were identified from 61 European centers. 1098 patients were included, excluding synchronous metastases, concurrent malignancies, R2 resection or GIST recurrence. Tumour rupture (TR) was reclassified according to the Oslo sarcoma classification. Cox proportional hazards ratio and Kaplan-Meier survival estimates were used to analyse 5-year recurrence-free survival (RFS). RESULTS: Of 1098 patients, 38 (3%) underwent R1 resection with a risk of TR of 11%. The 5-year RFS was 89.6% with a median follow-up of 81 months [range: 31.2-152 months]. On univariate analysis, lower RFS was significantly associated with R1 resection [HR = 2.13; p = 0.04], high risk score according to the modified NIH classification, administration of adjuvant therapy [HR = 2.24; p < 0.001] and intraoperative complications [HR = 2.82; p < 0.001]. Only intraoperative complications [HR = 1.79; p = 0.02] and high risk according to the modified NIH classification including the updated definition of TR [HR = 3.43; p = 0.04] remained significant on multivariate analysis. CONCLUSION: This study shows that positive microscopic margins are not an independent predictive factor for RFS in GIST when taking into account the up-dated classification of TR. R1 resection may be considered a reasonable alternative to avoid major functional sequelae and should not lead to reoperation.

Early-stage clinical evaluation of real-time artificial intelligence assistance for laparoscopic cholecystectomy.

The growing availability of surgical digital data and developments in analytics such as artificial intelligence (AI) are being harnessed to improve surgical care. However, technical and cultural barriers to real-time intraoperative AI assistance exist. This early-stage clinical evaluation shows the technical feasibility of concurrently deploying several AIs in operating rooms for real-time assistance during procedures. In addition, potentially relevant clinical applications of these AI models are explored with a multidisciplinary cohort of key stakeholders.

Latent Graph Representations for Critical View of Safety Assessment.

Assessing the critical view of safety in laparoscopic cholecystectomy requires accurate identification and localization of key anatomical structures, reasoning about their geometric relationships to one another, and determining the quality of their exposure. Prior works have approached this task by including semantic segmentation as an intermediate step, using predicted segmentation masks to then predict the CVS. While these methods are effective, they rely on extremely expensive ground-truth segmentation annotations and tend to fail when the predicted segmentation is incorrect, limiting generalization. In this work, we propose a method for CVS prediction wherein we first represent a surgical image using a disentangled latent scene graph, then process this representation using a graph neural network. Our graph representations explicitly encode semantic information - object location, class information, geometric relations - to improve anatomy-driven reasoning, as well as visual features to retain differentiability and thereby provide robustness to semantic errors. Finally, to address annotation cost, we propose to train our method using only bounding box annotations, incorporating an auxiliary image reconstruction objective to learn fine-grained object boundaries. We show that our method not only outperforms several baseline methods when trained with bounding box annotations, but also scales effectively when trained with segmentation masks, maintaining state-of-the-art performance.

Endoscopic ultrasound-guided perivascular pancreatic radiofrequency ablation using a hydroxyethyl starch solution prior to pancreatectomy.

Background and study aims Pancreatic surgery remains complex, particularly for borderline resectable and locally advanced tumors. Vascular invasion compromises resectability, and vascular resection entails increased morbidity and mortality. Following a feasibility and safety demonstration of augmented endoscopic ultrasound (EUS)-guided radiofrequency ablation (RFA) using hydroxyethyl starch (HES) in porcine pancreatic parenchyma, the present study assesses whether this approach (EUS-sugar-RFA) in the pancreatic perivascular space is safe and creates a controllable margin of necrosis to enable a vessel-sparing resection. Methods EUS-sugar-RFA in the pancreatic parenchyma adjacent to the splenic artery and vein was performed in a live animal model. Following different survival periods (0-4 days) in the interventional group (n = 3), open pancreatectomy was carried out. The control group (n = 4) included open pancreatectomies in two pigs with non-treated pancreases and in two with pancreatic RFA alone on the same day. Results All procedures were completed successfully, without intraoperative or postoperative complications. Survival periods were uncomplicated. Histopathological examination showed local necrosis and inflammatory reaction at the ablation sites. Vascular wall integrity was preserved in all specimens. The untreated pancreatic zones in the interventional group were no different from the normal pancreases in the control group. Conclusions Preoperative perivascular augmented RFA using HES was safe, and in the pancreatic animal model, the best timeframe was within 24 hours before pancreatic surgery. This technique might improve resectability in selected borderline and locally advanced pancreatic cancers.

Risk factors for in-hospital mortality after emergency colorectal surgery in octogenarians: results of a cohort study from a referral center.

PURPOSE: The objective of this study was to investigate predictive factors of mortality in emergency colorectal surgery in octogenarian patients. METHODS: It is a retrospective cohort study conducted at a single-institution tertiary referral center. Consecutive patients who underwent emergency colorectal surgery between January 2015 and January 2020 were identified. The primary endpoint was 30-day mortality. Univariate and multivariate analyses were performed using a logistic regression model. RESULTS: A total of 111 patients were identified (43 men, 68 women). Mean age was 85.7 ± 3.7 years (80-96). Main diagnoses included complicated sigmoiditis in 38 patients (34.3%), cancer in 35 patients (31.5%), and ischemic colitis in 31 patients (27.9%). An ASA score of 3 or higher was observed in 88.3% of patients. The mean Charlson score was 5.9. The Possum score was 35.9% for mortality and 79.3% for morbidity. The 30-day mortality rate was 25.2%. Univariate analysis of preoperative risk factors for mortality shows that the history of valvular heart disease (p = 0.008), intensive care unit provenance (p = 0.003), preoperative sepsis (p < 0.001), diagnosis of ischemic colitis (p = 0.012), creatinine (p = 0.006) and lactate levels (p = 0.01) were significantly associated with 30-day mortality, and patients coming from home had a lower 30-day mortality rate (p = 0.018). Intraoperative variables associated with 30-day mortality included ileostomy creation (p = 0.022) and temporary laparostomy (p = 0.004). At multivariate analysis, only lactate (p = 0.032) and creatinine levels (p = 0.027) were found to be independent predictors of 30-day mortality, home provenance was an independent protective factor (p = 0.004). Mean follow-up was 3.4 years. Survival at 1 and 3 years was 57.6 and 47.7%. CONCLUSION: Emergency colorectal surgery is challenging. However, age should not be a contraindication. The 30-day mortality rate (25.2%) is one of the lowest in the literature. Hyperlactatemia (> 2mmol/L) and creatinine levels appear to be independent predictors of mortality.

CholecTriplet2022: Show me a tool and tell me the triplet - An endoscopic vision challenge for surgical action triplet detection.

Formalizing surgical activities as triplets of the used instruments, actions performed, and target anatomies is becoming a gold standard approach for surgical activity modeling. The benefit is that this formalization helps to obtain a more detailed understanding of tool-tissue interaction which can be used to develop better Artificial Intelligence assistance for image-guided surgery. Earlier efforts and the CholecTriplet challenge introduced in 2021 have put together techniques aimed at recognizing these triplets from surgical footage. Estimating also the spatial locations of the triplets would offer a more precise intraoperative context-aware decision support for computer-assisted intervention. This paper presents the CholecTriplet2022 challenge, which extends surgical action triplet modeling from recognition to detection. It includes weakly-supervised bounding box localization of every visible surgical instrument (or tool), as the key actors, and the modeling of each tool-activity in the form of triplet. The paper describes a baseline method and 10 new deep learning algorithms presented at the challenge to solve the task. It also provides thorough methodological comparisons of the methods, an in-depth analysis of the obtained results across multiple metrics, visual and procedural challenges; their significance, and useful insights for future research directions and applications in surgery.

Preserving privacy in surgical video analysis using a deep learning classifier to identify out-of-body scenes in endoscopic videos.

Surgical video analysis facilitates education and research. However, video recordings of endoscopic surgeries can contain privacy-sensitive information, especially if the endoscopic camera is moved out of the body of patients and out-of-body scenes are recorded. Therefore, identification of out-of-body scenes in endoscopic videos is of major importance to preserve the privacy of patients and operating room staff. This study developed and validated a deep learning model for the identification of out-of-body images in endoscopic videos. The model was trained and evaluated on an internal dataset of 12 different types of laparoscopic and robotic surgeries and was externally validated on two independent multicentric test datasets of laparoscopic gastric bypass and cholecystectomy surgeries. Model performance was evaluated compared to human ground truth annotations measuring the receiver operating characteristic area under the curve (ROC AUC). The internal dataset consisting of 356,267 images from 48 videos and the two multicentric test datasets consisting of 54,385 and 58,349 images from 10 and 20 videos, respectively, were annotated. The model identified out-of-body images with 99.97% ROC AUC on the internal test dataset. Mean ± standard deviation ROC AUC on the multicentric gastric bypass dataset was 99.94 ± 0.07% and 99.71 ± 0.40% on the multicentric cholecystectomy dataset, respectively. The model can reliably identify out-of-body images in endoscopic videos and is publicly shared. This facilitates privacy preservation in surgical video analysis.

Deep sedation versus orotracheal intubation for endoscopic sleeve gastroplasty (ESG): preliminary experience.

BACKGROUND: Endoscopic sleeve gastroplasty (ESG) is an emerging bariatric procedure currently performed under general anaesthesia with orotracheal intubation (OTI). Several studies have shown the feasibility of advanced endoscopic procedures under deep sedation (DS) without impacting patient outcomes or adverse event rates. Our goal was to perform an initial comparative analysis of ESG in DS with ESG under OTI. METHODS: A prospective institutional registry was reviewed for ESG patients between 12/2016 and 1/2021. Patients were stratified into OTI or DS cohorts, and the 1st 50 cases performed in each cohort were included for comparability. Univariate analysis was performed on demographics, intraoperative, and postoperative outcomes (up to 90 days). Multivariate analyses evaluated the relationship between anesthesia type, preclinical and clinical variables. RESULTS: Of the 50 DS patients, 21(42%) underwent primary and 29 (58%) revisional surgery. There was no significant differences in Mallampati score across groups. No DS patient required intubation. DS patients were younger (p = 0.006) and lower BMI (p = 0.002) than OTI. As expected, DS patients overall and in the primary subgroup had shorter operative time (p ≤ 0.001 and p = 0.003, respectively) and higher rates (84% DS vs. 20% OTI, p ≤ 0.001) of ambulatory procedures. There were no significant differences in the sutures used between groups (p = 0.616). DS patients required less postoperative opioids (p ≤ 0.001) and antiemetics (p = 0.006) than OTI. There were no significant differences in 3-month postoperative weight loss across cohorts. There was no rehospitalization in either group. In primary ESG cases, we found DS patients were more likely younger (p = 0.006), female (p = 0.001), and had a lower BMI (p = 0.0027). CONCLUSIONS: ESG under DS is safe and feasible in select patients. We found DS safely increased rates of outpatient care, reduced use of opioids and antiemetics, and provided the same results of postoperative weight loss. Patient selection for DS may be more clearer for durable weight loss.

Live laparoscopic video retrieval with compressed uncertainty.

Searching through large volumes of medical data to retrieve relevant information is a challenging yet crucial task for clinical care. However the primitive and most common approach to retrieval, involving text in the form of keywords, is severely limited when dealing with complex media formats. Content-based retrieval offers a way to overcome this limitation, by using rich media as the query itself. Surgical video-to-video retrieval in particular is a new and largely unexplored research problem with high clinical value, especially in the real-time case: using real-time video hashing, search can be achieved directly inside of the operating room. Indeed, the process of hashing converts large data entries into compact binary arrays or hashes, enabling large-scale search operations at a very fast rate. However, due to fluctuations over the course of a video, not all bits in a given hash are equally reliable. In this work, we propose a method capable of mitigating this uncertainty while maintaining a light computational footprint. We present superior retrieval results (3%-4% top 10 mean average precision) on a multi-task evaluation protocol for surgery, using cholecystectomy phases, bypass phases, and coming from an entirely new dataset introduced here, surgical events across six different surgery types. Success on this multi-task benchmark shows the generalizability of our approach for surgical video retrieval.

Rendezvous in time: an attention-based temporal fusion approach for surgical triplet recognition.

PURPOSE: One of the recent advances in surgical AI is the recognition of surgical activities as triplets of [Formula: see text]instrument, verb, target[Formula: see text]. Albeit providing detailed information for computer-assisted intervention, current triplet recognition approaches rely only on single-frame features. Exploiting the temporal cues from earlier frames would improve the recognition of surgical action triplets from videos. METHODS: In this paper, we propose Rendezvous in Time (RiT)-a deep learning model that extends the state-of-the-art model, Rendezvous, with temporal modeling. Focusing more on the verbs, our RiT explores the connectedness of current and past frames to learn temporal attention-based features for enhanced triplet recognition. RESULTS: We validate our proposal on the challenging surgical triplet dataset, CholecT45, demonstrating an improved recognition of the verb and triplet along with other interactions involving the verb such as [Formula: see text]instrument, verb[Formula: see text]. Qualitative results show that the RiT produces smoother predictions for most triplet instances than the state-of-the-arts. CONCLUSION: We present a novel attention-based approach that leverages the temporal fusion of video frames to model the evolution of surgical actions and exploit their benefits for surgical triplet recognition.

Weakly Supervised Temporal Convolutional Networks for Fine-Grained Surgical Activity Recognition.

Automatic recognition of fine-grained surgical activities, called steps, is a challenging but crucial task for intelligent intra-operative computer assistance. The development of current vision-based activity recognition methods relies heavily on a high volume of manually annotated data. This data is difficult and time-consuming to generate and requires domain-specific knowledge. In this work, we propose to use coarser and easier-to-annotate activity labels, namely phases, as weak supervision to learn step recognition with fewer step annotated videos. We introduce a step-phase dependency loss to exploit the weak supervision signal. We then employ a Single-Stage Temporal Convolutional Network (SS-TCN) with a ResNet-50 backbone, trained in an end-to-end fashion from weakly annotated videos, for temporal activity segmentation and recognition. We extensively evaluate and show the effectiveness of the proposed method on a large video dataset consisting of 40 laparoscopic gastric bypass procedures and the public benchmark CATARACTS containing 50 cataract surgeries.

In Vitro Antibody Quantification with Hyperspectral Imaging in a Large Field of View for Clinical Applications.

Hyperspectral imaging (HSI) is a non-invasive, contrast-free optical-based tool that has recently been applied in medical and basic research fields. The opportunity to use HSI to identify exogenous tumor markers in a large field of view (LFOV) could increase precision in oncological diagnosis and surgical treatment. In this study, the anti-high mobility group B1 (HMGB1) labeled with Alexa fluorophore (647 nm) was used as the target molecule. This is the proof-of-concept of HSI's ability to quantify antibodies via an in vitro setting. A first test was performed to understand whether the relative absorbance provided by the HSI camera was dependent on volume at a 1:1 concentration. A serial dilution of 1:1, 10, 100, 1000, and 10,000 with phosphatase-buffered saline (PBS) was then used to test the sensitivity of the camera at the minimum and maximum volumes. For the analysis, images at 640 nm were extracted from the hypercubes according to peak signals matching the specificities of the antibody manufacturer. The results showed a positive correlation between relative absorbance and volume (r = 0.9709, p = 0.0013). The correlation between concentration and relative absorbance at min (1 µL) and max (20 µL) volume showed r = 0.9925, p < 0.0001, and r = 0.9992, p < 0.0001, respectively. These results demonstrate the HSI potential in quantifying HMGB1, hence deserving further studies in ex vivo and in vivo settings.

Educational Scoring System in Laparoscopic Cholecystectomy: Is It the Right Time to Standardize?

Background and Objectives: Laparoscopic cholecystectomy (LC) is one of the most performed surgeries worldwide. Procedure difficulty and patient outcomes depend on several factors which are not considered in the current literature, including the learning curve, generating confusing and subjective results. This study aims to create a scoring system to calculate the learning curve of LC based on hepatobiliopancreatic (HPB) experts' opinions during an educational course. Materials and Methods: A questionnaire was submitted to the panel of experts attending the HPB course at Research Institute against Digestive Cancer-IRCAD (Strasbourg, France) from 27-29 October 2022. Experts scored the proposed variables according to their degree of importance in the learning curve using a Likert scale from 1 (not useful) to 5 (very useful). Variables were included in the composite scoring system only if more than 75% of experts ranked its relevance in the learning curve assessment ≥4. A positive or negative value was assigned to each variable based on its effect on the learning curve. Results: Fifteen experts from six different countries attended the IRCAD HPB course and filled out the questionnaire. Ten variables were finally included in the learning curve scoring system (i.e., patient body weight/BMI, patient previous open surgery, emergency setting, increased inflammatory levels, presence of anatomical bile duct variation(s), and appropriate critical view of safety (CVS) identification), which were all assigned positive values. Minor or major intraoperative injuries to the biliary tract, development of postoperative complications related to biliary injuries, and mortality were assigned negative values. Conclusions: This is the first scoring system on the learning curve of LC based on variables selected through the experts' opinions. Although the score needs to be validated through future studies, it could be a useful tool to assess its efficacy within educational programs and surgical courses.

TRandAugment: temporal random augmentation strategy for surgical activity recognition from videos.

PURPOSE: Automatic recognition of surgical activities from intraoperative surgical videos is crucial for developing intelligent support systems for computer-assisted interventions. Current state-of-the-art recognition methods are based on deep learning where data augmentation has shown the potential to improve the generalization of these methods. This has spurred work on automated and simplified augmentation strategies for image classification and object detection on datasets of still images. Extending such augmentation methods to videos is not straightforward, as the temporal dimension needs to be considered. Furthermore, surgical videos pose additional challenges as they are composed of multiple, interconnected, and long-duration activities. METHODS: This work proposes a new simplified augmentation method, called TRandAugment, specifically designed for long surgical videos, that treats each video as an assemble of temporal segments and applies consistent but random transformations to each segment. The proposed augmentation method is used to train an end-to-end spatiotemporal model consisting of a CNN (ResNet50) followed by a TCN. RESULTS: The effectiveness of the proposed method is demonstrated on two surgical video datasets, namely Bypass40 and CATARACTS, and two tasks, surgical phase and step recognition. TRandAugment adds a performance boost of 1-6% over previous state-of-the-art methods, that uses manually designed augmentations. CONCLUSION: This work presents a simplified and automated augmentation method for long surgical videos. The proposed method has been validated on different datasets and tasks indicating the importance of devising temporal augmentation methods for long surgical videos.

Surgical Models of Liver Regeneration in Pigs: A Practical Review of the Literature for Researchers.

The remarkable capacity of regeneration of the liver is well known, although the involved mechanisms are far from being understood. Furthermore, limits concerning the residual functional mass of the liver remain critical in both fields of hepatic resection and transplantation. The aim of the present study was to review the surgical experiments regarding liver regeneration in pigs to promote experimental methodological standardization. The Pubmed, Medline, Scopus, and Cochrane Library databases were searched. Studies evaluating liver regeneration through surgical experiments performed on pigs were included. A total of 139 titles were screened, and 41 articles were included in the study, with 689 pigs in total. A total of 29 studies (71% of all) had a survival design, with an average study duration of 13 days. Overall, 36 studies (88%) considered partial hepatectomy, of which four were an associating liver partition and portal vein ligation for staged hepatectomy (ALPPS). Remnant liver volume ranged from 10% to 60%. Only 2 studies considered a hepatotoxic pre-treatment, while 25 studies evaluated additional liver procedures, such as stem cell application, ischemia/reperfusion injury, portal vein modulation, liver scaffold application, bio-artificial, and pharmacological liver treatment. Only nine authors analysed how cytokines and growth factors changed in response to liver resection. The most used imaging system to evaluate liver volume was CT-scan volumetry, even if performed only by nine authors. The pig represents one of the best animal models for the study of liver regeneration. However, it remains a mostly unexplored field due to the lack of experiments reproducing the chronic pathological aspects of the liver and the heterogeneity of existing studies.

Towards automatic verification of the critical view of the myopectineal orifice with artificial intelligence.

BACKGROUND: Visualization of key anatomical landmarks is required during surgical Trans Abdominal Pre Peritoneal repair (TAPP) of inguinal hernia. The Critical View of the MyoPectineal Orifice (CVMPO) was proposed to ensure correct dissection. An artificial intelligence (AI) system that automatically validates the presence of key and marks during the procedure is a critical step towards automatic dissection quality assessment and video-based competency evaluation. The aim of this study was to develop an AI system that automatically recognizes the TAPP key CVMPO landmarks in hernia repair videos. METHODS: Surgical videos of 160 TAPP procedures were used in this single-center study. A deep neural network-based object detector was developed to automatically recognize the pubic symphysis, direct hernia orifice, Cooper's ligament, the iliac vein, triangle of Doom, deep inguinal ring, and iliopsoas muscle. The system was trained using 130 videos, annotated and verified by two board-certified surgeons. Performance was evaluated in 30 videos of new patients excluded from the training data. RESULTS: Performance was validated in 2 ways: first, single-image validation where the AI model detected landmarks in a single laparoscopic image (mean average precision (MAP) of 51.2%). The second validation is video evaluation where the model detected landmarks throughout the myopectineal orifice visual inspection phase (mean accuracy and F-score of 77.1 and 75.4% respectively). Annotation objectivity was assessed between 2 surgeons in video evaluation, showing a high agreement of 88.3%. CONCLUSION: This study establishes the first AI-based automated recognition of critical structures in TAPP surgical videos, and a major step towards automatic CVMPO validation with AI. Strong performance was achieved in the video evaluation. The high inter-rater agreement confirms annotation quality and task objectivity.

Proposal and multicentric validation of a laparoscopic Roux-en-Y gastric bypass surgery ontology.

BACKGROUND: Phase and step annotation in surgical videos is a prerequisite for surgical scene understanding and for downstream tasks like intraoperative feedback or assistance. However, most ontologies are applied on small monocentric datasets and lack external validation. To overcome these limitations an ontology for phases and steps of laparoscopic Roux-en-Y gastric bypass (LRYGB) is proposed and validated on a multicentric dataset in terms of inter- and intra-rater reliability (inter-/intra-RR). METHODS: The proposed LRYGB ontology consists of 12 phase and 46 step definitions that are hierarchically structured. Two board certified surgeons (raters) with > 10 years of clinical experience applied the proposed ontology on two datasets: (1) StraBypass40 consists of 40 LRYGB videos from Nouvel Hôpital Civil, Strasbourg, France and (2) BernBypass70 consists of 70 LRYGB videos from Inselspital, Bern University Hospital, Bern, Switzerland. To assess inter-RR the two raters' annotations of ten randomly chosen videos from StraBypass40 and BernBypass70 each, were compared. To assess intra-RR ten randomly chosen videos were annotated twice by the same rater and annotations were compared. Inter-RR was calculated using Cohen's kappa. Additionally, for inter- and intra-RR accuracy, precision, recall, F1-score, and application dependent metrics were applied. RESULTS: The mean ± SD video duration was 108 ± 33 min and 75 ± 21 min in StraBypass40 and BernBypass70, respectively. The proposed ontology shows an inter-RR of 96.8 ± 2.7% for phases and 85.4 ± 6.0% for steps on StraBypass40 and 94.9 ± 5.8% for phases and 76.1 ± 13.9% for steps on BernBypass70. The overall Cohen's kappa of inter-RR was 95.9 ± 4.3% for phases and 80.8 ± 10.0% for steps. Intra-RR showed an accuracy of 98.4 ± 1.1% for phases and 88.1 ± 8.1% for steps. CONCLUSION: The proposed ontology shows an excellent inter- and intra-RR and should therefore be implemented routinely in phase and step annotation of LRYGB.