Artificial intelligence-based automated preprocessing and classification of impacted maxillary canines in panoramic radiographs

OBJECTIVES

Automating the digital workflow for diagnosing impacted canines using panoramic radiographs (PRs) is challenging. This study explored feature extraction, automated cropping, and classification of impacted and nonimpacted canines as a first step.

METHODS

A convolutional neural network with SqueezeNet architecture was first trained to classify two groups of PRs (91with and 91without impacted canines) on the MATLAB programming platform. Based on results, the need to crop the PRs was realized. Next, artificial intelligence (AI) detectors were trained to identify specific landmarks (maxillary central incisors, lateral incisors, canines, bicuspids, nasal area, and the mandibular ramus) on the PRs. Landmarks were then explored to guide cropping of the PRs. Finally, improvements in classification of automatically cropped PRs were studied.

RESULTS

Without cropping, the area under the curve (AUC) of the receiver operating characteristic (ROC) curve for classifying impacted and nonimpacted canine was 84%. Landmark training showed that detectors could correctly identify upper central incisors and the ramus in ∼98% of PRs. The combined use of the mandibular ramus and maxillary central incisors as guides for cropping yielded the best results (∼10% incorrect cropping). When automatically cropped PRs were used, the AUC-ROC improved to 96%.

CONCLUSIONS

AI algorithms can be automated to preprocess PRs and improve the identification of impacted canines.

Characterizing the Inflammatory Profile of Neutrophil-Rich Triple-Negative Breast Cancer

Breast cancer (BC) is one of the most common types of cancer in women in the United Arab Emirates. Immunogenic tumours, such as triple-negative breast cancer (TNBC), show increased neutrophil infiltration, which is associated with poor prognosis and limited efficacy of immunotherapy. This study aims to investigate in vitro the bidirectional effect of neutrophils on metastatic TNBC (MDA-MB-231) compared to less-metastatic luminal breast cancer (MCF-7) cell lines. We found that BC cells or their conditioned medium (CM) reduced the viability of neutrophil-like cells (HL60). This was supported by increased cellular stress and NETosis in differentiated HL60 cells (dHL60) upon exposure to MDA-MB-231 compared to MCF-7-CM using nucleic acid staining essays. Flow cytometry showed comparable expression of inflammatory markers by polymorphonuclear cells (PMN) when treated with MDA-MB-231-CM and standard polarizing cocktails. Furthermore, MDA-MB-231-CM triggered an inflammatory pattern with evidence of stronger adhesion (CD62L) and degranulation (CD11b and CD66b) phenotypes. The proinflammatory polarization of dHL60 by MDA-MB-231-CM was additionally confirmed by the elevated CD54 expression, myeloperoxidase, and CD11b protein levels, which matched an increased transwell migratory capacity. In conclusion, BC might use neutrophils to their benefit through NETosis and complement system activation, which makes this crosstalk a potential mechanism for understanding tumour progression.

Mutational spectrum and phenotypic variability of Duchenne muscular dystrophy and related disorders in a Bangladeshi population

Duchenne muscular dystrophy (DMD) is a severe rare neuromuscular disorder caused by mutations in the X-linked dystrophin gene. Several mutations have been identified, yet the full mutational spectrum, and their phenotypic consequences, will require genotyping across different populations. To this end, we undertook the first detailed genotype and phenotype characterization of DMD in the Bangladeshi population. We investigated the rare mutational and phenotypic spectrum of the DMD gene in 36 DMD-suspected Bangladeshi participants using an economically affordable diagnostic strategy involving initial screening for exonic deletions in the DMD gene via multiplex PCR, followed by testing PCR-negative patients for mutations using whole exome sequencing. The deletion mapping identified two critical DMD gene hotspot regions (near proximal and distal ends, spanning exons 8-17 and exons 45-53, respectively) that comprised 95% (21/22) of the deletions for this population cohort. From our exome analysis, we detected two novel pathogenic hemizygous mutations in exons 21 and 42 of the DMD gene, and novel pathogenic recessive and loss of function variants in four additional genes: SGCD, DYSF, COL6A3, and DOK7. Our phenotypic analysis showed that DMD suspected participants presented diverse phenotypes according to the location of the mutation and which gene was impacted. Our study provides ethnicity specific new insights into both clinical and genetic aspects of DMD.

Overlapping pathogenic de novo CNVs in neurodevelopmental disorders and congenital anomalies impacting constraint genes regulating early development

Neurodevelopmental disorders (NDDs) and congenital anomalies (CAs) are rare disorders with complex etiology. In this study, we investigated the less understood genomic overlap of copy number variants (CNVs) in two large cohorts of NDD and CA patients to identify de novo CNVs and candidate genes associated with both phenotypes. We analyzed clinical microarray CNV data from 10,620 NDD and 3176 CA cases annotated using Horizon platform of GenomeArc Analytics and applied rigorous downstream analysis to evaluate overlapping genes from NDD and CA CNVs. Out of 13,796 patients, only 195 cases contained 218 validated de novo CNVs. Eighteen percent (31/170) de novo CNVs in NDD cases and 40% (19/48) de novo CNVs in CA cases contained genomic overlaps impacting developmentally constraint genes. Seventy-nine constraint genes (10.1% non-OMIM entries) were found to have significantly enriched genomic overlap within rare de novo pathogenic deletions (P value = 0.01, OR = 1.58) and 45 constraint genes (13.3% non-OMIM entries) within rare de novo pathogenic duplications (P value = 0.01, OR = 1.97). Analysis of spatiotemporal transcriptome demonstrated both pathogenic deletion and duplication genes to be highly expressed during the prenatal stage in human developmental brain (P value = 4.95 X 10-6). From the list of overlapping genes, EHMT1, an interesting known NDD gene encompassed pathogenic deletion CNVs from both NDD and CA patients, whereas FAM189A1, and FSTL5 are new candidate genes from non-OMIM entries. In summary, we have identified constraint overlapping genes from CNVs (including de novo) in NDD and CA patients that have the potential to play a vital role in common disease etiology.

MERS-CoV Found in Hyalomma dromedarii Ticks Attached to Dromedary Camels at a Livestock Market, United Arab Emirates, 2019

The main mode of transmission of Middle East respiratory syndrome-related coronavirus (MERS-CoV) between dromedaries is likely via the respiratory route. However, there must be other modes to explain how the infection is brought to MERS-CoV-negative closed herds, such as transmission by ticks. Here, we present a study performed at three different locations in the United Arab Emirates (UAE) involving 215 dromedary camels (Camelus dromedarius) and the ticks attached to them. We tested the camels and ticks via RT-(q)PCR for the presence of MERS-CoV nucleic acids, as well as flaviviruses that may be present in the region (e.g., Alkhumra hemorrhagic fever virus). Camel sera were additionally analyzed for evidence of previous exposure to MERS-CoV. In total, 8 out of 242 tick pools were positive for MERS-CoV RNA (3.3%; C_t 34.6-38.3), 7 of which contained Hyalomma dromedarii ticks, and one contained a Hyalomma sp. tick (species not identified). All of the virus-positive ticks' host camels were also positive for MERS-CoV RNA in their nasal swab samples. Short sequences established in the N gene region from two positive tick pools were identical to viral sequences from their hosts' nasal swabs. In total, 59.3% of dromedaries at the livestock market had MERS-CoV RNA in their nasal swabs (C_t 17.7-39.5). While dromedaries at all locations were negative for MERS-CoV RNA in their serum samples, antibodies were detected in 95.2% and 98.7% of them (tested by ELISA and indirect immunofluorescence test, respectively). Given the probably transient and/or low level of MERS-CoV viremia in dromedaries and the rather high C_t values observed in the ticks, it seems unlikely that Hyalomma dromedarii is a competent vector for MERS-CoV; however, its role in mechanical or fomite transmission between camels should be investigated.

Development and Comparison of Conventional and 3D-Printed Laboratory Models of Maxillary Defects

BACKGROUND

Recording accurate impressions from maxillary defects is a critical and challenging stage in the prosthetic rehabilitation of patients following maxillectomy surgery. The aim of this study was to develop and optimize conventional and 3D-printed laboratory models of maxillary defects and to compare conventional and digital impression techniques using these models.

METHODS

Six different types of maxillary defect models were fabricated. A central palatal defect model was used to compare conventional silicon impressions with digital intra-oral scanning in terms of dimensional accuracy and total time taken to record the defect and produce a laboratory analogue.

RESULTS

Digital workflow produced different results than the conventional technique in terms of defect size measurements which were statistically significant (p < 0.05). The time taken to record the arch and the defect using an intra-oral scanner was significantly less compared with the traditional impression method. However, there was no statistically significant difference between the two techniques in terms of the total time taken to fabricate a maxillary central defect model (p > 0.05).

CONCLUSIONS

The laboratory models of different maxillary defects developed in this study have the potential to be used to compare conventional and digital workflow in prosthetic treatment procedures.

Estimating aspect ratio of the distal femur and proximal tibia in the Emirati population from MRI scans of the knee: a preliminary experience

Size and shape of knee implants play an important role in the success of total knee arthroplasty. Several studies have identified anthropometric differences of the distal femur and proximal tibia between the genders and ethnicities. Ill-fitting prosthesis can cause overhang or under-fit resulting in persistence of pain, periprosthetic fracture and decreased range of motion. The purpose of this study was to estimate the aspect ratio of distal femur and proximal tibia in the Emirati population and determine whether gender differences exist within this group. Magnetic resonance imaging datasets of unilateral knees scans performed on adult Emirati patients at a tertiary care hospital were retrospectively examined. Knee parameters were obtained from 65 males and 46 females (n = 111). Females showed significantly smaller AP and ML dimensions of distal femur and lower aspect ratios compared to males (p < 0.001). Proximal tibial dimensions (AP and ML) of Emirati women are also significantly smaller compared to men. However, aspect ratio of proximal tibia did not show gender variation (p = 0.956) within the Emirati population. Emirati knees showed significant gender differences in bony dimensions and aspect ratio of the knee, and also have smaller aspect ratios when compared with most other population groups.

Crimean-Congo Hemorrhagic Fever Virus Past Infections Are Associated with Two Innate Immune Response Candidate Genes in Dromedaries

Dromedaries are an important livestock, used as beasts of burden and for meat and milk production. However, they can act as an intermediate source or vector for transmitting zoonotic viruses to humans, such as the Middle East respiratory syndrome coronavirus (MERS-CoV) or Crimean-Congo hemorrhagic fever virus (CCHFV). After several outbreaks of CCHFV in the Arabian Peninsula, recent studies have demonstrated that CCHFV is endemic in dromedaries and camel ticks in the United Arab Emirates (UAE). There is no apparent disease in dromedaries after the bite of infected ticks; in contrast, fever, myalgia, lymphadenopathy, and petechial hemorrhaging are common symptoms in humans, with a case fatality ratio of up to 40%. We used the in-solution hybridization capture of 100 annotated immune genes to genotype 121 dromedaries from the UAE tested for seropositivity to CCHFV. Through univariate linear regression analysis, we identified two candidate genes belonging to the innate immune system: FCAR and CLEC2B. These genes have important functions in the host defense against viral infections and in stimulating natural killer cells, respectively. This study opens doors for future research into immune defense mechanisms in an enzootic host against an important zoonotic disease.