Comparison between Saliva and Nasopharyngeal Swab Specimens for Detection of Respiratory Viruses by Multiplex Reverse Transcription-PCR.

Nasopharyngeal swabs (NPSs) are being widely used as specimens for multiplex real-time reverse transcription (RT)-PCR for respiratory virus detection. However, it remains unclear whether NPS specimens are optimal for all viruses targeted by multiplex RT-PCR. In addition, the procedure to obtain NPS specimens causes coughing in most patients, which possibly increases the risk of nosocomial spread of viruses. In this study, paired NPS and saliva specimens were collected from 236 adult male patients with suspected acute respiratory illnesses. Specimens were tested for 16 respiratory viruses by multiplex real-time RT-PCR. Among the specimens collected from the 236 patients, at least 1 respiratory virus was detected in 183 NPS specimens (77.5%) and 180 saliva specimens (76.3%). The rates of detection of respiratory viruses were comparable for NPS and saliva specimens (P = 0.766). Nine virus species and 349 viruses were isolated, 256 from NPS specimens and 273 from saliva specimens (P = 0.1574). Adenovirus was detected more frequently in saliva samples (P < 0.0001), whereas influenza virus type A and human rhinovirus were detected more frequently in NPS specimens (P = 0.0001 and P = 0.0289, respectively). The possibility of false-positive adenovirus detection from saliva samples was excluded by direct sequencing. In conclusion, neither of the sampling methods was consistently more sensitive than the other. We suggest that these cost-effective methods for detecting respiratory viruses in mixed NPS-saliva specimens might be valuable for future studies.

Whole-genome sequencing in clinically diagnosed Charcot-Marie-Tooth disease undiagnosed by whole-exome sequencing.

Whole-genome sequencing is the most comprehensive form of next-generation sequencing method. We aimed to assess the additional diagnostic yield of whole-genome sequencing in patients with clinically diagnosed Charcot-Marie-Tooth disease when compared with whole-exome sequencing, which has not been reported in the literature. Whole-genome sequencing was performed on 72 families whose genetic cause of clinically diagnosed Charcot-Marie-Tooth disease was not revealed after the whole-exome sequencing and 17p12 duplication screening. Among the included families, 14 (19.4%) acquired genetic diagnoses that were compatible with their phenotypes. The most common factor that led to the additional diagnosis in the whole-genome sequencing was genotype-driven analysis (four families, 4/14), in which a wider range of genes, not limited to peripheral neuropathy-related genes, were analysed. Another four families acquired diagnosis due to the inherent advantage of whole-genome sequencing such as better coverage than the whole-exome sequencing (two families, 2/14), structural variants (one family, 1/14) and non-coding variants (one family, 1/14). In conclusion, an evident gain in diagnostic yield was obtained from whole-genome sequencing of the whole-exome sequencing-negative cases. A wide range of genes, not limited to inherited peripheral neuropathy-related genes, should be targeted during whole-genome sequencing.

Comparing intra-observer variation and external variations of a fully automated cephalometric analysis with a cascade convolutional neural net.

The quality of cephalometric analysis depends on the accuracy of the delineating landmarks in orthodontic and maxillofacial surgery. Due to the extensive number of landmarks, each analysis costs orthodontists considerable time per patient, leading to fatigue and inter- and intra-observer variabilities. Therefore, we proposed a fully automated cephalometry analysis with a cascade convolutional neural net (CNN). One thousand cephalometric x-ray images (2 k × 3 k) pixel were used. The dataset was split into training, validation, and test sets as 8:1:1. The 43 landmarks from each image were identified by an expert orthodontist. To evaluate intra-observer variabilities, 28 images from the dataset were randomly selected and measured again by the same orthodontist. To improve accuracy, a cascade CNN consisting of two steps was used for transfer learning. In the first step, the regions of interest (ROIs) were predicted by RetinaNet. In the second step, U-Net detected the precise landmarks in the ROIs. The average error of ROI detection alone was 1.55 ± 2.17 mm. The model with the cascade CNN showed an average error of 0.79 ± 0.91 mm (paired t-test, p = 0.0015). The orthodontist's average error of reproducibility was 0.80 ± 0.79 mm. An accurate and fully automated cephalometric analysis was successfully developed and evaluated.