Validation of a Chicken Wing Training Model for Endoscopic Microsurgical Dissection

Transcription

1 The Laryngoscope VC 2014 The American Laryngological, Rhinological and Otological Society, Inc. Validation of a Chicken Wing Training Model for Endoscopic Microsurgical Dissection Daniel J. Kaplan, BA; Francisco Vaz-Guimaraes, MD; Juan C. Fernandez-Miranda, MD; Carl H. Snyderman, MD, MBA Objectives/Hypothesis: To determine if training with a chicken wing model improves performance of endoscopic endonasal surgery (EES) with microvascular dissection. Study Design: Randomized experimental study. Methods: A single-blinded randomized clinical trial of trainees with various levels of endoscopic experience was conducted to determine if prior training on a nonhuman model augments endoscopic skill and efficiency in a surrogate model for live surgery. Medical students, residents, and fellows were randomized to two groups: a control group that performed an endoscopic transantral internal maxillary artery dissection on a silicone-injected anatomical specimen, and an interventional group that underwent microvascular dissection training on a chicken wing model prior to performing the anatomic dissection on the cadaver specimen. Time to completion and quality of dissection were measured. Results: A Mann-Whitney test demonstrated a significant improvement in time and quality outcomes respectively across all interventional groups, with the greatest improvements seen in participants with less endoscopic experience: medical students (P 5.032, P 5.008), residents and fellows (P 5.016, P 5.032). Conclusions: Prior training on the chicken wing model improves surgical performance in a surrogate model for live EES. Key Words: Endoscopic endonasal surgery, microvascular dissection, training model. Level of Evidence: 1b. Laryngoscope, 125: , 2015 INTRODUCTION Endoscopic endonasal surgery (EES) utilizes endoscopic visualization through direct ventral corridors to minimize the need to manipulate neural and vascular structures when compared to open skull base approaches. 1,2 The endoscopic endonasal approach has been reported to reduce postoperative morbidity and recovery time, shorten hospitalization, and decrease cost of care. 2 Acquisition of endoscopic dissection techniques is a difficult task, and there is an extended learning curve. 3 7 The endonasal route requires the use of long surgical instruments with which the surgeon must perform microsurgical-like dissection under nonstereoscopic visualization provided by the endoscope. 8 Limitations on From the Department of Neurological Surgery (F.V.-G., J.C.F.-M., C.H.S.) and Department of Otolaryngology (C.H.S.), University of Pittsburgh School of Medicine (D.J.K.), Pittsburgh, Pennsylvania, U.S.A. Editor s Note: This Manuscript was accepted for publication September 24, Presented as a poster at the 24th Annual Meeting of the North American Skull Base Society, San Diego, California, U.S.A., February 14 16, C.H.S. is a consultant for SPIWay LLC. The authors have no other funding, financial relationships, or conflicts of interest to disclose. Send correspondence to Carl H. Snyderman, MD, Department of Otolaryngology, University of Pittsburgh Medical Center, 200 Lothrop Street, EEI Suite 500, Pittsburgh, PA snydermanch@upmc.edu DOI: /lary training hours established by the Accreditation Council for Graduate Medical Education 9 and the need to improve the efficiency of the operating theater could exacerbate the challenge by reducing the surgical experience of residents. For these reasons, a surgical training model for EES is needed. 10,11 The use of surgical simulation allows trainees to develop basic technical skills in a risk-free environment for patients without wasting valuable operating room time. The gold standard is anatomical dissection, but this is an expensive resource with limited availability. 12 Training with virtual reality (VR) simulators has shown great promise, 12 but existing models are expensive or have technical limitations. For example, the endoscopic sinus surgery (ES3) simulator has been shown to be a very effective training tool with particularly strong content, construct, and concurrent validity, 13 but it is no longer in production as it was cost prohibitive for most medical centers. The less expensive VR models have limited options for instrumentation and tend to lack haptic feedback, which significantly impairs their ability to mimic actual surgery. The goal of this study was to demonstrate the efficacy of a previously described ex vivo endoscopic surgical training model using chicken wings. 14 A chicken wing model was selected for its low cost, availability, feasibility, reproducibility, and vascular properties. Efficacy was primarily evaluated in terms of content validity (skill development), but also in terms of construct validity (differentiates 571

2 Fig. 1. Timeline of experiment. between expert and novice) and concurrent validity (outcomes equal performance). MATERIALS AND METHODS Subjects were assigned to four groups: a control and intervention group of medical students, and a control and intervention group of residents and fellows. Intervention groups were trained using the chicken wing model first and then performed the cadaver dissection. Control groups only performed the cadaver dissection (Fig. 1). A cadaver dissection model was used to compare outcome to performance instead of a live surgery due to the lack of endoscopic experience of medical students and junior residents, ethical considerations, and limited clinical opportunities to conduct such a study. Construction of the Surgical Environment Details of the chicken wing dissection have been previously described. 14 A Styrofoam box with a depth of 5 cm and an open top was used. A chicken wing was placed inside of the box in a standardized central position with the vascular pedicle oriented horizontally; right and left wings were randomized. A rubber facial mask was placed over the box; the mask s nostrils were 1.2 cm across and 1.5 cm from top to bottom. The angled distance from the opening of the nostril to the floor of the box was 15 cm. Operating distances and angles were based on a retrospective radiological analysis of fine-cut computed tomography scans of patients and corresponded to the transcribriform, transplanum, transellar, and transclival endoscopic endonasal approaches. A nasal endoscope was placed in the mask s right nostril and held in position by a mechanical scope holder. This setup was done by a member of the research team prior to participant arrival to standardize visualization of the operative field. Participants/Group Sizes Sample sizes were derived by using estimates of means and standard deviations from a preliminary experiment conducted previously at the University of Pittsburgh Medical Center using the chicken wing model. Using these estimates, we determined that a sample size of five in each group (i.e., 10 total medical students and 10 total residents/fellows) would allow us sufficient statistical power to detect a 56% difference in mean times between the control and intervention groups, within each group (medical student and resident/fellow). Sample sizes were limited by the availability of medical students and residents. Ten University of Pittsburgh School of Medicine students were recruited, along with 10 residents and fellows from the 572 University of Pittsburgh Medical Center (UPMC) otolaryngology and neurosurgery departments. Because none of the medical students had any prior surgical experience, they were randomly assigned to the control and the interventional groups (five per group). Among the residents and fellows, four had significant prior experience (defined as performance of at least 100 endoscopic cases or current fellowship training in endoscopic surgical techniques). Randomly, two were assigned to the control and two to the interventional group. Of the six remaining residents, the three residents without prior endoscopic experience were placed in the intervention group, whereas the three residents with limited experience were placed in the control group. By placing the slightly more experienced residents in the control group, we ensured that any superiority in results demonstrated by the interventional group would be due to the intervention, not the prior experience. Training Procedure Chicken wings were randomized to left or right orientation inside the box. All dissections were conducted using a 0 Hopkins rod lens endoscope with high-definition camera monitor and a xenon light source (Karl Storz, Culver City, CA). Appropriate endoscopic microsurgical instrumentation was included (Pittsburgh Black & Gold set; KLS Martin, Jacksonville, FL). Video and still images of the training sessions were recorded using an AIDA video recording system (Karl Storz). All procedures were performed using a simulated two-surgeon technique with two hands free for dissection 15 with the endoscope held by a mechanical scope holder at the 12 o clock position in the simulated patient s right nostril, the surgical instruments at the 6 o clock position in the right and left nostrils, and the chicken wing in the transplanum position at the bottom of the box using the reference marks described previously. 14 Medical students completed six training sessions over the course of 2 weeks, with no more than one session allowed per day. Due to time constraints, residents and fellows were asked to complete five training sessions in 2.5 weeks, with at most one per day. Prior to the start of training, subjects were shown a training video demonstrating the chicken wing dissection performed by a skilled surgeon, and read a standardized script of the video procedure prior to their first session. During the training, subjects were asked to perform five representative tasks. 14 The goal of task 1 was opening the skin from joint to joint. Task 2 involved connective tissue dissection with separation of the two main groups of muscles and exposure of the main perineurovascular sheath. Task 3 was defined as opening the perivascular sheath and exposing the dorsal wall of the vessel along its entire length. In task 4, the trainee was directed to completely dissect the nerve from the artery without injuring either structure (Fig. 2). Task 5 consisted of clipping the vessel with a pair of aneurysm clips and a clip applier. Time to complete each individual task and the total time for completion of the five tasks combined (from skin to neurovascular dissection) was measured. Cadaver Dissection Approval for this study was obtained from the University of Pittsburgh s institutional review board (PRO ) and the Committee for Oversight of Research and Clinical Training Involving Decedents (CORID No. 419). All study participants performed the cadaver dissection following the completion of the chicken wing training session by the intervention groups. Medical students and residents/fellows in the training group

3 Fig. 2. Chicken wing dissections step. Reprinted with permission from Jusue-Torres et al. 14 Fig. 4. Resident and fellow dissection times. performed the cadaver dissection within 1 and 2 weeks, respectively, of completion of training. Anatomical specimens (Research for Life, Chandler, AZ) were fixed in formalin. Subjects were asked to perform a transnasal dissection of the internal maxillary artery (IMA) within the pterygopalatine space due to its clinical relevance and moderate complexity. Laterality was randomly assigned to each participant by a computer program. Specimens were prepared by the research team prior to participant dissection. This included a medial maxillectomy and exposure of the sphenopalatine foramen to provide a standardized surgical field for dissection. A posterior septectomy provided binarial access. Subjects were shown a training video made by a skilled surgeon, and read a standardized script immediately prior to their dissection. The scope was placed in the ipsilateral nostril of the target IMA and positioned with a mechanical endoscope holder. The participants worked through both nostrils, using a simulated two-surgeon, two-handed technique. Participants were given four tasks to complete. Task 1 consisted of removing the bone of the posterior wall using a Kerrison rongeur while maintaining the periosteal membrane. Task 2 consisted of incising the periosteal membrane. Task 3 was dissection of three branches of the IMA. Task 4 challenged participants to clip each vessel with one aneurysm clip using an applier. Time to completion of the entire procedure was measured using video-recorded timestamps. These discs were then blindly graded by an expert surgeon (F.V.-G.), using the Modified Global Rating Scale of Operative Performance (MGRS) tool. 16 This is a 35-point scale with seven categories (respect for tissue, time and motion, instrument handling, flow of operation, use of assistants or endoscope, overall performance, and quality of final product), each worth a maximum of five points. Statistical Analysis All quantitative variables were expressed in average and standard deviation and analyzed with the Mann-Whitney U test for nonparametric data. Descriptive statistics were also gathered. RESULTS Chicken Wing Training Medical students. The results of the training sessions are depicted in Figure 3. The average total time to completion of all five tasks improved from minutes for the first trial to minutes for trial 6 (paired t test, P ). Maximum improvement was noted by trial 3. Residents/fellows. The results of this cohort s training sessions are depicted in Figure 4. The average total time to completion of all five tasks improved from minutes for the first trial to minutes for trial 6 (paired t test, P ). Improvement in dissection times was consistent across trials, and dissection times for residents/fellows were significantly better than the medical student group at all trials (P in the final trial). Left and right orientation data were not recorded for the chicken wing phase. Maximum improvement again was noted by trial 3. Fig. 3. Medical student dissection times. Cadaver Dissection Medical students. Medical student cadaver dissection performance is reported in Table I. Differences in average time between the experimental (chicken wingtrained) and control groups were statistically significant according to a Mann-Whitney U test (P 5.032) (Fig. 5). 573

4 TABLE I. Medical Student Cadaver Dissection Performance. Intervention Time, min MGRS Control Time, min MGRS 8-R L L R L L R L R R L 5 left; MGRS 5 Modified Global Rating Scale of Operative Performance; R 5 right. Fig. 5. Cadaver dissection times by treatment and experience level. Fig. 6. Cadaver dissection quality scores by treatment and experience level. MGRS 5 Modified Global Rating Scale of Operative Performance. The untrained group (control) had a mean time of minutes (95% confidence interval: minutes). The trained group had a mean time of minutes (95% confidence interval: minutes). Additionally, the Spearman rho test showed correlation between participants final chicken wing dissection time and cadaver dissection time (P 5.037). There was no statistical difference between right and left sides (P ). Differences in quality between the two groups were also statistically significant according to a Mann- Whitney U test (P 5.008) (Fig. 6). Using the MGRS tool, the control group had a mean of (95% confidence interval: ). The trained group had a mean score of (95% confidence interval: ). Residents and fellows. Resident and fellow cadaver dissection performance is reported in Table II. Differences in average time between the intervention (chicken wing-trained) and the control groups were statistically significant according to a Mann-Whitney U test (P 5.016) (Fig. 5). The control group had a mean time of minutes (95% confidence interval: ), whereas the trained group had a mean time of minutes (95% confidence interval: ). In addition, dissection times for trained residents/fellows were much less than trained medical student times. There was no statistical difference between right and left sides (P ). Differences in quality were also statistically significant according to a Mann-Whitney U test (P 5.032) (Fig. 6). The untrained group had a mean score of (95% confidence interval: ). The trained group had mean score of (95% confidence interval: ). Quality scores of trained residents and fellows were superior to trained medical students. DISCUSSION EES offers tremendous benefit to patients, but has an exceptionally long learning curve. Currently, live surgery and cadaveric dissections are the only accepted paradigms of training. Live surgery exposes patients to unnecessary risk and is an inefficient use of a scarce resource (operating room). The result is that most residents are not being exposed to advanced endoscopic endonasal approach techniques until further into their training. Thus, there is a need for a safe environment for trainees to learn at earlier stages of their career. Vivisection of animals is costly, unappealing, and strictly regulated. Anatomical specimens are expensive, have limited availability (donation, cost, regulation, and societal attitudes), and pose a biological risk. Virtual reality simulators lack the operative experience of haptic feedback with real instruments, and are prohibitively expensive. The clinically validated ES3 model is limited to a few large academic centers. New models like the German Virtual Reality-Functional Endoscopic Sinus Surgery model are promising, but will also be outside the price range of the majority of centers, and has not been validated. Here, we introduce an ex vivo surgical training model for EES of the skull base. After exposure to the model, we hypothesized that all groups should improve to some degree, with the most skillful individuals improving the least. 17 We believe that the chicken wing model addresses the four traditional domains of validity (face, content, 574

5 TABLE II. Resident/Fellow Cadaver Dissection Performance. Intervention Time, min MGRS Control Time, min MGRS Sig. Exp. 17-L Sig. Exp 20-R Sig. Exp. 14-L Sig. Exp 16-R None 12-R Limited 18-R None 13-R Limited 19-L None 15-L Limited 11-L L 5 left; MGRS 5 Modified Global Rating Scale of Operative Performance; R 5 right; Sig. Exp. 5 significant experience. construct, concurrent). In particular, we have demonstrated a significant increase in surgical skill and operative performance using cadaver dissection (typical operative procedure) as a surrogate for live surgery (content validity). Statistically significant improvements in time and quality were observed for both the medical student and resident/fellow groups. On average, trained medical students performed 45% faster than untrained medical students, and operated at 125% higher quality. Trained residents and fellows performed 42% faster than their untrained cohort, and performed at 44% higher quality. Despite starting with various levels of skills and experience, subjects who lacked previous significant endoscopic experience and received training intervention (medical students, inexperienced residents) performed dramatically better in both time and quality on their cadaver dissection than their untrained counterparts. Residents without endoscopic experience benefited the most from the training. Those with the least experience (medical students) also showed dramatic improvements. As expected, study participants with significant prior endoscopic experience (fellows) benefited less from the training. A valid model would be expected to discriminate between medical students and residents/fellows based on their varying levels of experience (construct validity). In our study, participants with more experience consistently performed better on the chicken wing model during their first attempt. Importantly, all participants were able to reach a somewhat similar level of proficiency regardless of starting skill level in a relatively short time. Most improvement occurred during the first three sessions. The optimal length of training and the ideal frequency of repetitive sessions is unknown. Based on the correlation between time required for trial 6 of the CW dissection and time on the cadaveric dissection, it is very possible that performance on the model can directly reflect ability to perform in surgery (concurrent validity). The results of this study suggest that the chicken wing trainer is an effective model for teaching EES skills, and these skills may be translatable to similar live surgery. The benefits of such simulation are most pronounced for those with no significant prior endoscopic surgical exposure. In a training program, demonstrated proficiency on the chicken wing model or in a cadaveric dissection should be demonstrated before performing a live surgery. The normative data provided here are useful in determining when a trainee has achieved proficiency (three to five trials over 1 week). Ideally, this would then be confirmed with a closer approximation for live surgery using cadaveric dissection. Cadaveric dissection has the added benefit of reinforcing anatomical relationships. Learning retention is most effective when there is spacing of learning episodes; the optimal spacing and frequency is unknown. Long gaps between EES during residency should be preceded by a repeat assessment. The chicken wing model also lends itself to self-paced learning with minimal involvement of other personnel. This model is a very low-cost alternative to the more traditional cadaveric or live specimen dissections used to train residents as well as the newer option of virtual simulators. Additional economic benefits accrue from increased operative efficiency for the hospital with greater throughput of cases, better use of human resources, and improved patient and employee satisfaction. A comparison of pituitary surgeries at our institution demonstrated an approximately 50% increase in operative duration with the involvement of trainees. 18 A modest increase in operative efficiency due to better training outside of the operating room could have a large economic impact. A potential limitation of this study is the absence of objective data regarding the level of endoscopic skill at entry into the study. This is unlikely to be significant due to the marked differences in nascent surgical skills between medical students and residents/fellows. Anatomical variations of the specimens may have introduced time delays, but these would not have been systematic. Completion of the surgical tasks was judged by an observer in an unblinded fashion. A strong positive correlation of time and quality scores suggests that this did not significantly alter the results (negative correlation). Last, videos were judged by only one expert reviewer, and internal consistency of the reviewer was not assessed. However, to minimize any temporal bias, the reviewer was given all dissection videos at the same time and reviewed them in a randomly generated order. Future studies will include validation of the chicken wing model in the operating room with live surgeries and the retention of acquired skills in the absence of regular endoscopic experience. The model could also be used as an evaluation tool for residents prior to moving to the next training level or for the assessment of surgical skills of established surgeons (credentialing and ongoing professional practice evaluation). It is likely that multiple simulation models similar to the chicken wing model will be necessary to cover the variety of surgical skills that are acquired during a surgical residency. CONCLUSION In this single-center, randomized, single-blind study, a chicken wing model for EES training was validated. Although the training proved helpful for individuals of all skill levels, it was particularly useful for individuals without significant prior endoscopic training (medical students and junior residents). Compared to alternative models, the chicken wing is effective and inexpensive. It has wide applicability to the training and evaluation of surgical residents and practicing surgeons. 575

6 In a value-based system of healthcare, training models such as this have the potential to improve quality of care while reducing costs. BIBLIOGRAPHY 1. Maroon JC. Skull base surgery: past, present, and future trends. Neurosurg Focus 2005;19: Prevedello D, Kassam A, Gardner P, Carrau R, Snyderman C. Expanded endoscopic endonasal approaches to the skull base. In: Cappabianca P, Califano L, Iaconetta G, eds. Cranial, Craniofacial and Skull Base Surgery. Milan, Italy: Springer; 2008: Sonnenburg RE, White D, Ewend MG, Senior B. The learning curve in minimally invasive pituitary surgery. Am J Rhinol 2004;18: Koc K, Anik I, Ozdamar D, Cabuk B, Keskin G, Ceylan S. The learning curve in endoscopic pituitary surgery and our experience. Neurosurg Rev 2006;29: O Malley BW Jr, Grady MS, Gabel BC, et al. Comparison of endoscopic and microscopic removal of pituitary adenomas: single-surgeon experience and the learning curve. Neurosurg Focus 2008;25:E Leach P, Abou-Zeid AH, Kearney T, Davis J, Trainer PJ, Gnanalignham KK. Endoscopic transsphenoidal pituitary surgery: evidence of an operative learning curve. Neurosurgery 2010;67: Smith SJ, Eralil G, Woon K, Sama A, Dow G, Robertson I. Light at the end of the tunnel: the learning curve associated with endoscopic transsphenoidal skull base surgery. Skull Base 2010;20: Snyderman CH, Pant H, Carrau RL, Prevedello D, Gardner P, Kassam AB. What are the limits of endoscopic sinus surgery?: the expanded endonasal approach to the skull base. Keio J Med 2009;58: Pepper DJ, Schweinfurth M, Herrin VE. The effect of new duty hours on resident academic performance and adult resuscitation outcomes. Am J Med 2014;127: Snyderman C, Kassam A, Carrau R, Mintz A, Gardner P, Prevedello DM. Acquisition of surgical skills for endonasal skull base surgery: a training program. Laryngoscope 2007;117: Prevedello DM, Kassam AB, Snyderman C, et al. Endoscopic cranial base surgery: ready for prime time? Clin Neurosurg 2007;54: Wiet G, Stredney D, Wan I. Training and simulation in otolaryngology. Otolaryngol Clin North Am 2011;44: Fried MP, Wiet GJ, Sadoughi B. Simulation and haptics in otolaryngology training. In Flint PW, Haughey BH, Lund VJ, et al., eds. Cummings Otolaryngology-Head & Neck Surgery. 5th ed. Philadelphia, PA: Mosby Elsevier; 2010: Jusue-Torres I, Sivakanthan S, Fernandez-Miranda JC, Pinheiro-Neto CD, Gardner PA, Snyderman CH. Chicken wing training model for endoscopic microsurgery. J Neurol Surg B 2013;74: Mamelak AN, Carmichael J, Bonert VH, Cooper O, Melmed S. Single-surgeon fully endoscopic endonasal transsphenoidal surgery: outcomes in three-hundred consecutive cases. Pituitary 2013;16: Pandey VA, Wolfe JHN, Black SA, Cairols M, Liapis CD, Berggvist D. Self-assessment of technical skill in surgery: the need for expert feedback. Ann R Coll Surg Engl 2008;90: Downing SM. Validity: on meaningful interpretation of assessment data. Med Educ 2003;37: Dedhia RC, Lord CA, Pinheiro-Neto CD, et al. Endoscopic endonasal pituitary surgery: impact of surgical education on operation length and patient morbidity. J Neurol Surg B 2012;73: