Search Results

The XR-Series module (XR-10) is a quantitative multi-parameter automated hematology analyzer intended for in vitro diagnostic use in screening patient populations found in clinical laboratories.

The XR-Series module classifies and enumerates the following parameters in whole blood: WBC, RBC, HGB, HCT, MCV, MCH, MCHC, PLT (PLT-I, PLT-F), NEUT%/#, LYMPH%/#, MONO%/#, EO%/#, BASO%/#, IG%/#, RDW-CV, RDW-SD, MPV, NRBC%/#, RET%/#, IPF, IPF#, IRF, RET-He and has a Body Fluid mode for body fluids. The Body Fluid mode enumerates the WBC-BF, RBC-BF, MN%/#, PMN%/#, and TC-BF# parameters in cerebrospinal fluid (CSF), serous fluids (peritoneal, pleural) and synovial fluids. Whole blood should be collected in K2EDTA or K3EDTA anticoagulant, and serous and synovial fluids in K2EDTA anticoagulant to prevent clotting of fluid. The use of anticoagulants with CSF specimens is neither required nor recommended.

The Sysmex XR-Series module (XR-10) is a quantitative multi-parameter hematology analyzer intended to perform tests on whole blood samples collected in K2 or K3EDTA and body fluids (pleural, peritoneal and synovial) collected in K2EDTA anticoagulant. The analyzers can also perform tests on CSF, which should not be collected in any anticoagulant. The XR-Series analyzer consist of four principal units: (1) One Main Units (XR-10) which aspirates, dilutes, mixes, and analyzes blood and body fluid samples; (2) Two Auto Sampler Units (SA-10, SA-01) which supply samples to the Main Unit automatically; (3) IPU (Information Processing Unit) which processes data from the Main Unit and provides the operator interface with the system; (4) Pneumatic Unit which supplies pressure and vacuum from the Main Unit.

This document describes the acceptance criteria and the studies conducted to prove that the Sysmex XR-Series (XR-10) Automated Hematology Analyzer meets these criteria, demonstrating substantial equivalence to its predicate device, the Sysmex XN-20.

1. Table of Acceptance Criteria and Reported Device Performance

The FDA 510(k) clearance letter does not explicitly present a neatly formatted table of acceptance criteria alongside the reported performance for all parameters. Instead, it describes various performance studies (Precision, Linearity, Analytical Specificity/Interferences, Sample Stability, Detection Limit, Carry-Over, Comparison Studies, Matrix Studies, Bridging Studies, Clinical Studies, and Expected Values/Reference Range) and states that the XR-10 met "manufacturer's specifications or predefined acceptance criteria requirements" for each.

However, based on the provided data, we can infer some general acceptance criteria, particularly from the method comparison study which uses correlation coefficient (r) and percent bias (%Bias) as metrics. The clinical sensitivity and specificity tables also present implied acceptance criteria based on the demonstrated performance.

Inferred Acceptance Criteria & Reported Performance (Selection of Key Metrics)

2. Sample Size Used for the Test Set and Data Provenance

• Test Set Sample Size & Provenance:
  • Precision (Repeatability - Whole Blood): Residual K2EDTA whole blood samples for 10 replicates for target values, and three samples for other parameters. This was across three US clinical sites (Site 01, 05, 24).
  • Precision (Reproducibility - Whole Blood): XN CHECK whole blood control material, 90 results per control level (3 levels x 3 replicates x 2 runs x 5 days). Conducted at three US clinical sites.
  • Precision (Body Fluid): Residual peritoneal, pleural, and synovial fluid samples (K2EDTA) and CSF (no anticoagulant) for 10 replicates for target values. Conducted at three US clinical sites.
  • Linearity (Whole Blood & Body Fluid): Minimum of seven sample dilutions. Performed at one internal site.
  • Analytical Specificity/Interferences: Whole blood K2EDTA samples from donors. Number of samples not specified, but collected for this study purpose.
  • Sample Stability (Whole Blood): 8 unique leftover samples and 12 prospectively collected K2EDTA venous whole blood samples (20 samples total). Conducted at one internal site.
  • Sample Stability (Body Fluid): 12 unique de-identified leftover body fluid samples (3-CSF, 3-peritoneal, 3-pleural, 3-synovial). Conducted at 1 external site.
  • Detection Limit: Four blank samples and four low concentration samples for each parameter. Conducted across 2 XR-10 analyzers (implied internal or multi-site for the overall study context).
  • Carry-Over: High and low target concentration samples (number not specified). Conducted at three US clinical sites.
  • Comparison Studies (Whole Blood): 865 unique residual whole blood samples from pediatrics (<21 years) and adults (≥21 years), including various disease states. Collected across 3 US clinical sites. Samples were "residual" (retrospective).
  • Comparison Studies (Body Fluid): 397 residual body fluid samples (CSF, peritoneal, pleural, and synovial). Collected across three US sites. Samples were "residual" (retrospective).
  • Matrix Studies (Anticoagulant Comparison): 46 paired K2 and K3EDTA venous whole blood samples. Conducted at 1 internal site.
  • Matrix Studies (Venous vs. Capillary): 70 paired venous whole blood samples. Conducted at one internal site.
  • Matrix Studies (Normal vs. Micro-collection Tubes): 70 paired venous whole blood samples. Conducted at one internal site.
  • Bridging Studies (Whole Blood Mode to Pre-dilute Mode): 45 de-identified residual whole blood samples. Conducted at 1 internal site.
  • Bridging Studies (Predilute Mode Normal Tube to Micro-collection Tube): 40 de-identified residual whole blood samples. Conducted at one internal site.
  • Bridging Studies (Low WBC Mode Normal Tube to Micro collection Tube): 43 residual de-identified venous whole blood samples. Conducted at one internal site.
  • Clinical Studies (Sensitivity and Specificity): Patient samples representing a variety of abnormal conditions. 705 to 780 samples for manual microscopy comparison; 834 to 845 samples for XN-20 predicate comparison. Conducted at three external clinical sites (same as method comparison). This data appears retrospective as it involves "patient samples" and "manual differential counts and peripheral blood smear review".
  • Expected Values/Reference Range (Adult): 132 samples (58 males, 74 females). Data provenance implied from "verification of adult reference intervals" which suggests comparison against pre-existing intervals for a predicate device.
  • Expected Values/Reference Range (Pediatric): 196 pediatric samples (40 neonates, 55 infants, 60 children, 41 adolescents). Data provenance implied from "verification study" using established literature.

3. Number of Experts Used to Establish the Ground Truth for the Test Set and Their Qualifications

• For the Clinical Sensitivity and Specificity study (Abnormal Flagging – Manual Microscopy), ground truth was established by "manual differential counts and peripheral blood smear review by experienced examiners using light microscopy (reference method) at each of the three external clinical sites".
• The exact number of experts and their specific qualifications (e.g., "radiologist with 10 years of experience") are not specified in the provided text. It merely states "experienced examiners."

4. Adjudication Method for the Test Set

• For the Clinical Sensitivity and Specificity study, an adjudication method for discrepancies between examiners or between the device and manual microscopy is not explicitly described. It states that the manual differential counts and peripheral blood smear review were the "reference method," implying these were the accepted ground truth. No mention of 2+1, 3+1, or any other formal adjudication process is present.

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

• No, a traditional MRMC comparative effectiveness study where human readers' performance with and without AI assistance is measured and compared was not performed, nor is it applicable to this type of device (an automated hematology analyzer). This device classifies and enumerates blood parameters automatically, without human "reading" or AI assistance in the human interpretation loop. The "clinical studies" described evaluate the device's flagging capabilities against manual microscopy (human expert consensus) and against a predicate automated device, not the improvement of human readers using the device's AI capabilities.

6. If a Standalone (i.e. algorithm only without human-in-the-loop performance) was done

• Yes, the entire performance evaluation of the Sysmex XR-Series (XR-10) is essentially a standalone (algorithm only) performance assessment. It is an automated hematology analyzer that quantifies and classifies blood parameters without direct human intervention or interpretive assistance during the measurement process. The "Comparison Studies" directly assess the XR-10's performance against a predicate automated device (XN-20), representing a standalone comparison. The "Clinical Sensitivity and Specificity" also evaluates the XR-10's automated flagging against manual microscopy (human expert consensus of ground truth slides), demonstrating its standalone diagnostic performance.

7. The Type of Ground Truth Used

• The ground truth varied depending on the type of study:
  • For Analytical Performance (Precision, Linearity, Detection Limit, Carry-Over): The ground truth was based on the performance of the reference methods, control materials, calibrators, and linearity materials, as well as manufacturer's specifications and CLSI guidelines. For LoD/LoQ, the Sysmex XN-20 (the predicate device) was used to assign reference values to low-level samples.
  • For Method Comparison Studies: The ground truth was implicitly the predicate device (Sysmex XN-20). The study aimed to show substantial equivalence between the new device and the predicate, not necessarily against a "true" biological gold standard across all parameters.
  • For Clinical Sensitivity and Specificity: The ground truth for abnormal flagging was established by manual differential counts and peripheral blood smear review by experienced examiners using light microscopy. This represents an expert consensus (from potentially multiple experts, though not detailed) based on visual examination.
  • For Reference Interval Verification: The ground truth was established by literature sources (Wong et al., 2021 for pediatric; Kjeldsberg's Body Fluids, 1993 for body fluids) or pre-established reference intervals for a predicate device (Sysmex XE-5000 for adult whole blood).

8. The Sample Size for the Training Set

• The FDA 510(k) clearance letter does not specify the sample size used for the training set for the Sysmex XR-Series (XR-10) Automated Hematology Analyzer. It focuses solely on the validation/test performance data, which is typical for 510(k) submissions where the device's performance is demonstrated against established methods or predicate devices. This indicates that information on the internal development and training of the device's algorithms is not part of the publicly available clearance letter.

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

• Since the training set size is not provided, the method for establishing its ground truth is also not described in this document. It is generally assumed that manufacturers use robust internal processes for data collection, annotation, and ground truth establishment during the development phase of such automated diagnostic systems (e.g., using a combination of expert review, reference methods, and potentially pathological or clinical outcomes data, depending on the specific parameter and its clinical relevance). However, this specific information is not included in the 510(k) summary.

Ask a specific question about this device