Synthesis of fracture radiographs with deep neural networks.

PURPOSE: We describe a machine learning system for converting diagrams of fractures into realistic X-ray images. We further present a method for iterative, human-guided refinement of the generated images and show that the resulting synthetic images can be used during training to increase the accuracy of deep classifiers on clinically meaningful subsets of fracture X-rays. METHODS: A neural network was trained to reconstruct images from programmatically created line drawings of those images. The images were then further refined with an optimization-based technique. Ten physicians were recruited into a study to assess the realism of synthetic radiographs created by the neural network. They were presented with mixed sets of real and synthetic images and asked to identify which images were synthetic. Two classifiers were trained to detect humeral shaft fractures: one only on true fracture images, and one on both true and synthetic images. RESULTS: Physicians were 49.63% accurate in identifying whether images were synthetic or real. This is close to what would be expected by pure chance (i.e. random guessing). A classifier trained only on real images detected fractures with 67.21% sensitivity when no fracture fixation hardware was present. A classifier trained on both real images and synthetic images was 75.54% sensitive. CONCLUSION: Our method generates X-rays realistic enough to be indistinguishable from real X-rays. We also show that synthetic images generated using this method can be used to increase the accuracy of deep classifiers on clinically meaningful subsets of fracture X-rays.

Tanzania IR Initiative: Training the First Generation of Interventional Radiologists.

Despite a population of nearly 60 million, there is currently not a single interventional radiologist in Tanzania. Based on an Interventional Radiology (IR) Readiness Assessment, the key obstacles to establishing IR in Tanzania are the lack of training opportunities and limited availability of disposable equipment. An IR training program was designed and initiated, which relies on US-based volunteer teams of IR physicians, nurses, and technologists to locally train radiology residents, nurses, and technologists. Preliminary results support this strategy for addressing the lack of training opportunities and provide a model for introducing IR to other resource-limited settings.

Predictive value of hyperthermia and intracranial hypertension on neurological outcomes in patients with severe traumatic brain injury.

BACKGROUND: Intracranial hypertension (ICH) and hyperthermia are common after traumatic brain injury (TBI) and associated with worse neurological outcomes. This study sets out to determine the combined power of temperature and intracranial pressure (ICP) for predicting neurologic outcomes and prolonged length of stay (LOS) following severe TBI. METHODS: High resolution (every 6 seconds) temperature and ICP data were collected in adults with severe TBI from 2008-2010. Temperatures were plotted against concurrent ICP and divided based on breakpoints (Temperature: <36, 36-38.5 or >38.5 °C, ICP: <20, 20-30 or >30 mmHg). The percentage of time spent in each section, as well as several pooled unfavourable conditions (hyperthermia ± ICH), were then evaluated for predictive value for ICU-LOS > 7 days and short-term (<6 months) vs. long-term (>6 months) dichotomized neurologic outcomes. RESULTS: Fifty patients were included for analysis with severe TBI. Evaluation of the area under the operating receiver curve (AUC) showed significant periods of fever and high ICP (<30 mmHg) had a strong association with poor long-term neurological outcomes (Day 3, AUC = 0.71, p = 0.04) and were higher than either condition alone. ICU-LOS > 7 days was increased when hyperthermia and/or ICH remained uncontrolled by Day 5 (AUC = 0.82, p = 0.02). SUMMARY: Hyperthermia combined with ICH were shown to be significant prognostic indicators of future poor neurologic outcomes in patients with severe traumatic brain injury.