Objectives: Congestive heart failure is showing an increasing incidence in modern society with progressively higher life expectancy. More than 64 million people worldwide and 47,2% of those over the age of 95 years are affected. 

Common causes for heart failure include valvular heart diseases such as mitral valve insufficiency. Mitral valve insufficiency is the second most common heart disease to be treated worldwide. The surgical reconstruction is often the best treatment option, at which the minimally invasive procedure has shown far better outcomes than the open-heart procedure.

Minimally invasive mitral valve repair (MVR) is a highly complex procedure, which requires plenty of practice. Previous studies showed that surgeons must perform at least 75-125 operations to overcome the learning curve. Benefits of training in this area have been poorly investigated. We conducted a prospective study to assess the training effect of different tasks in MVR on a patient-specific simulator.

Methods: 25 medical students, in their first to sixth year, without previous experience in MVR were recruited for this study and divided into two groups. Group A (Gr. A) attended three training sessions while Group B (Gr. B) only two. By each session, participants were asked to entirely perform two neo-chordae implantation and ring-annuloplasty on silicone patient-specific valve replicas following precise instruction. Three different models with different pathologies were used (posterior, anterior and bi-leaflet prolapse). During the session the anatomical accuracy (papillary muscle and leaflet segment identification for neo-chordae implantation, annulus detection for annuloplasty), functional accuracy (length of neo-chordae, quality of knot-tying) and time of execution for each step were evaluated to quantify the learning progress. After every session participants received feedback from an expert to enhance their improvement. 

Results: The following results refer to a comparison between the first and last training session. Regarding anatomical and functional accuracy, irregular stitches by annuloplasty decreased from 52% to 20% generally. For the implantation of two neo-chordae participants saved an average time of 24.64% (354 s to 364 s, p<0.049). Gr. A improved by 30.02% (344 s to 240 s, p<0.024) and Gr. B by 19.71% (363s to 286s, p<0.151). Annuloplasty-stich time decreased by 15.31% (53 s to 43 s, p<0.089) for Gr. A while Gr. B showed no significant changes (both 56 s, p<0.976). For knot-tying the average improvement was 8.72% (79 s to 72 s, p<0.007). Gr. A enhanced their performance by 12.97% (75 s to 65 s, p<0.013) whereas Gr. B by 4.49% (82 s to 78 s, p<0.179). Differences between both groups were most significant for annuloplasty-stitching (Gr. A 43 s, Gr. B 56 s, p<0.006) and knot-tying (Gr. A 65 s, Gr. B 78 s, p<0.001). 

Conclusion: All participants overall became faster and more precise after completing all training sessions, showing that surgical training in MVR can flatten the steep learning curve. More training is beneficial for achieving better results in performing neo-chordae implantation and ring-annuloplasty. Training on patient-specific simulators proves to be an effective method for learning complex procedures of MVR and should be implemented into clinical routines of both cardiac surgery as well as cardiology.