RapidNAM: Algorithm for the semi-automated generation of nasoalveolar molding device designs for the presurgical treatment of bilateral cleft lip and palate.

Objective: Nasoalveolar molding (NAM) is an accepted presurgical treatment modality for newborns with cleft lip and palate (CLP). However, the therapy is time-consuming and requires high expertise. To facilitate the treatment, we reveal an algorithm for the automated generation of patient individual NAM devices for neonates with bilateral cleft lip and palate (BCLP) and present results of software validation. Methods: The algorithm was implemented utilizing Python 2.7 and Blender 2.78a based on 17 digitized (3D-scanning) impressions of maxillae with BCLP. The algorithm segments alveolar structures, bridges clefts, and generates a series of NAM device designs, destined for 3D-printing for subsequent treatment. The datasets were used for first software tests. For validation, a follow-up study was carried out using six new, independent maxilla models. The generated NAM plate designs were examined regarding their potential clinical usability. Furthermore, a deviation analysis was carried out, which measured the plate models’ and upper jaw models’ surface deviations. Results: Series of NAM devices were generated automatically in 21 out of 23 cases. We calculated an average surface deviation of 0.140 mm (SD: 0.016 mm). Four out of six plate series (follow-up trials) were assessed as probably usable with minor adjustments. Conclusion: The algorithm generates 3D-printable series of NAM device designs reliably. We expect most of the series to be clinically usable and that the first plates of each series will fit the patients’ maxillae. Significance: The proposed algorithm has the potential to reduce the therapist's manual work and therefore time effort/costs related to NAM.