What happened
16 Bit of Toronto makes Rho, a locked algorithm trained by machine learning that reviews x-rays of patients aged 50 and older for possible low bone density. It reads frontal x-rays of the spine, chest, pelvis, knee or hand and wrist, with the patient's age and sex. The company tested it retrospectively on 3,729 cases from six True North Imaging centers in Canada and on two US datasets. DXA scans of the same patients served as the reference. FDA reviewers report that the device met its performance goals of specificity and AUC in all three datasets. Specificity was 0.90 to 0.94, while sensitivity was 0.67 in the Canadian data and 0.45 and 0.36 in the US data. Sensitivity fell below 30% in some subgroups, such as knee x-rays in the US data. FDA granted the De Novo request on 9 April 2024. FDA judged that the low sensitivity points to a small probable benefit and that missed cases carry no more risk than current care.
Sources
-
Regulatory document, by authors with no tie to the system.
U.S. Food and Drug Administration, Center for Devices and Radiological Health. De Novo classification request for Rho (DEN230023): decision summary. Granted 2024-04-09.
The device met its performance goals of specificity and AUC in all three datasets.
Details
AI system
- Workflow
-
The software read frontal x-rays taken for other clinical reasons, with the patient's age and sex. A report went to the radiologist only for positive cases.
Follow-up
- Conditions for replication
-
The patient is 50 or older and has an eligible frontal x-ray. A negative result must not be read as ruling out low bone density.
Entry details
- Sector
-
Healthcare.
- Purpose
-
16 Bit built Rho to flag patients with possible low bone density on x-rays taken for other reasons, to prompt an assessment of bone health.
Not yet checked: model, release, provider, role, configuration, inputs and outputs, autonomy, human oversight, AI contribution, beneficiaries, reach, distribution and access, duration, every measurement question, every money question, ownership and control, dependencies, cross-border connections, geopolitical implications, replication attempts, availability, institutional decision, grounds, action and result, success criterion, every evidence question, observed or intended, alternative assessments, related entries, other identifiers, category mappings.
Cite this record
Peter H. Vartanian. (2026). Rho flags low bone density on routine x-rays but misses many cases (AIBD-0769). Artificial Intelligence Benefits Database, edition of 31 October 2026. https://petervartanian.xyz/aibd/records/AIBD-0769.html (archived at https://doi.org/10.7910/DVN/AADSII, version 4.0)