Virtual reality in planning complex heart surgery.

Objectives

Intraventricular tunneling is a pivotal step in biventricular or 1.5 ventricle repair of complex congenital heart disease, conventionally reliant on surgeon's intraoperative inspection and judgement. We evaluated novel assistive processes and translational tools for aiding intraventricular tunneling during complex surgical repairs using 3-dimensional virtual reality and 3-dimensional printing technologies.

Methods

Computed tomography data of prospective complex congenital heart disease cases considered for biventricular/1.5 ventricle repair with intraventricular tunneling were converted to "digital twins" and analyzed in 3-dimensional virtual reality using medical modeling software. Digital dissection, annotation, and measurement tools helped identify critical anatomic elements: interventricular septum to great vessel orientation, optimal pathway least likely to obstruct, dynamic changes, conduit requirement, and so forth. Virtual baffle design tool simulated baffles with morphometrics. Novel workflows helped convert virtual baffles to physical templates: 3-dimensionally printed in flexible resin, digitally unfolded and printed in rigid resin, and on plain paper. Templates were compared and used as surgical guides. Adherence to simulation-based plans and surgical baffle placement were assessed.

Results

Twenty cases (mean age 8.61 ± 7.61 years) included double outlet right ventricles (n = 12; 2 with superior-inferior ventricles, 2 with common atrioventricular canal), double outlet both ventricles (1), corrected transposition (5), and D-transposition (2) variants. Situs abnormalities (5/20), dextrocardia (7/20), and additional defects added to the complexity. Operative findings were as predicted, and repair was achieved in 100% (2 ventricle repair in 14 patients and 1.5 ventricle repair in 6 patients). Morphometrics of virtual, 3-dimensional printed, and paper-printed baffles were identical and successfully used as surgical guides; deviation from predicted geometry was less than 10%. All baffles were single patch, placed as per digital plan, with no residuals. Mean cardiopulmonary bypass time was 322 ± 109 minutes, and crossclamp time was 159 ± 54 minutes. Median hospital stay was 10 days (interquartile range, 8-16) with no operative mortality. No reinterventions or residuals were observed at 20 months median follow-up.

Conclusions

Early-stage exploration of translational impact of 3-dimensional virtual reality-based baffle planning is encouraging. Workflow for creating paper-printable baffle templates shows great promise.