The development of a deep learning model for automated segmentation of the robotic pancreaticojejunostomy.

BACKGROUND: Minimally invasive surgery provides an unprecedented opportunity to review video for assessing surgical performance. Surgical video analysis is time-consuming and expensive. Deep learning provides an alternative for analysis. Robotic pancreaticoduodenectomy (RPD) is a complex and morbid operation. Surgeon technical performance of pancreaticojejunostomy (PJ) has been associated with postoperative pancreatic fistula. In this work, we aimed to utilize deep learning to automatically segment PJ RPD videos. METHODS: This was a retrospective review of prospectively collected videos from 2011 to 2022 that were in libraries at tertiary referral centers, including 111 PJ videos. Each frame of a robotic PJ video was categorized based on 6 tasks. A 3D convolutional neural network was trained for frame-level visual feature extraction and classification. All the videos were manually annotated for the start and end of each task. RESULTS: Of the 100 videos assessed, 60 videos were used for the training the model, 10 for hyperparameter optimization, and 30 for the testing of performance. All the frames were extracted (6 frames/second) and annotated. The accuracy and mean per-class F1 scores were 88.01% and 85.34% for tasks. CONCLUSION: The deep learning model performed well for automated segmentation of PJ videos. Future work will focus on skills assessment and outcome prediction.

Developing a Robotic Surgery Curriculum: Selection of Virtual Reality Drills for Content Alignment.

INTRODUCTION: Despite the importance of simulation-based training for robotic surgery, there is no consensus about its training curricula. Recently, a virtual reality (VR) platform (SimNow, Intuitive, Inc) was introduced with 33 VR drills but without evidence of their validity. As part of our creating a new robotic VR curriculum, we assessed the drills' validity through content mapping and the alignment between learning goals and drill content. METHODS: Three robotically trained surgeons content-mapped all 33 drills for how well the drills incorporated 15 surgery skills and also rated the drills' difficulty, usefulness, relevance, and uniqueness. Drills were added to the new curriculum based on consensus about ratings and historic learner data. The drills were grouped according to similar skill sets and arranged in order of complexity. RESULTS: The 33 drills were judged to have 12/15 surgery skills as primary goals and 13/15 as secondary goals. Twenty of the 33 drills were selected for inclusion in the new curriculum; these had 11/15 skills as primary goals and 11/15 as secondary goals. However, skills regarding energy sources, atraumatic handling, blunt dissection, fine dissection, and running suturing were poorly represented in the drills. Three previously validated inanimate drills were added to the curriculum to address lacking skill domains. CONCLUSIONS: We identified 20 of the 33 SimNow drills as a foundation for a robotic surgery curriculum based on content-oriented evidence. We added 3 other drills to address identified gaps in drill content.

Defining the Price Tag of Complications Following Pancreatic Surgery: A US National Perspective.

INTRODUCTION: Pancreatic surgery tends to have a high rate of postoperative complications due to its complex nature, significantly increasing hospital costs. Our aim was to describe the true association between complications and hospital costs in a national cohort of US patients. METHODS: The National Inpatient Sample was used to conduct a retrospective analysis of elective pancreatic resections performed between 2004 and 2017, categorizing them based on whether patients experienced major complications (MaC), minor complications (MiC), or no complications (NC). Multivariable quantile regression was used to analyze how costs varied at different percentiles of the cost curve. RESULTS: Of 37,893 patients, 45.3%, 28.6%, and 26.1% experienced NC, MiC, and MaC, respectively. Factors associated with MaC were a Charlson Comorbidity Index of ≥4, prolonged length of stay, proximal pancreatectomy, older age, male sex, and surgery performed at hospitals with a small number of beds or at urban nonteaching hospitals (all P < 0.01). Multivariable quantile regression revealed significant variation in MiC and MaC across the cost curve. At the 50th percentile, MiC increased the cost by $3352 compared to NC while MaC almost doubled the cost of the surgery, increasing it by $20,215 (both P < 0.01). The association between complications and cost was even greater at the 95th percentile, increasing the cost by $10,162 and $108,793 for MiC and MaC, respectively (P < 0.01). CONCLUSIONS: MiC and MaC were significantly associated with increased hospital costs. Furthermore, the relationship between MaC and costs was especially apparent at higher percentiles of the cost curve.

Feasibility, effectiveness and transferability of a novel mastery-based virtual reality robotic training platform for general surgery residents.

BACKGROUND: The annual number of robotic surgical procedures is on the rise. Robotic surgery requires unique skills compared to other surgical approaches. Simulation allows basic robot skill acquisition and enhances patient safety. The purpose of this study was to evaluate the feasibility, effectiveness, and transferability of a mastery-based curriculum using a new virtual reality (VR) robotic simulator for surgery resident training. METHODS: Nineteen PGY2s and 22 PGY4s were enrolled. Residents completed a pretest and posttest consisting of five VR and three previously validated inanimate tasks. Training included practicing 33 VR tasks until a total score ≥ 90% ("mastery") was achieved using automated metrics (time, economy of motion). Inanimate performance was evaluated by two trained, blinded raters using video review metrics (time, errors, and modified OSATS). Outcomes were defined as: curriculum feasibility (completion rate, training time, repetitions), training effectiveness (pre/post training skill improvement), and skill transferability (skill transfer to validated inanimate drills). Wilcoxon signed-rank and Mann-Whitney U tests were used; median (IQR) reported. RESULTS: Thirty-four of 41 residents (83%) achieved mastery on all 33 VR tasks; median training time was 7 h (IQR: 5'26″-8'52″). Pretest vs. post-test performance improved (all p < 0.001) according to all VR and Inanimate metrics for both PGY2 and PGY4 residents. Significant pretest performance differences were observed between PGY2 and PGY4 residents for VR but not inanimate tasks; no PGY2 vs. PGY4 posttest performance differences were observed for both VR and inanimate tasks. CONCLUSION: This mastery-based VR curriculum was associated with a high completion rate and excellent feasibility. Significant performance improvements were noted for both the VR and inanimate tasks, supporting training effectiveness and skill transferability. Additional studies examining validity evidence may help further refine this curriculum.

Safety and Feasibility of Single-Port Robotic-Assisted Nipple-Sparing Mastectomy.

Importance: Robotic-assisted nipple-sparing mastectomies with multiport robots have been described in the US since 2015; however, significant hurdles to multiport robotic surgery exist in breast surgery. Objective: To demonstrate that the single-port da Vinci SP (Intuitive Surgical) robotic system is feasible in patients undergoing robotic nipple-sparing mastectomy (rNSM). Design, Setting, and Participants: An initial case series of 20 patients at a large university hospital underwent bilateral single-port robotic nipple-sparing mastectomies (SPrNSM) with tissue expander reconstruction from February 1, 2020, through January 4, 2023. Participants included women who met surgical criteria for nipple-sparing mastectomies, per standard of care. Intervention: Surgery using a single-port robot and the surgical technique of the authors. Main Outcomes and Measures: Age, indication, body mass index, breast size, operative time, conversion to open surgery, systemic complications, postoperative skin necrosis, and reported skin and nipple areolar complex (NAC) sensation. Results: Twenty women aged 29 to 63 years (median, 40 years) underwent bilateral SPrNSM. Eleven patients completed prophylactic surgery due to a high risk for breast cancer (more than 20% lifetime risk) and 9 patients had breast cancer. Breast size ranged from A through D cup with median B cup and a body mass index range of 19.7 through 27.8 (median 24.4). The total duration of the procedure from incision to skin closure for both sides ranged from 205 minutes to 351 minutes (median, 277). The median robotic time for bilateral SPrNSM was 116 minutes and varied by cup size (A cup, 95 minutes; B cup, 140 minutes; C cup, 118 minutes; D cup, 114 minutes) with no inflection point in learning curve. No cases were converted to open and no immediate complications, such as hematoma, positive margins, or recurrence, were seen. In the first 10 patients prior to routine sensation testing, 20 resected breasts had measurable NAC sensation at a range from 4 to 36 months post-index resection (65%). In the second 10 patients of the cohort, measurable NAC was preserved in 13 of 20 resected breasts 2 weeks following the index operation (65%). Conclusion and Relevance: In this case series, SPrNSM with immediate reconstruction was feasible and performed safely by an experienced breast surgeon with limited previous robotic training. Further studies confirming the preliminary data demonstrating improved NAC and skin sensation following SPrNSM are warranted. Trial Registration: ClinicalTrials.gov Identifier: NCT05245812.

Proficiency Levels and Validity Evidence for Scoring Metrics for a Virtual Reality and Inanimate Robotic Surgery Simulation Curriculum.

OBJECTIVE: Our institution recently implemented a virtual reality (VR) skills curriculum for general surgery residents using the SimNow simulator. Based on a content alignment study, we revised the curriculum to include only 20 of 33 VR tasks and we added 3 previously validated inanimate tasks. The purpose of this study was to establish expert-derived proficiency levels for all tasks and to evaluate the validity of the scoring for the VR tasks. DESIGN: Two expert robotic surgeons performed 5 repetitions of each VR and inanimate task. The trimmed mean (lowest scoring attempt and outliers [>2 standard deviations] were eliminated) was defined as the expert level for each task. For the VR tasks, expert levels were compared to resident performance to evaluate validity. SETTING: This study was conducted at the University of Texas Southwestern Medical Center (Dallas, TX), a tertiary care academic teaching hospital. PARTICIPANTS: Two expert robotic surgeons participated in this study. The data from 42 residents (PGY2-4) who completed the original curriculum was used to represent novice performance. RESULTS: Comparison of expert levels and resident performance was statistically significant for 15 VR tasks (supporting validity) and approached significance (p = 0.06, 0.09) for 2 VR tasks; expert levels were designated as proficiency levels for these 17 tasks. Group comparisons were clearly not significant (p = 0.2-0.8) for 3 VR tasks; 2 of these 3 tasks were retained as introductory exercises (with 3 repetitions required) and 1 was excluded. For the 3 inanimate tasks, expert levels minus 2 standard deviations were designated as proficiency levels. CONCLUSIONS: This analysis generated validity evidence for 15 VR tasks and established expert-derived proficiency levels for 17 VR tasks and 3 inanimate tasks. Our proposed curriculum now consists of 19 VR and 3 inanimate tasks using the selected proficiency levels. We anticipate that this design will maximize curriculum efficiency and effectiveness.

Adapting to a Robotic Era: The Transferability of Open and Laparoscopic Skills to Robotic Surgery.

BACKGROUND: The learning curve of robotic surgical skills is poorly understood. There is a lack of data on the transferability of skills from open and laparoscopic training to robotic surgery. In this retrospective cohort study, we investigated the impact of training acquired during intern year on the development of robotic skills in general surgery residents, prior to formal robotic training. METHODS: Between 2019 and 2021, novice general surgery residents underwent robotic skill assessment using 3 validated inanimate drills before starting intern year. After completing basic open and laparoscopic proficiency-based curricula, they completed an identical robotic skill assessment at the end of intern year. Pre and post intern year robotic performances were recorded and analyzed by 2 blinded graders. Video-based assessment included completion time, errors, and the modified Objective Structured Assessment of Technical Skills (mOSATS) score. RESULTS: The total time needed to complete all 3 robotic drills decreased from a mean of 26 to 17 minutes after intern year (p < 0.001). The number of errors decreased from a mean of 2.16 to 0.56 errors per subject (p < 0.001). The aggregated mOSATS score increased by an average of 41% (p < 0.001), with a greater increase in technical skill domains compared to the knowledge-based domain. The interrater intraclass correlation coefficient was 0.91. CONCLUSIONS: Baseline robotic surgical skills are limited without formal training. Our findings suggest that acquiring basic open and laparoscopic skills, such as knot tying, needle driving, and tissue handling results in improved performance on the robotic platform, prior to formal robotic training. Therefore, requiring trainees to complete fundamental open and laparoscopic training prior to robotic training may be an efficient and effective strategy within a surgical residency curriculum.

Impact of Neoadjuvant Treatment and Minimally Invasive Surgery on Perioperative Outcomes of Pancreatoduodenectomy: an ACS NSQIP Analysis.

BACKGROUND: There is an increasing use of neoadjuvant treatment (NAT) for pancreatic cancer (PC) followed by minimally invasive pancreatoduodenectomy (MIPD). We evaluate the impact of the surgical approach on 30-day outcomes in PC patients who underwent NAT. METHODS: Patients with PC who had NAT followed by MIPD or open pancreatoduodenectomy (OPD) were identified from a pancreatectomy-targeted dataset (2014-2020) of the National Surgical Quality Improvement Program. Comparisons were made between MIPD and OPD within NAT groups. RESULTS: A total of 5588 patients were analyzed. Of those, 4907 underwent OPD and 476 underwent MIPD. In addition, 3559 patients received neoadjuvant chemotherapy alone and 1830 received neoadjuvant chemoradiation. In the chemotherapy-alone group, the MIPD subgroup had lower rates of any complication (38.2% vs. 45.8%, P = 0.005), but there were no differences in mortality (2.1% for MIPD vs 1.9% for OPD, P=0.8) or serious complication (11.8% for MIPD vs 15% for OPD, P=0.1). On multivariable analysis, MIPD was independently predictive of lower rates of any complication (OR: 0.74, 95% CI 0.6-0.93, P = 0.0009), CR-POPF (OR: 0.58, 95% CI 0.35-0.96, P = 0.04), and shorter LOS (estimate: -1.03, 95% CI -1.73 to -0.32, P = 0.004). In the chemoradiation group, patients undergoing MIPD had higher rates of preoperative diabetes (P < 0.05), but there were no significant differences in any outcomes between the two approaches in this group. CONCLUSION: MIPD is safe and feasible after NAT. Patients having neoadjuvant chemotherapy alone followed by MIPD had lower rates of complications, shorter LOS, and fewer CR-POPFs compared to OPD.