The introduction of endovascular techniques has progressively reduced the number of open surgeries for cerebrovascular diseases. Consequently, neurosurgical residents and young trainees have fewer opportunities to see, learn and practice cerebrovascular procedures. Nonetheless, bypass surgery still plays a crucial role in treating complex aneurysms, moyamoya disease, and in restoring adequate cerebral circulation when dealing with difficult intracranial tumors or other occlusive vascular lesions (1–8).

Microvascular anastomosis is one of the most challenging neurosurgical technique because of the small diameter of the vessels (less than 2 mm in most of the cases) as well as the depth and the narrowness of the surgical field. Hence, microvascular anastomosis techniques require dedicated and effective training to be mastered. Several training models, such as chicken wings, rats, placental vessels, human cadavers, and plastic tubing have been suggested for improving and refining microsurgical skills (2, 9–15). In the last years, virtual reality models and web-based simulators have also proved to be very powerful tools in enhancing trainees' learning experience at various levels (1, 16–22).

However, the proposed training models do not often adequately reflect the realism of an actual cerebrovascular bypass surgery and/or are not easily accessible. Here, we describe a simplified, ready-to-use, reusable, ergonomic simulator for training microvascular anastomosis techniques (1, 16–22).

A total of 336 bypasses were performed (24 for each trainee, 8EE, 8ES, 8SS). Detailed results are displayed in Tables 1–3.

Novice group average TPB (minutes) at the first vs. at the eighth attempt was 37.25 vs. 21.25 min for EE (p < 0.0005), 41.58 vs. 28.83 min for ES (p < 0.0005), 40.67 vs. 28.83 min for SS (p < 0.0005).

The mean NOMAT score for the expert group at the first vs. at the eighth attempt was 44 vs. 45.5 for EE (p = 0.746), 44.25 vs. 48 for ES (p = 0.095), 39.25 vs. 43 for SS (p = 0.095). The mean NOMAT score for the novice group at the first vs. at the eighth attempt was 20.5 vs. 33.75 for EE (p < 0.0005), 20.50 vs. 20.58 for ES (p = 0.959), 33.75 vs. 33.75 for SS (p = 1.00).

The mean Likert score after the last attempt of microanastomosis was 25 for the expert group and 24.58 for the novice group out of a maximum score of 30 (Figure 3).

Residents and young surgeons require a continuous training process in order to develop eye-hand coordination and dexterity in performing microvascular anastomosis techniques. However, the proposed training models do not often adequately reflect the realism of an actual cerebrovascular bypass surgery and/or are not easily accessible (Table 4).

In the past years live animals (in particular rats) were the most adopted training model. The main advantages of this model are obviously represented by the pulsatile blood flow and natural viscosity. Furthermore, the size and the feel of animal arteries resemble those of human vessels (1, 25). However, nowadays, increasingly strict regulations concerning animal welfare, appropriate shelter, economic and ethical issues, and surgical management have progressively imposed significant but necessary limitations of the use of live animal models. In addition, animals models have to be adequately prepared before starting a microvascular anastomosis training: a careful dissection of the surrounding fat and connective tissue is required to expose the vessels, and sometimes the vessels may not have the desirable size and/or length (1, 25).

For these reasons, other bypass training models have been evaluated in the last decades. Fresh human and bovine placentas proved to be excellent tissue models for bypass training, because their vessels are greatly similar to the cerebral ones, and a placenta may provide material for dozens of microsurgical anastomoses in a single session. Several placenta training models have been proposed and described in the medical literature (26–28). Oliveira et al. proposed a human placenta simulator, a high-fidelity and easily available simulator for training neurosurgeons in vascular microsurgery (17, 29). Belykh et al. demonstrated also that human and bovine placental vessels are convenient, anatomically relevant, and beneficial models for microneurosurgical training. Hence, the microanastomosis simulation using these models has high face, content, and construct validities (1). In 2020, Ferrarez et al. described a superficial temporal artery-middle cerebral artery bypass training simulator placing a human placenta in an artificial skull in which a fronto-temporal approach had been previously performed (18).

On the other hand, the use of placenta vessels has also several disadvantages: (1) scarce opportunities to work on these models on an individual basis (educational programs are often required); (2) a specific distribution route has to be arranged with the Obstetrics and Gynaecology unit; (3) storage issues; (4) placenta vessels, like animals' models, have to be adequately prepared before starting a microvascular anastomosis training.

Beyond live animals and placentas, several artificial models have been proposed and described in the medical literature. In 2011, Mori et al. proposed a simulation model for training posterior circulation revascularization: a three-dimensional skull model with artificial brain, where superior cerebellar and posterior inferior cerebellar arteries were made from artificial blood vessels and glued on the brain (30). In 2020, Cikla et al. described the “grapefruit training model” which may provide a realistic simulation of side-to-side distal anterior cerebral artery bypass procedure using a “dissected grapefruit” to simulate the interhemispheric fissure, chicken wing vessels or synthetic tubing for pericallosal arteries, and an aquarium pump to mimic circulation (31). The main disadvantage of this model is that its assembly is not so immediate (in particular if the vessels are harvested from chicken wings). In addition, the model is specifically designed for training only the distal anterior cerebral artery side-to-side bypass (31).

In consideration of the above, even if it is evident that several and different bypass training models have been proposed so far, few reflect the realism of an actual cerebrovascular bypass surgery, even fewer are reusable, most are not easily accessible, and the setting procedure before starting the bypass training is not so fast in almost all the models.

The setting procedure of the Iatrotek® bypass simulator is really quick, usually less than 5 min before the training could start. Our training model is reusable, and may be easily transported and used in any available place.

Obviously, the Kezlex PVA hydrogel synthetic vessels do not allow to train dissection skills, but they are more readily available, and easier to handle and store than live animals, placentas or artificial models. Then, the PVA hydrogel vessels are produced using three-dimensional computed tomography/magnetic resonance imaging scanning data, and their mechanical properties (e.g., surface friction and elasticity) and transparency are very similar to real arteries. In addition, the patency and the tightness of the bypass may be tested irrigating the vessels or using the indocyanine green.

Finally, when using the Iatrotek® bypass simulator, the degree of difficulty of the training may be easily and quickly adjusted modifying the depth and the width of the working field, as well as the position of the interchangeable approximators.

Our experience showed that the training with the Iatrotek® bypass simulator was able to determine a performance improvement in both the novice and expert groups. When comparing the results of the first and the last attempts, we registered an improvement in the mean TPB in both groups for all the three types of microanastomosis. The improvement was obviously more pronounced in the novice group, always reaching a statistical significance, while in the expert group the improvement was statistically significant only for the ES bypass. The NOMAT score also tended to improve in both groups, showing a statistically significant progress in the novice group for the EE bypass. The mean number of leakages after bypass execution, as well as the relative time for their resolution, also tended to slightly and progressively reduce in both groups by increasing the number of attempts, thus again underlining the potential effectiveness of the training model in refining microsurgical skills. Experienced trainees expressed a slightly better subjective evaluation of the training model than novices (Likert score of 25 vs. 24.58 out of 30, respectively).

Iatrotek® didactic platform and PVA hydrogel vessels, despite the above described advantages, are more expensive than other models presented in our discussion. On the other hand this bypass simulator could be used not only by neurosurgeons but also by other physicians (e.g., vascular, plastic surgeons, etc.). In this way, by purchasing only one simulator, surgeons from different specialties could improve their microvascular skills. As a consequence, PVA hydrogel vessels could be purchased in large amounts and at a more convenient price from the supplier. However, as above mentioned, PVA hydrogel vessels are more expensive than placental ones. On the other hand, PVA hydrogel vessels have no storage issues and no costs of conservation. In addition, they are ready to use and realistic because of their technique of production and mechanical properties. Hence, according to our analysis, this bypass simulator could be practical and cost-effective when compared with the alternative solutions on the market.

In conclusion, our proposed bypass training model may represent a simplified, ready-to-use, reusable, ergonomic and efficient system to improve eye-hand coordination and dexterity in performing microanastomosis.