Computer vision coaching microsurgical laboratory training: PRIME (Proficiency Index in Microsurgical Education) proof of concept.

Computer vision (CV) feedback could be aimed as a constant tutor to guide ones proficiency during microsurgical practice in controlled environments. Five neurosurgeons with different levels of microsurgical expertise performed simulated vessel dissection and micro-suture in an ex vivo model for posterior computer analysis of recorded videos. A computer program called PRIME (Proficiency Index of Microsurgical Education) used in this research recognized color-labeled surgical instruments, from downloading videos into a platform, with a range of motion greater than 3 mm, for objective evaluation of number of right and left hand movements. A proficiency index of 0 to 1 was pre-established in order to evaluate continuous training improvement. PRIME computer program captured all hand movements executed by participants, except for small tremors or inconsistencies that have a range of motion inferior to 3 mm. Number of left and right hand movements were graphically expressed in order to guide more objective and efficacious training for each trainee, without requiring body sensors and cameras around the operating table. Participants with previous microsurgical experience showed improvement from 0.2 to 0.6 (p < 0.05), while novices had no improvement. Proficiency index set by CV was suggested, in a self-challenge and self-coaching manner. PRIME would offer the capability of constant laboratory microsurgical practice feedback under CV guidance, opening a new window for oriented training without a tutor or specific apparatus regarding all levels of microsurgical proficiency. Prospective, large data study is needed to confirm this hypothesis.

Stroke Microsurgical Thrombectomy Human Placenta Simulator.

BACKGROUND: Stroke microsurgical cerebrovascular thrombectomy reports are limited, although this technique could be used in many centers as a primary treatment or a salvage intervention option. It requires great ability, so our aim is to describe and validate a stroke microsurgical thrombectomy ex vivo simulator with operative nuances analysis. METHODS: Human placenta (HP) models simulated middle cerebral artery vessels with intraluminal thrombus to be microsurgically excised. Six neurosurgeons performed 1-mm and 2-mm longitudinal and transverse arteriotomy in different arteries to remove a 1.5-cm length thrombus. Validation through construct validity compared time to complete the task, complete vessel cleaning, vessel manipulation, vessel stenosis, and leakage in both techniques. RESULTS: All 6 HP models reproduced with fidelity stroke microsurgical thrombectomy, so participants completed 24 sessions, 4 for each neurosurgeon on the same model in different arteries. Construct validity highlighted microsurgical technical difficulties with positive results obtained by parameters variation during performance. Transverse arteriotomy with 1-mm length had best results (P < 0.05) allowing complete thrombus removal, less stenosis, and minor leakage in abbreviated time. CONCLUSIONS: A HP simulator can reproduce with high fidelity all stroke microsurgical thrombectomy part tasks. Transverse 1-mm arteriotomy followed by thrombectomy and 2 simple sutures can fulfill all quality assurance aspects in such intervention accordingly to training model, due to easier vessel opening, complete thrombus removal, no stenosis, and faster microsuture.

Development and validation of the Skills Assessment in Microsurgery for Brain Aneurysms (SAMBA) instrument for predicting proficiency in aneurysm surgery.

OBJECTIVE: Surgical performance evaluation was first described with the OSATS (Objective Structured Assessment of Technical Skills) and modified for aneurysm microsurgery simulation with the OSAACS (Objective Structured Assessment of Aneurysm Clipping Skills). These methods rely on the subjective opinions of evaluators, however, and there is a lack of objective evaluation for proficiency in the microsurgical treatment of brain aneurysms. The authors present a new instrument, the Skill Assessment in Microsurgery for Brain Aneurysms (SAMBA) scale, which can be used similarly in a simulation model and in the treatment of unruptured middle cerebral artery (MCA) aneurysms to predict surgical performance; the authors also report on its validation. METHODS: The SAMBA scale was created by consensus among 5 vascular neurosurgeons from 2 different neurosurgical departments. SAMBA results were analyzed using descriptive statistics, Cronbach's alpha indexes, and multivariate ANOVA analyses (p < 0.05). RESULTS: Expert, intermediate-level, and novice surgeons scored, respectively, an average of 33.9, 27.1, and 16.4 points in the real surgery and 33.3, 27.3, and 19.4 points in the simulation. The SAMBA interrater reliability index was 0.995 for the real surgery and 0.996 for the simulated surgery; the intrarater reliability was 0.983 (Cronbach's alpha). In both the simulation and the real surgery settings, the average scores achieved by members of each group (expert, intermediate level, and novice) were significantly different (p < 0.001). Scores among novice surgeons were more diverse (coefficient of variation = 12.4). CONCLUSIONS: Predictive validation of the placenta brain aneurysm model has been previously reported, but the SAMBA scale adds an objective scoring system to verify microsurgical ability in this complex operation, stratifying proficiency by points. The SAMBA scale can be used as an interface between learning and practicing, as it can be applied in a safe and controlled environment, such as is provided by a placenta model, with similar results obtained in real surgery, predicting real surgical performance.

Learning brain aneurysm microsurgical skills in a human placenta model: predictive validity.

OBJECTIVE Surgery for brain aneurysms is technically demanding. In recent years, the process to learn the technical skills necessary for these challenging procedures has been affected by a decrease in the number of surgical cases available and progressive restrictions on resident training hours. To overcome these limitations, surgical simulators such as cadaver heads and human placenta models have been developed. However, the effectiveness of these models in improving technical skills is unknown. This study assessed concurrent and predictive validity of brain aneurysm surgery simulation in a human placenta model compared with a "live" human brain cadaveric model. METHODS Two human cadaver heads and 30 human placentas were used. Twelve neurosurgeons participated in the concurrent validity part of this study, each operating on 1 human cadaver head aneurysm model and 1 human placenta model. Simulators were evaluated regarding their ability to simulate different surgical steps encountered during real surgery. The time to complete the entire aneurysm task in each simulator was analyzed. The predictive validity component of the study involved 9 neurosurgical residents divided into 3 groups to perform simulation exercises, each lasting 6 weeks. The training for the 3 groups consisted of educational video only (3 residents), human cadaver only (3 residents), and human placenta only (3 residents). All residents had equivalent microsurgical experience with superficial brain tumor surgery. After completing their practice training, residents in each of the 3 simulation groups performed surgery for an unruptured middle cerebral artery (MCA) aneurysm, and their performance was assessed by an experienced vascular neurosurgeon who watched the operative videos. RESULTS All human cadaver heads and human placentas were suitable to simulate brain aneurysm surgery. In the concurrent validity portion of the experiment, the placenta model required a longer time (p < 0.001) than cadavers to complete the task. The placenta model was considered more effective than the cadaver model in simulating sylvian fissure splitting, bipolar coagulation of oozing microvessels, and aneurysm neck and dome dissection. Both models were equally effective in simulating neck aneurysm clipping, while the cadaver model was considered superior for simulation of intraoperative rupture and for reproduction of real anatomy during simulation. In the predictive validity portion of the experiment, residents were evaluated for 4 tasks: sylvian fissure dissection, microvessel bipolar coagulation, aneurysm dissection, and aneurysm clipping. Residents trained in the human placenta simulator consistently had the highest overall performance scores when compared with those who had trained in the cadaver model and those who had simply watched operative videos (p < 0.001). CONCLUSIONS The human placenta biological simulator provides excellent simulation for some critical tasks of aneurysm surgery such as splitting of the sylvian fissure, dissection of the aneurysm neck and dome, and bipolar coagulation of surrounding microvessels. When performing surgery for an unruptured MCA aneurysm, residents who had trained in the human placenta model performed better than residents trained with other simulation scenarios/models. In this age of reduced exposure to aneurysm surgery and restrictions on resident working hours, the placenta model is a valid simulation for microneurosurgery with striking similarities with real surgery.