BoneXpert is a medical software device running on a Windows PC or virtual machine in the hospital network.
BoneXpert is FDA 510(k) cleared (K262390) as a Class II device, product code QIH, 21 CFR 892.2050.
Indications for Use
BoneXpert is designed to view and quantify bone age from 2D Posterior Anterior (PA) hand radiographs using machine learning techniques to aid in the analysis of bone age assessment of patients between 2-21 years old for pediatric radiologists. The results should not be relied upon alone to make diagnostic decisions. The images shall be without any major bone deformity and without excessive image postprocessing (edge enhancement).
BoneXpert is installed as a DICOM node. Images are sent from PACS or the X-ray modality, and an annotated image with the Greulich-Pyle bone age is returned to the same study in PACS for review by the radiologist. The hand radiograph must be provided as a DICOM file.
The US version of BoneXpert does not include Tanner-Whitehouse bone age, the Bone Health Index or the BoneXpert adult height prediction model.
BoneXpert is licensed through an annual license fee based on an estimate of the number of annual analyses made. Contact us for a free trial period.
The Instructions for Use are available in the Help menu of BoneXpert. Adverse events may be reported to the FDA through MedWatch (www.fda.gov/medwatch).
Performance
| Range of application | Bone age range | Greulich-Pyle: approximately 2-19 years for males and 2-18 years for females |
| Patient age | 2-21 years | |
| Intended users | Pediatric radiologists | |
| Contraindications | Contraindicated for use in children with severe anatomical abnormalities, such as achondroplasia, and in children with severely abnormal poses of the hand, including patients with cerebral palsy. | |
| Bone Age | Bone age scale | Greulich-Pyle (GP), including carpal bone age for younger children |
| Bone age accuracy | 0.52 years (RMS error relative to the average of 4 raters) 0.41 years (relative to the average of infinitely many raters; RMS error, model-based estimate) |
|
| Large deviations | 1.7% of bone ages deviate by more than 1.5 years from the average of 3 raters (6.4% for human raters) | |
| Bone age precision, SD | 0.08 years (repeat X-rays of the same child taken on the same day with the same modality) | |
| Bias from change of modality or post-processing | Up to 0.2 years | |
| Images accepted for analysis | 99% (images without severe abnormal anatomy or excessive post-processing) | |
| Supplementary outputs | Bone age SDS and SD | Bone age standard deviation scores for Caucasian, African American, Hispanic and Asian reference populations, and bone age SD from the Greulich-Pyle atlas (Brush) |
| Percent Mature Height | Bayley-Pinneau tables (1952) |