A Deep Learning-Based System for the Assessment of Dental Caries Using Colour Dental Photographs.

D1ental caries remains the most common chronic disease in childhood, affecting almost half of all children globally. Dental care and examination of children living in remote and rural areas is an ongoing challenge that has been compounded by COVID. The development of a validated system with the capacity to screen large numbers of children with some degree of automation has the potential to facilitate remote dental screening at low costs. In this study, we aim to develop and validate a deep learning system for the assessment of dental caries using color dental photos. Three state-of-the-art deep learning networks namely VGG16, ResNet-50 and Inception-v3 were adopted in the context. A total of 1020 child dental photos were used to train and validate the system. We achieved an accuracy of 79% with precision and recall respectively 95% and 75% in classifying 'caries' versus 'sound' with inception-v3.

A multidisciplinary quality improvement program increases the inferior vena cava filter retrieval rate.

Published reports indicate low retrieval rates for retrievable inferior vena cava (IVC) filters. We performed a historic-controlled study of a 5-year intervention (March 2007 to February 2012) to improve IVC filter retrieval rates at a university medical center serving a rural area. All adults with a retrievable filter placed were included, except those with a life expectancy <6 months. The intervention included initial verbal counseling and printed educational materials, correspondence after discharge, and a hematology consultation. The control group included patients with retrievable filters placed in the 15 months preceding study initiation. In the control group, 116 filters were placed and 27 (23%) were removed, compared to 378 filters placed and 169 (45%) removed during the intervention. Adjusting for patient characteristics, the odds ratio of retrieval during the intervention was 3.03 (95% CI 1.85-4.27) compared to the control period. An intervention including patient education and hematology follow-up appeared to significantly improve IVC filter retrieval rates.

The Assessment of Digital Study Models Using the GOSLON Yardstick Index.

Objective : To investigate the reproducibility and repeatability of digital models of patients with a unilateral cleft lip and palate (UCLP) using the GOSLON yardstick. Design : Reproducibility and repeatability study. Method : Two examiners used the GOSLON yardstick to assess the intermaxillary dentoalveolar relationship of 30 consecutive UCLP patients by analyzing their 9-year (±3 months) dental study casts and digital study models. The records were rated 1 week apart to avoid bias. The process was repeated 1 month later as a measure of reproducibility. Reliability was assessed by comparing the GOSLON score achieved between the two modalities. Patient dental study casts were sent to 3M Unitek Australia to be scanned using the 3M Unitek Lava(TM) system to produce digital study models. The accuracy of the dental study cast occlusal registration was assessed by both raters prior to sending the study models for scanning. Statistical analysis : The Linear Weighted Kappa statistic and Kendall's Coefficient of Concordance statistic were used to determine the levels of agreement within and between raters. Results : The linear weighted Kappa statistic for intrarater repeatability of digital study models scores were very high (0.89 and 0.97). This compared favorably to the intrarater repeatability of study model casts scores (0.86 and 0.97). There was very good agreement for interrater digital study model scores (0.80 and 0.87) and also for the interrater study model casts scores (0.80 and 0.90). Kendall's Coefficient of Concordance statistic (0.99) and Correlation Coefficient (0.86) support the weighted Kappa results of the digital study model scores. Conclusion : Digital models can be used for GOSLON scoring with a high degree of reproducibility and repeatability.

Effects of blood conservation on the incidence of anemia and transfusions in pediatric parapneumonic effusion: a hospitalist perspective.

BACKGROUND: Children with pneumonia may develop parapneumonic effusion (PNE). The associated inflammatory process and nutritional compromise can blunt erythropoesis. Traditional treatment for these children with PNE includes repeated phlebotomy and surgical intervention, resulting in ongoing blood losses. Blood transfusions used to treat acquired anemia are associated with multiple complications. OBJECTIVES: This study evaluated the effect of hospitalists' implementation of blood conservation guidelines (BCG) on the incidence of anemia and transfusion requirements in children with PNE. DESIGN: Retrospective cohort study of hospitalized children with PNE. SETTINGS: University affiliated Children's Hospital. PATIENTS: Children who were admitted to the hospital with PNE and managed using BCG (Group I) were compared to simultaneous no intervention group (S) and historical no intervention group (H). Group (I) and (S) were admitted from year 2000 to 2004 and the Group (H) were admitted from year 1997 to 1999. MEASUREMENTS: Phlebotomy frequency and volume, measured hemoglobin (Hgb) levels, and the need for red blood transfusions. RESULTS: Children in the BCG group (n = 24) compared to simultaneous no intervention group (n = 28) and historical no intervention group (n = 29) had lesser phlebotomy volumes (14 ± 8, 18 ± 14 and 69 ± 66 mL; P = 0.001), trend toward lesser Hgb drop (1.7 ± 1.4, 2.1 ± 1.2 and 2 ± 1.4 gm%; P ≤ 0.37), and lesser incidence of transfusion (8%, 18% and 31%; P = 0.11). Transfused children were younger (3.5 ± vs. 6.4 ± 4 years; P = 0.001) and had lower initial Hgb (9.9 ± 1 vs. 11.4 ± 1 gm%; P = 0.001), more phlebotomy (5.9 ± 7 vs. 1.1 ± 1 mL/kg., P = 0.001), longer hospitalization (18.7 ± 5 vs. 11.1 ± days; P = 0.001), and slightly higher (pediatric risk of mortality [PRISM]) scores (3.4 ± 5.7 vs. 1.6 ± 2.7; P = 0.25). CONCLUSION: Implementing BCG lowers phlebotomy losses and the need for transfusion.