Search

Objective To provide an overview of ML models and data streams utilized for automated surgical phase recognition. Background Phase recognition identifies different steps and phases of an operation. ML is an evolving technology that allows analysis and interpretation of huge data sets. Automation of phase recognition based on data inputs is essential for optimization of workflow, surgical training, intraoperative assistance, patient safety, and efficiency. Methods A systematic review was performed according to the Cochrane recommendations and the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement. PubMed, Web of Science, IEEExplore, GoogleScholar, and CiteSeerX were searched. Literature describing phase recognition based on ML models and the capture of intraoperative signals during general surgery procedures was included. Results A total of 2254 titles/abstracts were screened, and 35 full-texts were included. Most commonly used ML models were Hidden Markov Models and Artificial Neural Networks with a trend towards higher complexity over time. Most frequently used data types were feature learning from surgical videos and manual annotation of instrument use. Laparoscopic cholecystectomy was used most commonly, often achieving accuracy rates over 90%, though there was no consistent standardization of defined phases. Conclusions ML for surgical phase recognition can be performed with high accuracy, depending on the model, data type, and complexity of surgery. Different intraoperative data inputs such as video and instrument type can successfully be used. Most ML models still require significant amounts of manual expert annotations for training. The ML models may drive surgical workflow towards standardization, efficiency, and objectiveness to improve patient outcome in the future. Registration prospero CRD42018108907.

Background:. Minimally invasive pancreatic surgery (MIPS) has developed over the last 3 decades and is nowadays experiencing an increased interest from the surgical community. With increasing awareness of both the public and the surgical community on patient safety, optimization of training has gained importance. For implementation of MIPS we propose 3 training phases. The first phase focuses on developing basic skills and procedure specific skills with the help of simulation, biotissue drills, video libraries, live case observations, and training courses. The second phase consists of index procedures, fellowships, and proctoring programs to ensure patient safety during the first procedures. During the third phase the surgeons aim is to safely implement the procedure into standard practice while minimizing learning curve related excess morbidity and mortality. Case selection, skills assessment, feedback, and mentoring are important methods to optimize this phase. The residual learning curve can reach up to 100 cases depending on the surgeons’ previous experience, selection of cases, and definition of the parameters used to assess the learning curve. Adequate training and high procedural volume are key to implementing MIPS safely.

To compare perioperative outcomes of laparoscopic pancreaticoduodenectomy (LPD) to open pancreaticoduodenectomy (OPD) using evidence from randomized controlled trials (RCTs).LPD is used more commonly, but this surge is mostly based on observational data.We searched CENTRAL, Medline and Web of Science for RCTs comparing minimally invasive to OPD for adults with benign or malignant disease requiring elective pancreaticoduodenectomy. Main outcomes were 90-day mortality, Clavien-Dindo ≥3 complications, and length of hospital stay (LOS). Secondary outcomes were postoperative pancreatic fistula (POPF), delayed gastric emptying (DGE), postpancreatectomy hemorrhage (PPH), bile leak, blood loss, reoperation, readmission, oncologic outcomes (R0-resection, lymph nodes harvested), and operative times. Data were pooled as odds ratio (OR) or mean difference (MD) with a random-effects model. Risk of bias was assessed using the Cochrane Tool and the GRADE approach (Prospero registration ID: CRD42019120363).Three RCTs with a total of 224 patients were included. Meta-analysis showed there were no significant differences regarding 90-day mortality, Clavien-Dindo ≥3 complications, LOS, POPF, DGE, PPH, bile leak, reoperation, readmission, or oncologic outcomes between LPD and OPD. Operative times were significantly longer for LPD {MD [95% confidence interval (CI)] 95.44 minutes (24.06-166.81 minutes)}, whereas blood loss was lower for LPD [MD (CI) -150.99 mL (-168.54 to -133.44 mL)]. Certainty of evidence was moderate to very low.At current level of evidence, LPD shows no advantage over OPD. Limitations include high risk of bias and moderate to very low certainty of evidence. Further studies should focus on patient safety during LPD learning curves and the potential role of robotic surgery.