Pneumothorax Triage Algorithm Prioritizes X-Ray Scheduling
By MedImaging International staff writers Posted on 26 Sep 2019 |

Image: A novel X-Ray scheduling algorithm prioritizes pneumothorax triage X-Ray scheduling (Photo courtesy of GE Healthcare).
A novel artificial intelligence (AI) algorithm analyzes chest x-rays for pneumothorax and helps flag suspected cases for radiologists to prioritize reading.
The GE Healthcare (GE; Little Chalfont, United Kingdom) Critical Care Suite is designed to automatically scan images for pneumothorax immediately upon acquisition, without any additional infrastructure or IT networking. Notifications arrive in picture archiving and communication systems (PACS) at the exact same time as the original digital imaging and communications in medicine (DICOM) image, causing no additional delay or processing time, thus enabling the escalation of image review and shorter turn-around times for prioritized cases.
Developed using more than 12,000 images from six total data sources in three different countries, the pneumothorax triage algorithm detects nearly all large pneumothoraxes, with 96% sensitivity, 75% sensitivity, and 94% specificity. The positive predictive value (PPV) ranges from 35%-70%, depending on the prevalence of pneumothorax; for 4% prevalence, the algorithm had one true notification per two false notifications, and for 15% prevalence, it had seven true notifications per three false notifications. The negative predictive value ranged from 97-99%, based on the same prevalence range.
“Since thousands of chest x-rays are taken in a hospital every day, looking for pneumothorax in critical patients is kind of like trying to find a needle in a haystack,” said Rachael Callcut, MD, MSPH, director of data science for the Center for Digital Health Innovation at the University of California, San Francisco (UCSF), who was involved in the development of the Critical Care Suite. “The algorithms themselves are not making diagnoses, but draw everyone's attention to finding something potentially life-threatening for patients.”
Primary spontaneous pneumothorax is an abnormal accumulation of air in the pleural space that can result in the partial or complete collapse of a lung. It is likely due to the formation of small sacs of air (blebs) in lung tissue that rupture, causing air to leak into the pleural space, creating pressure that is manifest as chest pain on the side of the collapsed lung and shortness of breath. Often, people who experience a primary spontaneous pneumothorax have no prior sign of illness; the blebs themselves typically do not cause any symptoms and are visible only on medical imaging. Affected individuals may have one to more than thirty blebs.
The GE Healthcare (GE; Little Chalfont, United Kingdom) Critical Care Suite is designed to automatically scan images for pneumothorax immediately upon acquisition, without any additional infrastructure or IT networking. Notifications arrive in picture archiving and communication systems (PACS) at the exact same time as the original digital imaging and communications in medicine (DICOM) image, causing no additional delay or processing time, thus enabling the escalation of image review and shorter turn-around times for prioritized cases.
Developed using more than 12,000 images from six total data sources in three different countries, the pneumothorax triage algorithm detects nearly all large pneumothoraxes, with 96% sensitivity, 75% sensitivity, and 94% specificity. The positive predictive value (PPV) ranges from 35%-70%, depending on the prevalence of pneumothorax; for 4% prevalence, the algorithm had one true notification per two false notifications, and for 15% prevalence, it had seven true notifications per three false notifications. The negative predictive value ranged from 97-99%, based on the same prevalence range.
“Since thousands of chest x-rays are taken in a hospital every day, looking for pneumothorax in critical patients is kind of like trying to find a needle in a haystack,” said Rachael Callcut, MD, MSPH, director of data science for the Center for Digital Health Innovation at the University of California, San Francisco (UCSF), who was involved in the development of the Critical Care Suite. “The algorithms themselves are not making diagnoses, but draw everyone's attention to finding something potentially life-threatening for patients.”
Primary spontaneous pneumothorax is an abnormal accumulation of air in the pleural space that can result in the partial or complete collapse of a lung. It is likely due to the formation of small sacs of air (blebs) in lung tissue that rupture, causing air to leak into the pleural space, creating pressure that is manifest as chest pain on the side of the collapsed lung and shortness of breath. Often, people who experience a primary spontaneous pneumothorax have no prior sign of illness; the blebs themselves typically do not cause any symptoms and are visible only on medical imaging. Affected individuals may have one to more than thirty blebs.
Latest Radiography News
- World's Largest Class Single Crystal Diamond Radiation Detector Opens New Possibilities for Diagnostic Imaging
- AI-Powered Imaging Technique Shows Promise in Evaluating Patients for PCI
- Higher Chest X-Ray Usage Catches Lung Cancer Earlier and Improves Survival
- AI-Powered Mammograms Predict Cardiovascular Risk
- Generative AI Model Significantly Reduces Chest X-Ray Reading Time
- AI-Powered Mammography Screening Boosts Cancer Detection in Single-Reader Settings
- Photon Counting Detectors Promise Fast Color X-Ray Images
- AI Can Flag Mammograms for Supplemental MRI
- 3D CT Imaging from Single X-Ray Projection Reduces Radiation Exposure
- AI Method Accurately Predicts Breast Cancer Risk by Analyzing Multiple Mammograms
- Printable Organic X-Ray Sensors Could Transform Treatment for Cancer Patients
- Highly Sensitive, Foldable Detector to Make X-Rays Safer
- Novel Breast Cancer Screening Technology Could Offer Superior Alternative to Mammogram
- Artificial Intelligence Accurately Predicts Breast Cancer Years Before Diagnosis
- AI-Powered Chest X-Ray Detects Pulmonary Nodules Three Years Before Lung Cancer Symptoms
- AI Model Identifies Vertebral Compression Fractures in Chest Radiographs
Channels
MRI
view channel
AI Tool Tracks Effectiveness of Multiple Sclerosis Treatments Using Brain MRI Scans
Multiple sclerosis (MS) is a condition in which the immune system attacks the brain and spinal cord, leading to impairments in movement, sensation, and cognition. Magnetic Resonance Imaging (MRI) markers... Read more
Ultra-Powerful MRI Scans Enable Life-Changing Surgery in Treatment-Resistant Epileptic Patients
Approximately 360,000 individuals in the UK suffer from focal epilepsy, a condition in which seizures spread from one part of the brain. Around a third of these patients experience persistent seizures... Read more
AI-Powered MRI Technology Improves Parkinson’s Diagnoses
Current research shows that the accuracy of diagnosing Parkinson’s disease typically ranges from 55% to 78% within the first five years of assessment. This is partly due to the similarities shared by Parkinson’s... Read more
Biparametric MRI Combined with AI Enhances Detection of Clinically Significant Prostate Cancer
Artificial intelligence (AI) technologies are transforming the way medical images are analyzed, offering unprecedented capabilities in quantitatively extracting features that go beyond traditional visual... Read moreUltrasound
view channel
AI Identifies Heart Valve Disease from Common Imaging Test
Tricuspid regurgitation is a condition where the heart's tricuspid valve does not close completely during contraction, leading to backward blood flow, which can result in heart failure. A new artificial... Read more
Novel Imaging Method Enables Early Diagnosis and Treatment Monitoring of Type 2 Diabetes
Type 2 diabetes is recognized as an autoimmune inflammatory disease, where chronic inflammation leads to alterations in pancreatic islet microvasculature, a key factor in β-cell dysfunction.... Read moreNuclear Medicine
view channel
Novel PET Imaging Approach Offers Never-Before-Seen View of Neuroinflammation
COX-2, an enzyme that plays a key role in brain inflammation, can be significantly upregulated by inflammatory stimuli and neuroexcitation. Researchers suggest that COX-2 density in the brain could serve... Read more
Novel Radiotracer Identifies Biomarker for Triple-Negative Breast Cancer
Triple-negative breast cancer (TNBC), which represents 15-20% of all breast cancer cases, is one of the most aggressive subtypes, with a five-year survival rate of about 40%. Due to its significant heterogeneity... Read moreGeneral/Advanced Imaging
view channel
AI-Powered Imaging System Improves Lung Cancer Diagnosis
Given the need to detect lung cancer at earlier stages, there is an increasing need for a definitive diagnostic pathway for patients with suspicious pulmonary nodules. However, obtaining tissue samples... Read more
AI Model Significantly Enhances Low-Dose CT Capabilities
Lung cancer remains one of the most challenging diseases, making early diagnosis vital for effective treatment. Fortunately, advancements in artificial intelligence (AI) are revolutionizing lung cancer... Read moreImaging IT
view channel
New Google Cloud Medical Imaging Suite Makes Imaging Healthcare Data More Accessible
Medical imaging is a critical tool used to diagnose patients, and there are billions of medical images scanned globally each year. Imaging data accounts for about 90% of all healthcare data1 and, until... Read more
Global AI in Medical Diagnostics Market to Be Driven by Demand for Image Recognition in Radiology
The global artificial intelligence (AI) in medical diagnostics market is expanding with early disease detection being one of its key applications and image recognition becoming a compelling consumer proposition... Read moreIndustry News
view channel
GE HealthCare and NVIDIA Collaboration to Reimagine Diagnostic Imaging
GE HealthCare (Chicago, IL, USA) has entered into a collaboration with NVIDIA (Santa Clara, CA, USA), expanding the existing relationship between the two companies to focus on pioneering innovation in... Read more
Patient-Specific 3D-Printed Phantoms Transform CT Imaging
New research has highlighted how anatomically precise, patient-specific 3D-printed phantoms are proving to be scalable, cost-effective, and efficient tools in the development of new CT scan algorithms... Read more
Siemens and Sectra Collaborate on Enhancing Radiology Workflows
Siemens Healthineers (Forchheim, Germany) and Sectra (Linköping, Sweden) have entered into a collaboration aimed at enhancing radiologists' diagnostic capabilities and, in turn, improving patient care... Read more