(123 days)
The Spectra Optia Apheresis System, a blood component separator, is intended for use in therapeutic apheresis applications, and may be used to perform Red Blood Cell Exchange, Depletion, and Depletion/Exchange (RBCX) procedures.
The Spectra Optia Apheresis System, a blood component separator, can be used to perform Red Blood Cell Exchange (RBCx) procedures for the transfusion management of Sickle Cell Disease in adults and children.
The Spectra Optia Apheresis System is a centrifugal system that separates whole blood into its cellular and plasma components. The device is comprised of three major sub-systems: (1) the apheresis machine itself (centrifuge, pumps, valves. etc.). (2) sterile, single-use, disposable tubing sets and, (3) embedded software.
Modifications to the disposable Exchange Set and embedded software have been made to enable Red Blood Cell Exchange (RBCx) procedures on the Spectra Optia system.
Here's an analysis of the acceptance criteria and the study proving the device meets them, based on the provided text:
Spectra Optia® Apheresis System for Red Blood Cell Exchange (RBCx)
1. Table of Acceptance Criteria and Reported Device Performance
| Acceptance Criteria (Implied) | Reported Device Performance |
|---|---|
| Primary Endpoints: Consistently achieve target HbS levels as prescribed by the physician in the target population. | "The study resulted in all primary endpoints being met..." |
| Safety: No Serious Adverse Events (SAEs) or Unanticipated Adverse Device Effects (UADEs). | "...with no serious adverse events (SAEs) or unanticipated adverse device effects (UADEs) reported." |
| Performance Equivalence (vs. predicate COBE Spectra): Ability to achieve patient hematocrit targets. | "In both a "simulated-use" laboratory validation study and human clinical trial, Spectra Optia's RBCx protocol was found to perform the same as the COBE Spectra RBCx protocol, with respect to the system's ability to achieve patient hematocrit targets..." |
| Performance Equivalence (vs. predicate COBE Spectra): Ability to maintain patient fluid balance. | "...and to maintain patient fluid balance." |
2. Sample Size Used for the Test Set and Data Provenance
- Sample Size: Not explicitly stated. The document mentions "a prospective, multi-center, single-arm, open-label study," which indicates a clinical study with real patients, but the exact number of patients is not provided.
- Data Provenance: The study was a "prospective, multi-center" clinical study. This implies data was collected from multiple clinical sites (likely within the US, given the FDA submission) actively as the study progressed (prospective).
3. Number of Experts Used to Establish the Ground Truth for the Test Set and Qualifications of Those Experts
- The concept of "experts establishing ground truth" as seen in diagnostic imaging for example, is not directly applicable here. The device is an apheresis system for treatment, not a diagnostic device.
- The "ground truth" for the primary endpoint (achieving target HbS) would be derived from objective lab measurements of the patients' HbS levels before and after the RBCx procedure, based on the physician's prescription. The "ground truth" for safety would be identified by clinical observation and reporting of adverse events by the clinical staff involved in the study.
4. Adjudication Method for the Test Set
- Adjudication methods (like 2+1, 3+1) are typically used in studies where there's subjectivity in interpreting results (e.g., image reading). This device's primary endpoints (HbS levels, fluid balance, adverse events) are objective measurements or clinical observations. Therefore, a specific "adjudication method" in that sense is not mentioned or likely applicable. The study's design (multi-center, open-label) implies standardized protocols for data collection and reporting.
5. If a Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study Was Done, What was the Effect Size of How Much Human Readers Improve with AI vs Without AI Assistance
- No, an MRMC comparative effectiveness study was not done. This device is an automated apheresis system and does not involve "human readers" or "AI assistance" in the typical sense of a diagnostic or interpretive task. The comparison was between the modified Spectra Optia system and its predicate device (COBE Spectra), both being automated systems.
6. If a Standalone (i.e., algorithm only without human-in-the-loop performance) Was Done
- The device itself is an "algorithm only" in the sense that it is an automated system with embedded software. The clinical study evaluated the performance of this automated system in a real-world clinical setting without direct human intervention in the core RBCx procedure execution by the device itself (though human operators initiate and monitor the procedure).
- The comparison against the predicate COBE Spectra system (also an automated system) effectively acts as a standalone performance comparison between two automated systems.
7. The Type of Ground Truth Used
- Clinical Outcomes/Measurements: The ground truth was based on objective clinical measurements and outcomes.
- HbS Levels: Objective laboratory measurements of hemoglobin S in the patients' blood, compared against physician-prescribed targets.
- Safety Data: Clinical observation and reporting of adverse events (SAEs, UADEs).
- Hematocrit Targets & Fluid Balance: Objective physiological measurements taken during the procedures.
8. The Sample Size for the Training Set
- This device is not a machine learning or AI model in the common sense that requires a "training set" for model development. The software algorithms are likely rule-based or control-system based.
- The software was "verified through a variety of verification testing; including Functional, Reliability, Usability, Exploratory, and Robustness." This indicates traditional software engineering testing rather than machine learning training. Therefore, a "training set" in the context of data for model learning is not applicable or stated.
9. How the Ground Truth for the Training Set Was Established
- As concluded in point 8, there isn't a "training set" in the typical machine learning sense for this device. The software was developed using standard engineering practices, and its performance was then validated through non-clinical verification and a prospective clinical study.
N/A