Assist radiologists and emergency clinicians in detecting fractures during review of radiographs of MSK system
Computer-assisted detection and diagnosis (CAD) software device to assist radiologists and emergency physicians in detecting fractures during the review of radiographs of the musculoskeletal system. Rayvolve is indicated for adult and pediatric population (≥ 2 years). Rayvolve is indicated for radiographs of the following industry-standard radiographic views and study types.
Radiographs of the musculoskeletal system
Fracture Detection in DICOM
A validation study involving 2634 radiography sets from 2549 children was conducted. The average age was 8.5 years (SD ± 4.5), with a nearly even gender split (1459 boys and 1090 girls). Rayvolve demonstrated high diagnostic performance:
Sensitivity: 95.7% (95% CI: 94.0–96.9)
Specificity: 91.2% (95% CI: 89.8–92.5)
Accuracy: 92.6% (95% CI: 91.5–93.6)
Positive Predictive Value (PPV): 89.9% (95% CI: 87.8–91.8), indicating the likelihood that detected fractures were true positives.
Negative Predictive Value (NPV): 96.5% (95% CI: 95.2–97.5), showing Rayvolve's ability to reliably rule out fractures when none are present
These metrics indicate Rayvolve’s high reliability in fracture detection, particularly in paediatric cases, across detection, enumeration, and localisation approaches, with the specificity slightly lower when children were wearing casts. Statistical tests included chi-square and Fisher’s exact test, yielding significant p-values (p < 0.001), especially in cases without casts where the algorithm maintained a high negative predictive value
A study found using Rayvolve in clinical workflows showed a notable reduction in radiologist workload by automating the detection and localisation of fractures on radiographs, although specific time-savings data weren’t detailed for each case.
Workflow Integration: Rayvolve was found effective in emergency settings, particularly for rapid paediatric assessments, as it integrates with PACS systems and automatically highlights suspected fracture regions, allowing radiologists to prioritise cases more efficiently.
A study found improving diagnostic accuracy and reducing false positives, Rayvolve minimises the need for follow-up imaging and additional diagnostic procedures. For example, with a false-positive reduction of 8.8% (from 17% to 8.2%), the system effectively decreases unnecessary follow-up costs and potential additional exposure to radiation for pediatric patients. The high sensitivity and specificity reduce the likelihood of missed fractures and associated complications. By preventing missed diagnoses, Rayvolve helps avoid the added costs of delayed fracture treatments and hospital readmissions, contributing to overall healthcare savings in pediatric emergency care.
Evaluation of the Performance of an Artificial Intelligence (AI)Algorithm in Detecting Thoracic Pathologies on Chest Radiographs; Hubert Bettinger; Gregory Lenczner; Jean Guigui; Luc Rotenberg; Elie Zerbib; Alexandre Attia; Julien Vidal and; Pauline Beaumel (2024)
https://www-sciencedirect-com.libproxy.ucl.ac.uk/science/article/pii/S1076633223003112
