Based on 3D images of Fragile X Syndrome in a retrospective cohort of young Chinese males

Click here to read the complete paper by Jieyi Chen, Siyuan Du, Yiting Zhu, Dongyun Li, Chunchun Hu, Lianni Mei, Yunqian Zhu, Huihui Chen, Sijia Wang, Xiu Xu, Xinran Dong, Wenhao Zhou, and Qiong Xu

A long and narrow face. A broad forehead. Mandibular Prognathism or “underbite”. Protruding ears. These are some of the most common results for searches involving common facial characteristics in individuals living with Fragile X Syndrome (FXS). What is lesser known, however, is that these results are based on studies focusing mostly on post-pubertal or adult individuals living with FXS, without digital images. In short, there is a dearth of literature studying the facial features of the younger population of those living with FXS, thereby limiting the ability to identify early FX facial features. 

3D images as opposed to 2D images allow stereoscopic and quantitative facial phenotypes to be extracted. This allows for the exploration of the link between subtle facial phenotypes and genotypes, including genomic mutations and methylation. The authors of this study were able to depict more subtle facial features of patients through the comparison and quantitative analysis of 3D images from FXS patients and controls at early childhood by using this technology. They also studied the possibility of better facilitating screening by machine learning using 3D images. Additionally, we explored how to better assist screening by machine learning utilizing 3D facial images. They also investigated whether different genetic genotypes and methylation subtypes in patients affect facial morphology.

The authors were able to qualitatively visualise the typical and subtle variations between 3D facial images. They found that the projection of patients and controls in Fragile X-linked vectors were significantly different. Overall, the study suggests that  3D facial images could assist to distinguish male patients living with FXS by machine learning, in which the selected regional features performed better than the global features and sparse landmarks. It also found that the genetic and methylation status may affect regional facial features differently.