(349 days)
The HD01-series monitors are intended for use by trained medical personnel on patients receiving hemodialysis treatment, for the routine monitoring of the following patient and blood handling diagnostic parameters:
Access recirculation: a check on proper dialysis needle placement and dialysis pump setting in relation to flow in the access device. Access recirculation would prevent the patient of receiving its full hemodialysis prescription. Access recirculation that cannot be mitigated by altering needle placement should be referred to the attending physician, for a change in the patient hemodialysis prescription and as a possible indication of deteriorating flow in the access device.
Access flow: allows the routine monitoring of patency of the access device. A substantial decrease in flow passing through the shunt or fistula from the initial level may indicate deterioration of the access, and should be referred to the attending physician for further diagnostic assessment and treatment.
Delivered bloodflow: a check on the hemodialysis pump setting, to confirm that the patient receives the hemodialysis treatment prescribed by the attending physician.
The HD01 uses transit-time ultrasound principles to make the flow measurements and register sound velocity indicator dilution curves. Standard Stewart-Hamilton equations are employed for the various calculations. Portions of these technologies are covered under Transonic Systems' USA and worldwide patents and patents pending.
All Transonic HD01-series hemodialysis monitor devices consists of the following components:
- "HD01 flow/dilution meter" or "HD01 meter": A bench-top, line power operated electronic measurement unit with serial RS232 data output link;
- "Dual flow/dilution sensor", individually referred to as "arterial sensor" and "venous sensor": Two plastic-encased ultrasonic sensors, connecting to the HD01 unit, to be clipped onto the patient arterial and venous hemodialysis blood lines;
- "Monitor software" also referred to as "software": Computer software to be installed on an IBM-PC compatible computer, which performs the various device function calculations (R%, QA etc.) from the signals produced by the HD01 meter.
Here's an analysis of the acceptance criteria and supporting studies for the Transonic HD01-SERIES HEMODIALYSIS MONITOR, based on the provided text:
1. A table of acceptance criteria and the reported device performance
| Device Function | Acceptance Criteria (Predicate Device) | Reported Device Performance (Transonic HD01) |
|---|---|---|
| Access Recirculation | 3-Sample BUN recirculation on B. M. Hitachi 911 Analyzer | Transonic HD01-R% |
| Absolute Accuracy | Clinical zero recirculation offset: 3 to 12%. Accuracy at higher recirculation levels: not available. | Clinical zero recirc. accuracy: ± 2%. Accuracy at higher recirculation levels: The larger of: ± 2% or ± 5% of reading (bench-validated). |
| Repeatability | Clinical data: 2% ± 14.4%, 20% ± 9.6%, 40% ± 3.6%. Clinical correlation coefficient .81. | Absolute, bench: 2% ± 2%, 20% ± 2%, 40% ± 2%. Clinical correlation coefficient .98. |
| Access Flow | Color Doppler Duplex Sonography Philips P700 w. CVI-Q option | Transonic HD01-QA |
| Accuracy | ± 14%, subject to tight alignment and operator training conditions. | ± 50 ml/min or ± 15% of access flow, whichever is larger. |
| Repeatability | Clinical: correlation coefficient = 0.52. | Clinical: correlation coefficient = 0.978, mean absolute error = 5.0 ± 3.8%. |
| Delivered Bloodflow | Fresenius 2008 Dialysis System (pump performance) | Transonic HD01-QB |
| Total Accuracy | Pump calibration by RPM to flow. No accuracy tolerance specified for changes in tubing ID and inlet pressure. Literature cites up to 50% error. | $\pm 6%$ of flow reading + zero offset; Max. zero offset = $\pm 8$ ml/min (adjustable to zero). Total combined error: $\pm 8%$ for a typical flow in 400 ml/min range. |
| Delivered Bloodflow (Comparison to Transonic T101D/T201D) | Transonic T101D/T201D with 8C clamp-on flow probe | Transonic HD01-QB with H4D clamp-on sensor |
| Total Accuracy | $\pm 5%$ of flow reading + zero offset; Max. zero offset = +50 ml/min. | $\pm 6%$ of flow reading + zero offset; Max. zero offset = +8 ml/min. |
| Max. relative error | $\pm 2%$ of flow reading + zero offset. | $\pm 2%$ of flow reading + zero offset. |
| Typical flow resolution | 5 ml/min. | 1 ml/min. |
| Flow range | -10 to 10 L/min. | -2 to +2 L/min. |
2. Sample size used for the test set and the data provenance
The document does not explicitly state a specific "test set" sample size or its provenance (country of origin, retrospective/prospective). Instead, it references numerous publications (14 distinct publications cited) for each device function (Access Recirculation, Access Flow, Delivered Bloodflow) as evidence for effectiveness. These publications would contain the specific study designs, sample sizes, and data provenance. For example, for "Access Recirculation," studies like Depner et al., 1995, and Krivitski, 1995, are cited.
3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts
This information is not provided in the summary. The studies referenced would likely detail how ground truth was established, but the summary document itself does not specify the number or qualifications of experts for defining ground truth in a general test set.
4. Adjudication method for the test set
The document does not describe a specific adjudication method for a test set. This type of information would typically be found within the detailed methodologies of the individual studies cited.
5. If a multi-reader multi-case (MRMC) comparative effectiveness study was done, If so, what was the effect size of how much human readers improve with AI vs without AI assistance
This document describes a medical device for measuring physiological parameters (blood flow, recirculation) and does not involve "human readers" interpreting images or data in a way that would typically necessitate an MRMC comparative effectiveness study with AI assistance. Therefore, no such study is described, and no effect size for AI assistance is provided. The device itself is the measurement tool.
6. If a standalone (i.e. algorithm only without human-in-the-loop performance) was done
The Transonic HD01 is a standalone device in the sense that it performs its measurements (flow, dilution curves, calculations) automatically based on the input from the sensors. The "monitor software" performs the calculations from the signals. While a trained medical operator selects the measurements and performs procedures like saline injection, the core measurement and calculation process for R%, QA, etc., is done by the algorithm without human intervention in the interpretation of raw sensor data into the reported values. The effectiveness studies cited are implicitly standalone algorithm performance studies, validated against established clinical methods.
7. The type of ground truth used
The ground truth used for validation varies depending on the function:
- Access Recirculation:
- Comparison with BUN recirculation measurements using a Boehringer Mannheim Hitachi 911 Analyzer.
- Comparison with Urea Dilution methods.
- Access Flow:
- Comparison with Magnetic Resonance Angiography (MRA).
- Comparison with Duplex Ultrasonography.
- Validation in animal models (sheep).
- Delivered Bloodflow:
- Comparison with hemodialysis pump readouts (Fresenius 2008 Dialysis System).
- Comparison with volumetric (beaker-stopwatch) calibration.
- Comparison with another Transonic flowmeter (T101D/T201D).
8. The sample size for the training set
The document does not explicitly mention a "training set" or its sample size. This summary focuses on the device's validation against predicate devices and published studies. If machine learning or AI models were used in the device development, the training set information would be detailed internally or in specific technical documents, not typically in a 510(k) summary focused on substantial equivalence. The "HD01 flow/dilution meter" and "Monitor software" are described, but the method of their development (e.g., if they involved a "training set" in a modern AI sense) is not elaborated.
9. How the ground truth for the training set was established
As no specific "training set" is mentioned in the context of machine learning, there is no information provided on how its ground truth was established. The device utilizes "transit-time ultrasound principles" and "Standard Stewart-Hamilton equations" for its calculations, which are established physical and mathematical principles, rather than being learned from a labeled training set in the way a modern AI would be.
§ 876.5820 Hemodialysis system and accessories.
(a)
Identification. A hemodialysis system and accessories is a device that is used as an artificial kidney system for the treatment of patients with renal failure or toxemic conditions and that consists of an extracorporeal blood system, a conventional dialyzer, a dialysate delivery system, and accessories. Blood from a patient flows through the tubing of the extracorporeal blood system and accessories to the blood compartment of the dialyzer, then returns through further tubing of the extracorporeal blood system to the patient. The dialyzer has two compartments that are separated by a semipermeable membrane. While the blood is in the blood compartment, undesirable substances in the blood pass through the semipermeable membrane into the dialysate in the dialysate compartment. The dialysate delivery system controls and monitors the dialysate circulating through the dialysate compartment of the dialyzer.(1) The extracorporeal blood system and accessories consists of tubing, pumps, pressure monitors, air foam or bubble detectors, and alarms to keep blood moving safely from the blood access device and accessories for hemodialysis (§ 876.5540) to the blood compartment of the dialyzer and back to the patient.
(2) The conventional dialyzer allows a transfer of water and solutes between the blood and the dialysate through the semipermeable membrane. The semipermeable membrane of the conventional dialyzer has a sufficiently low permeability to water that an ultrafiltration controller is not required to prevent excessive loss of water from the patient's blood. This conventional dialyzer does not include hemodialyzers with the disposable inserts (Kiil type) (§ 876.5830) or dialyzers of high permeability (§ 876.5860).
(3) The dialysate delivery system consists of mechanisms that monitor and control the temperature, conductivity, flow rate, and pressure of the dialysate and circulates dialysate through the dialysate compartment of the dialyzer. The dialysate delivery system includes the dialysate concentrate for hemodialysis (liquid or powder) and alarms to indicate abnormal dialysate conditions. This dialysate delivery system does not include the sorbent regenerated dialysate delivery system for hemodialysis (§ 876.5600), the dialysate delivery system of the peritoneal dialysis system and accessories (§ 876.5630), or the controlled dialysate delivery system of the high permeability hemodialysis system § 876.5860).
(4) Remote accessories to the hemodialysis system include the unpowered dialysis chair without a scale, the powered dialysis chair without a scale, the dialyzer holder set, dialysis tie gun and ties, and hemodialysis start/stop tray.
(b)
Classification. (1) Class II (performance standards) for hemodialysis systems and all accessories directly associated with the extracorporeal blood system and the dialysate delivery system.(2) Class I for other accessories of the hemodialysis system remote from the extracorporeal blood system and the dialysate delivery system, such as the unpowered dialysis chair, hemodialysis start/stop tray, dialyzer holder set, and dialysis tie gun and ties. The devices subject to this paragraph (b)(2) are exempt from the premarket notification procedures in subpart E of part 807 of this chapter subject to the limitations in § 876.9.