Search Results

Bindex measures apparent cortical bone thickness at the proximal tibia and can be used in conjunction with other clinical risk factors or patient characteristics as an aid to the physician in the diagnosis of osteoporosis and other medical conditions leading to reduced bone strength and in the determination of fracture risk.

Device Description

The Bindex system includes ultrasound pulser, transducer and software. Bindex is connected to the USB port of a computer and controlled with computer software. Bindex is used for measurement of cortical bone thickness and it provides Density Index (DI), a parameter which estimates bone mineral density at the hip as measured with DXA. For measurements, gel is applied on skin and ultrasound transducer is manually placed on the measurement location. Standardized measurement location is at proximal (1/3 length) of tibia. Transducer is manually oriented perpendicularly to the surface of the cortical bone to achieve accepted measurement. Measurement is repeated five times at each measurement location. Finally, transducer is disinfected by wiping gel off with disinfective solution moistened cloth or tissue.

AI/ML Overview

Here's a breakdown of the acceptance criteria and the study proving the device meets them, based on the provided text:

Acceptance Criteria and Device Performance

The core of the device's effectiveness lies in its ability to aid in the diagnosis of osteoporosis and determination of fracture risk, performing comparably to the predicate device and being able to identify osteoporotic and non-osteoporotic subjects correctly with certain thresholds.

Acceptance Criteria Category	Specific Criteria	Reported Device Performance
Primary Effectiveness Endpoint	Aid in diagnosis of osteoporosis and determination of fracture risk in conjunction with other clinical risk factors.	Device measures apparent cortical bone thickness at the proximal tibia and provides Density Index (DI), which estimates BMD at the hip as measured by DXA.
	Using Density Index (DI) in conjunction with DXA for patients identified by Bindex (yellow area) will identify 90% of osteoporotic and non-osteoporotic subjects correctly.	This was "realized" in Karjalainen et al. 2016 and "verified" in the U.S. study (Schousboe et al. 2016). (No precise numerical breakdown of sensitivity/specificity achieving 90% is given directly, but stated as met.)
	True sensitivity and specificity to be above 80% with 95% confidence, assuming thresholds are determined in a population with at least 70 osteoporotic and non-osteoporotic patients.	This was "realized" in Karjalainen et al. 2016 and "verified" in the U.S. study (Schousboe et al. 2016).
	Similar fracture prediction capability to DXA.	Similar odds ratios (OR) for clinical fractures with DXA (OR = 1.47) and Bindex DI (OR = 1.37), Ct.Th. (OR = 1.56) were reported, suggesting similar fracture prediction capability. For each standard deviation decrease in DI, there was a significant increase in odds for hip osteoporosis (OR=3.03).
	Approximately 30% of patients would need a DXA examination (implying Bindex can reduce unnecessary DXA scans).	This was "verified in both studies."
Primary Safety Endpoint	No adverse events during ultrasound measurement, including skin effects, inflammation, heating, adverse effect on operator, electrical malfunction, or other hazardous situations.	"No adverse events were reported with patient or operator" as documented in the clinical evaluation.
Non-clinical Performance	Biocompatibility (cytotoxicity, sensitization, irritation)	Conducted according to ISO 10993-1. Short skin contact duration (less than 10 minutes) poses very low risk.
	Electrical Safety and Electromagnetic Compatibility (EMC)	Complies with IEC 60601-2-37 and IEC 60601-1-2.
	Software Verification and Validation	Conducted according to FDA guidance; considered "medium" level of concern.
	Mechanical and Acoustic Testing (acoustic output, drop test, ball pressure test, moulding stress relief)	Tested according to IEC 62359 and IEC 60601-1.
	In vitro and In vivo Proof of Concept (accuracy of cortical bone thickness measurement)	High linear correlations (r ≥ 0.95) with reference methods (caliper, pQCT) for cortical thickness. In vivo accuracy of 6.6% and precision of 0.26 mm for the envelope method.

Device Study Details

2. Sample Sizes and Data Provenance

In vitro and In vivo Proof of Concept:
- Bovine cortical bone samples (n=6)
- Human volunteers (n=20)
- Provenance: Not explicitly stated for these initial proof-of-concept studies, but the reference to Karjalainen et al. 2008 suggests academic research.
Clinical Studies:
- Study 1 (Initial Clinical Testing): Sample size not explicitly given for the study that introduced DI (Karjalainen et al. Osteoporos Int 2012).
- Study 2 (Finnish Population): 572 subjects.
  - Provenance: Finland; Karjalainen J. et al. Osteoporos Int 2016. This was a prospective study used to determine diagnostic thresholds.
- Study 3 (U.S. Study): 560 subjects.
  - Provenance: United States; Schousboe et al. 2016. This study validated the thresholds developed in the Finnish study. It's likely prospective given the validation nature.

3. Number of Experts and Qualifications for Ground Truth

The document does not explicitly state the number or specific qualifications of experts used to establish ground truth for the test set. However, it heavily references DXA (Dual-energy X-ray Absorptiometry), which is the established gold standard for osteoporosis diagnosis. The criteria for osteoporosis classification (T-score threshold of -2.5) are based on World Health Organization (WHO) definitions and guidelines from the International Society of Clinical Densitometry (ISCD). This implies that ground truth was established by clinical diagnosis using DXA, interpreted by qualified medical professionals (e.g., radiologists, endocrinologists, or other specialists involved in bone health), rather than by individual experts marking images for AI.

4. Adjudication Method for the Test Set

Not applicable. The primary ground truth for clinical studies was established by DXA measurements and WHO/ISCD criteria for osteoporosis, not by human expert agreement on device output in a read-out session that would require adjudication.

5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study

No, a traditional MRMC study as understood in image interpretation (comparing human reader performance with/without AI assistance) was not performed. This device is a measurement tool (bone sonometer) that provides a quantitative output (Density Index) rather than creating images for human interpretation. Its effectiveness was evaluated by comparing its diagnostic capability (identifying osteoporotic patients, fracture risk prediction) to the DXA gold standard and to the predicate device.

The study aimed to show that the Bindex could reduce the need for DXA examinations by identifying a subset of patients who wouldn't need one (where Bindex's "yellow area" means they would). Approximately 30% of patients would still need a DXA. This indicates an effect on follow-up testing, but not a direct improvement in human reader performance on a given test.

6. Standalone (Algorithm Only Without Human-in-the-Loop Performance) Study

Yes, the primary effectiveness endpoint evaluation describes the performance of the Bindex device (algorithm) itself in making classifications related to osteoporosis, in conjunction with DXA. While the phrase "in conjunction with other clinical risk factors or patient characteristics as an aid to the physician" implies physician involvement, the performance claims ("identify 90% of osteoporotic and non-osteoporotic subjects correctly," "similar fracture prediction capability") are attributed to the device's DI and Ct.Th. values, reflecting its standalone diagnostic potential within a clinical framework. The device is a "bone sonometer" providing a numerical output, not an AI interpreting complex images for human assistance.

7. Type of Ground Truth Used

Clinical Studies: Clinical diagnosis of osteoporosis based on DXA measurements and World Health Organization (WHO) T-score criteria, as defined by the International Society of Clinical Densitometry (ISCD). This is the gold standard for osteoporosis diagnosis.
In vitro and In vivo Proof of Concept:
- In vitro: Caliper measurements of bovine bone samples.
- In vivo: Peripheral Quantitative Computed Tomography (pQCT) measurements of human volunteers.

8. Sample Size for the Training Set

The document does not explicitly state the sample size of a "training set" for the device's algorithm. It describes clinical studies where thresholds were determined and validated.

The Finnish study (n=572) was where the thresholds were "determined." This population effectively served as the development/training dataset for establishing the Density Index (DI) thresholds for osteoporosis classification.
The U.S. study (n=560) was then used for validation of these predetermined thresholds.

9. How the Ground Truth for the Training Set Was Established

For the "training" (threshold determination) set in the Finnish study (n=572), the ground truth for osteoporosis was established by conventional DXA measurements and the application of WHO/ISCD guidelines for osteoporosis diagnosis (T-score threshold of -2.5). This means patients in this cohort underwent DXA scans, and their osteoporosis status was determined based on the globally accepted clinical standard. The Bindex's DI and Ct.Th. values were then correlated with this established DXA-based ground truth to define clinical thresholds for the device.

Ask a Question

Ask a specific question about this device

Page 1 of 1