Reproducibility of Deep Learning Algorithms Developed for Medical Imaging Analysis: A Systematic Review.

Since 2000, there have been more than 8000 publications on radiology artificial intelligence (AI). AI breakthroughs allow complex tasks to be automated and even performed beyond human capabilities. However, the lack of details on the methods and algorithm code undercuts its scientific value. Many science subfields have recently faced a reproducibility crisis, eroding trust in processes and results, and influencing the rise in retractions of scientific papers. For the same reasons, conducting research in deep learning (DL) also requires reproducibility. Although several valuable manuscript checklists for AI in medical imaging exist, they are not focused specifically on reproducibility. In this study, we conducted a systematic review of recently published papers in the field of DL to evaluate if the description of their methodology could allow the reproducibility of their findings. We focused on the Journal of Digital Imaging (JDI), a specialized journal that publishes papers on AI and medical imaging. We used the keyword "Deep Learning" and collected the articles published between January 2020 and January 2022. We screened all the articles and included the ones which reported the development of a DL tool in medical imaging. We extracted the reported details about the dataset, data handling steps, data splitting, model details, and performance metrics of each included article. We found 148 articles. Eighty were included after screening for articles that reported developing a DL model for medical image analysis. Five studies have made their code publicly available, and 35 studies have utilized publicly available datasets. We provided figures to show the ratio and absolute count of reported items from included studies. According to our cross-sectional study, in JDI publications on DL in medical imaging, authors infrequently report the key elements of their study to make it reproducible.

Trends and racial disparities for acute kidney injury in premature infants: the US national database.

BACKGROUND: To assess prevalence and outcomes of acute kidney injury (AKI) in very-low-birth-weight infants. METHODS: This cross-sectional study utilized the National Inpatient Sample (NIS) dataset for years 2000-2017. All premature infants with birth weight (BW) <1500g and/or gestational age (GA) ≤32 weeks were included. Analyses were conducted for overall population and two BW categories: <1000g and 1000-1499g. Adjusted odds ratios were calculated after controlling for confounding variables in logistic regression analysis. Cochrane-Armitage test was used to assess for statistically significant trends in AKI frequency over the years. RESULTS: In total, 1,311,681 hospitalized premature infants were included; 19,603 (1.5%) were diagnosed with AKI. The majority (74.3%) were BW <1000g and 63.9% ≤28 weeks gestation. Prevalence of AKI differed by ethnicity; White had significantly less AKI than Black (OR=0.79, p<0.001) and Hispanic (OR=0.83, p<0.001). AKI was significantly associated with higher mortality compared to controls (35.1 vs. 3.0%, p<0.001). AKI was associated with comorbidities such as necrotizing enterocolitis, patent ductus arteriosus, bronchopulmonary dysplasia, intraventricular hemorrhage, and septicemia. In a regression model, AKI was associated with higher mortality after controlling confounding factors (aOR=7.79, p<0.001). AKI was associated with significant increase in length of stay (p<0.001) and cost of hospitalization in survivors (p<0.001). There is a significant trend for increased AKI frequency over the years (Z score=4.33, p<0.001). CONCLUSION: AKI is associated with increased mortality and comorbidities in preterm infants, especially in infants with BW <1000g. Further studies are needed to understand precipitating factors and assess preventative measures for this serious complication.

Congenital Diarrheal Diseases.

Early diagnosis of diarrhea is critical to prevent disease progression. Diarrhea in newborns can be congenital or acquired; acquired diarrheas are the major cause in infants. Congenital diarrheal diseases are rare and include defects in digestion, absorption, and transport of nutrients, and electrolytes; disorders of enterocyte differentiation and polarization; defects of enteroendocrine cell differentiation; dysregulation of the intestinal immune response; and dysfunction of the immune system. This review discusses the clinical approach that may help in early identification and management of different congenital diarrheal diseases.

Determination and assessment of total mercury levels in local, frozen and canned fish in Lebanon.

Fish is an important constituent of the Lebanese diet. However, very little attention in our area is given to bring awareness regarding the effect of the toxicity of mercury (Hg) mainly through fish consumption. This study aimed to report analytical data on total mercury levels in several fish species for the first time in thirty years and to also made individuals aware of the presence and danger from exposure to mercury through fish consumption. Fish samples were selected from local Lebanese markets and fisheries and included 94 samples of which were fresh, frozen, processed, and canned fish. All values were reported as microgram of mercury per gram of fish based on wet weight. The level of mercury ranged from 0.0190 to 0.5700 microg/g in fresh samples, 0.0059 to 0.0665 microg/g in frozen samples, and 0.0305 to 0.1190 microg/g in canned samples. The data clearly showed that higher levels of mercury were detected in local fresh fish as opposed to other types thus placing consumers at higher risk from mercury exposure. Moreover, the data revealed that Mallifa (yellowstripe barracuda/Sphyraena chrysotaenia), Sargous (white seabream/Diplodus sargus), Ghobbos (bogue/Boops boops), and shrimp (Penaeus sp.) were among the types containing the highest amounts of mercury. On the other hand, processed fish such as fish fillet, fish burger, small shrimp and crab are found to contain lower levels of mercury and are associated with lower exposure risks to mercury. Lebanese population should therefore, be aware to consume limited amounts of fresh local fish to minimize exposure to mercury.