Procedure-specific Training for Robot-assisted Distal Pancreatectomy.

OBJECTIVE: To train practicing surgeons in robot-assisted distal pancreatectomy (RADP) and assess the impact on 5 domains of healthcare quality. BACKGROUND: RADP may reduce the treatment burden compared with open distal pancreatectomy (ODP), but studies on institutional training and implementation programs are scarce. METHODS: A retrospective, single-center, cohort study evaluating surgical performance during a procedure-specific training program for RADP (January 2006 to September 2017). Baseline and unadjusted outcomes were compared "before training" (ODP only; June 2012). Exclusion criteria were neoadjuvant therapy, vascular- and unrelated organ resection. Run charts evaluated index length of stay (LOS) and 90-day comprehensive complication index. Cumulative sum charts of operating time (OT) assessed institutional learning. Adjusted outcomes after RADP versus ODP were compared using a secondary propensity-score-matched (1:1) analysis to determine clinical efficacy. RESULTS: After screening, 237 patients were included in the before-training (133 ODP) and after-training (24 ODP, 80 RADP) groups. After initiation of training, mean perioperative blood loss decreased (-255 mL, P<0.001), OT increased (+65 min, P < 0.001), and median LOS decreased (-1 day, P < 0.001). All other outcomes remained similar (P>0.05). Over time, there were nonrandom (P < 0.05) downward shifts in LOS, while comprehensive complication index was unaffected. We observed 3 learning curve phases in OT: accumulation (<31 cases), optimization (case 31-65), and a steady-state (>65 cases). Propensity-score-matching confirmed reductions in index and 90-day LOS and blood loss with similar morbidity between RADP and ODP. CONCLUSION: Supervised procedure-specific training enabled successful implementation of RADP by practicing surgeons with immediate improvements in length of stay, without adverse effects on safety.

Technique and outcomes of robot-assisted median arcuate ligament release for celiac artery compression syndrome.

OBJECTIVE: Celiac artery compression by the median arcuate ligament (MAL) is a potential cause of postprandial abdominal pain and weight loss that overlaps with other common syndromes. Robotic technology may alter the current paradigm for surgical intervention. Open MAL release is often performed with concurrent bypass for celiac stenosis due to the morbidity of reintervention, whereas the laparoscopic approach is associated with high rates of conversion to open due to bleeding. We hypothesized that a robot-assisted technique might minimize conversion events to open, decrease perioperative morbidity, and defer consideration of vascular bypass at the initial operative setting. METHODS: We retrospectively analyzed patients treated for MAL syndrome by a multidisciplinary team at a tertiary medical center between September 2012 and December 2013. Diagnosis was based on symptom profile and peak systolic velocity (PSV) >200 cm/s during celiac artery duplex ultrasound imaging. All patients underwent robot-assisted MAL release with simultaneous circumferential neurolysis of the celiac plexus. Postoperative celiac duplex and symptom profiles were reassessed longitudinally to monitor outcomes. RESULTS: Nine patients (67% women) were evaluated for postprandial pain (100%) and weight loss (100%). All patients had celiac stenosis by mesenteric duplex ultrasound imaging (median PSV, 342; range, 238-637 cm/s), and cross-sectional imaging indicated a fishhook deformity in five (56%). Robot-assisted MAL release was completed successfully in all nine patients (100%) using a standardized surgical technique. Estimated blood loss was <50 mL, with a median hospital stay of 2 days (range, 2-3 days). No postoperative complications of grade ≥3, readmissions or reoperations were observed. All patients (100%) improved symptomatically at the 25-week median follow-up. Three patients experienced complete resolution on postoperative celiac duplex ultrasound imaging, and six patients showed an improved but persistent stenosis (PSV >200 cm/s) compared with preoperative velocities (P < .05 by Wilcoxon signed rank). No patients required additional treatment. CONCLUSIONS: Robot-assisted MAL release can be performed safely and effectively with avoidance of conversion events and minimal morbidity. Potential factors contributing to success are patient selection by a multidisciplinary team and replication of the open surgical technique by means of robot-assisted dexterity and visualization. The need for delayed reintervention for persistently symptomatic celiac stenosis is uncertain.

Safety and Efficiency of an Articulating Needle Driver in Advanced Laparoscopic Abdominal Surgery.

Objective: Robotic platforms offer articulating needle drivers but are associated with high costs and lack tactile feedback. The recently developed mechanical Laparoscopic Articulating Needle Driver (LAND) (Flexdex®) for conventional laparoscopy offers enhanced dexterity without these limitations. The goal of this study was to assess safety and efficiency during the implementation of the LAND, and describe its learning curve, in an expert center for laparoscopic surgery. Methods: All LAND-assisted procedures after clinical implementation for a period of 16 months were included into this study. Primary outcome domains were safety (intra- and postoperative complications within 30 days) and efficiency (operative time, suturing, and knot tying time of staple defects of enteroenterostomy). CUmulative SUM (CUSUM) analysis was utilized to describe the enteroenterostomy suturing time learning curve across consecutive cases by plotting the deviation from the series mean. Results: Forty-five procedures (34 Roux-en-Y gastric bypass [of which 7 including diaphragmatic hernia repair], 2 diaphragmatic hernia repair with Nissen fundoplication, and 2 right-sided hemicolectomy) were included into this study. Median (range) operative time and enteroenterostomy suturing time were 68 (46-177) minutes and 161 (112-241) seconds, respectively, comparable with conventional needle driver standards. One procedure was converted to the conventional needle driver due to device malfunctioning and one patients' postoperative course was complicated by a Clavien-Dindo grade 3a complication (intraluminal bleeding requiring gastroscopy). CUSUM chart displays a progression toward the mean from case 22 onward, indicating a limited learning curve. Conclusion: The LAND can be implemented safely and efficiently at a center of excellence for laparoscopic surgery and is associated with a limited learning curve.