3D-2D Distance Maps Conversion Enhances Classification of Craniosynostosis

• Published in: IEEE transactions on biomedical engineering Vol. PP; no. 11; pp. 1 - 10
• Main Authors: Schaufelberger, Matthias, Kaiser, Christian, Kuhle, Reinald, Wachter, Andreas, Weichel, Frederic, Hagen, Niclas, Ringwald, Friedemann, Eisenmann, Urs, Hoffmann, Jorgen, Engel, Michael, Freudlsperger, Christian, Nahm, Werner
• Format: Journal Article
• Language: English
• Published: United States IEEE 01.11.2023; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: 2D conversion; Artificial neural networks; Birth defects; Classification; Classifiers; CNN; Computed tomography; Conversion; convolutional neural network; Coordinate transformations; Cranial sutures; Craniosynostosis; Data augmentation; Datasets; Diagnosis; distance map; Head; Image resolution; Ionizing radiation; Magnetic heads; Mapping; Medical imaging; Neural networks; Pediatrics; photogrammetric surface scans; Photogrammetry; Radiation; Radiation effects; Ray tracing; resolution; Sampling; Shape; Surgery; Three-dimensional displays; Tomography; Training
• ISSN: 0018-9294; 1558-2531
• DOI: 10.1109/TBME.2023.3278030

Objective: Diagnosis of craniosynostosis using photogrammetric 3D surface scans is a promising radiation-free alternative to traditional computed tomography. We propose a 3D surface scan to 2D distance map conversion enabling the usage of the first convolutional neural networks (CNNs)-based classification of craniosynostosis. Benefits of using 2D images include preserving patient anonymity, enabling data augmentation during training, and a strong under-sampling of the 3D surface with good classification performance. Methods: The proposed distance maps sample 2D images from 3D surface scans using a coordinate transformation, ray casting, and distance extraction. We introduce a CNNbased classification pipeline and compare our classifier to alternative approaches on a dataset of 496 patients. We investigate into low-resolution sampling, data augmentation, and attribution mapping. Results: Resnet18 outperformed alternative classifiers on our dataset with an F1-score of 0.964 and an accuracy of 98.4  %. Data augmentation on 2D distance maps increased performance for all classifiers. Under-sampling allowed 256-fold computation reduction during ray casting while retaining an F1-score of 0.92. Attribution maps showed high amplitudes on the frontal head. Conclusion: We demonstrated a versatile mapping approach to extract a 2D distance map from the 3D head geometry increasing classification performance, enabling data augmentation during training on 2D distance maps, and the usage of CNNs. We found that low-resolution images were sufficient for a good classification performance. Significance: Photogrammetric surface scans are a suitable craniosynostosis diagnosis tool for clinical practice. Domain transfer to computed tomography seems likely and can further contribute to reducing ionizing radiation exposure for infants.