An artificial intelligence deep learning model for identification of small bowel obstruction on plain abdominal radiographs.

OBJECTIVES: Small bowel obstruction is a common surgical emergency which can lead to bowel necrosis, perforation and death. Plain abdominal X-rays are frequently used as a first-line test but the availability of immediate expert radiological review is variable. The aim was to investigate the feasibility of using a deep learning model for automated identification of small bowel obstruction. METHODS: A total of 990 plain abdominal radiographs were collected, 445 with normal findings and 445 demonstrating small bowel obstruction. The images were labelled using the radiology reports, subsequent CT scans, surgical operation notes and enhanced radiological review. The data were used to develop a predictive model comprising an ensemble of five convolutional neural networks trained using transfer learning. RESULTS: The performance of the model was excellent with an area under the receiver operator curve (AUC) of 0.961, corresponding to sensitivity and specificity of 91 and 93% respectively. CONCLUSION: Deep learning can be used to identify small bowel obstruction on plain radiographs with a high degree of accuracy. A system such as this could be used to alert clinicians to the presence of urgent findings with the potential for expedited clinical review and improved patient outcomes. ADVANCES IN KNOWLEDGE: This paper describes a novel labelling method using composite clinical follow-up and demonstrates that ensemble models can be used effectively in medical imaging tasks. It also provides evidence that deep learning methods can be used to identify small bowel obstruction with high accuracy.

Comparative appearances of non-tuberculous mycobacteria species: a CT study.

Our objective was to characterise the CT features of the various species of non-tuberculous mycobacteria (NTM) and to identify differences, if any, between Mycobacterium avium intracellulare( MAI) and other species. Fifty-five patients, who were culture positive on at least two occasions for a single NTM species, were evaluated. All patients had CT scans performed within 6 months of NTM identification. The CT scans were assessed for the presence and severity of bronchiectasis, nodules, cavities, tree-in-bud pattern, consolidation and for evidence of pre-existing lung disease. Bronchiectasis was identified in most patients (52 of 55, 95%) and nodules were present in approximately half (29 of 55, 53%). Patients with MAI ( n=16) were found to have significantly higher bronchiectasis scores and higher prevalence of nodules than the other species (both p<0.01). Patients with M.kansasii ( n=9) and M.xenopi ( n=9) had cavities, tree-in-bud pattern, and pre-existing emphysema as the dominant CT features. Patients with M.chelonae and M.fortuitum were younger than the other groups and also had a high incidence of pre-existing lung disease. Patients with MAI infection have more severe bronchiectasis and more nodules on CT than the other NTM species. Morphological differences between the other species were identified but were less distinct.

Abnormal osteocalcin binding in rheumatoid arthritis.

Studies of osteocalcin in the serum and synovial fluid of patients with rheumatoid arthritis (RA) and osteoarthritis (OA) showed the presence of significant amounts of osteocalcin in synovial fluid and that the values in RA synovial fluid were significantly lower than in OA synovial fluid. In addition, the osteocalcin in OA synovial fluid bound almost completely to hydroxyapatite, whereas a significant proportion of the osteocalcin in RA synovial fluid did not. These studies suggest that patients with severe RA produce low amounts of active osteocalcin and higher than expected amounts of inactive osteocalcin in the synovial fluid. They provide some evidence that osteoblast function may be abnormal in the osteoporosis of RA.