(212 days)
BD BACTEC Myco/F Lytic culture medium when used with the BD BACTEC fluorescent series instruments is a nonselective culture medium to be used as an adjunct to aerobic blood culture media for the qualitative culture and recovery of mycobacteria, yeast and fungi from blood. This medium may also be used for the culture of sterile body fluids when veast or fungi are suspected.
BD BACTEC Myco/F Lytic culture medium is a Middlebrook 7H9 and Brain Heart Infusion broth formulation for the recovery of mycobacteria from blood specimens, and yeast and fungi from blood and sterile body fluids. Specific modifications were made to enhance the growth and recovery of mycobacteria, yeast and fungi. These modifications include ferric ammonium citrate to provide an iron source for specific strains of mycobacteria and fungi, the addition of saponin as a blood lysing agent, and the addition of specific proteins and sugars to provide nutritional supplements. Each vial contains a sensor which can detect decreases in oxygen concentration in the vial resulting from microorganism metabolism and growth. The sensor is monitored by the BD BACTEC fluorescent series instrument for increasing fluorescence, which is due to the decrease in oxygen. A positive determination indicates the presumptive presence of viable microorganisms in the vial.
BD BACTEC Myco/F Lytic Culture Vials are supplied in a carton containing 50 vials. It is a nonsterile product.
This document is a 510(k) Substantial Equivalence Determination Decision Summary for the BD BACTEC Myco/F Lytic Culture Vials (plastic). It evaluates the device's performance against a predicate device (BD BACTEC Myco/F Lytic Culture Vials (glass)) to determine if it is substantially equivalent. The studies described are analytical performance comparisons rather than clinical trials with human subjects for diagnostic accuracy.
Here's an analysis of the acceptance criteria and study information provided:
1. Table of Acceptance Criteria and Reported Device Performance
The document does not explicitly state "acceptance criteria" in a bulleted or numbered list with corresponding reported performance values within a formal table. Instead, it presents performance data for the modified (plastic) device compared to the predicate (glass) device across various analytical parameters. The implicit acceptance criteria appear to be substantial equivalence, meaning the performance of the new plastic vial is "equivalent to or better than" the predicate glass vial.
Here's a synthesized table based on the provided data, comparing the plastic device's performance to the predicate:
Table: Summary of Device Performance Compared to Predicate (Implicit Acceptance Criteria: Equivalent to or Better)
| Performance Parameter | Acceptance Criteria (Implied) | Reported Device Performance (Plastic vs. Glass) |
|---|---|---|
| Reproducibility | No statistically significant differences in TTD or % recovery. | Met: No statistically significant differences in organism detection times or percent recovery between three lots for blood and sterile body fluid volumes. |
| Microbial Detection Limit (MDL) - Blood | Equivalent or better performance (ratio of positive yields close to 1 or higher) at low inoculum levels. | Met (mostly): Ratio of positive yields close to 1 or higher for most organisms, except Blastomyces dermatitidis (0.33). This organism's inconsistent growth was noted as a limitation. |
| Microbial Detection Limit (MDL) - Sterile Body Fluid | Equivalent or better performance (ratio of positive yields close to 1 or higher) at low inoculum levels. | Met (mostly): Ratio of positive yields close to 1 or higher for most organisms, except Blastomyces dermatitidis (0.86). This organism's inconsistent growth was noted as a limitation. |
| Delayed Entry in blood (Mycobacteria) | 100% recovery showed in delayed conditions. | Met: 100% recovery for all 8 mycobacterium strains evaluated across various delay times (12-96 hours) and temperatures (20-37.5°C). The Product Insert still recommends prompt placement but provides this information. |
| Time to Detection (TTD) - Blood | Equivalent or better performance (95% CI for median TTD difference contains zero or only negative values for shorter TTD). | Met (mostly): The plastic device performed equivalently to or better than the predicate for most conditions. Exceptions noted were for Dimorphic Fungi (0-1 CFU/vial, 1 mL blood) and Mycobacteria (1-10 CFU/vial, 1mL blood) where median TTD for plastic was longer, but supporting statistical analysis found performance equivalent or better for other conditions. |
| Time to Detection (TTD) - Sterile Body Fluid (SBF) | Equivalent or better performance (95% CI for median TTD difference contains zero or only negative values for shorter TTD). | Met (mostly): The plastic device performed equivalently to or better than the predicate for most conditions. Exceptions noted for Dimorphic Fungi under some test conditions. |
| Percent Recovery (Detection) - Blood (10-100 CFU/vial) | Comparable recovery rates (ratio of positive yields close to 1). | Met: Plastic recovery 99.52% vs. Glass 100%. Ratio of positive yields was 1.00. One instance of negative result only in plastic vial (Blastomyces dermatitidis), noted as a limitation. |
| Percent Recovery (Detection) - SBF (10-100 CFU/vial) | Comparable recovery rates (ratio of positive yields close to 1). | Met: Plastic recovery 100% vs. Glass 99.24%. Ratio of positive yields was 1.01. One instance of negative result only in glass vial (Candida auris). |
| False Positive Rate - Blood | Lower or comparable to predicate. | Met (superior): Plastic: 0.70% (all volumes combined) vs. Glass: 25.87%. Majority of false positives in glass were at higher blood volumes. |
| False Positive Rate - SBF | Lower or comparable to predicate. | Met (lower): Plastic: 0.47% (all SBF types combined) vs. Glass: 0.71%. |
| False Negative Rate - Blood | Comparable rates (not statistically significant difference). | Met: Plastic: 0.65% (4/619) vs. Glass: 0.48% (3/619). Difference not statistically significant (P = 1). |
| False Negative Rate - SBF | Comparable rates (not statistically significant difference). | Met: Plastic: 1.52% (6/396) vs. Glass: 1.01% (4/396). Difference not statistically significant (P = 0.7523). |
| Instrument Compatibility - Blood | Equivalent performance (similar TTD, comparable recovery). | Met: All four instruments (FX, FX-40, 9240, 9050) showed 100% recovery for both plastic and glass vials across organisms and volumes (with one exception for plastic at 5mL on FX-40 (94.44% vs 100% for glass)). TTD results were also comparable with median differences near zero. |
| Instrument Compatibility - SBF | Equivalent performance (similar TTD, comparable recovery). | Met: All four instruments showed 100% recovery for both plastic and glass vials across organisms and SBF volumes. TTD results were also comparable with median differences near zero. |
2. Sample Size Used for the Test Set and Data Provenance
The studies described are analytical performance studies performed in a laboratory setting, rather than clinical studies with patient samples. The samples used are seeded samples (i.e., known microorganisms inoculated into blood or sterile body fluid).
- Test Set (Seeded Samples):
- Reproducibility: Not explicitly stated as a separate "test set" size, but involved testing across three lots of media with varying blood/SBF volumes and inoculum levels.
- Microbial Detection Limit (MDL) - Blood:
- Sample Size: 414 paired vials (plastic vs. glass). 2 pairs discarded due to contamination, resulting in 412 paired vials. (23 organisms x 3 lots x 3 blood vol x 2 inoculum levels x 1 instrument).
- Data Provenance: This is in vitro analytical data obtained from controlled laboratory experiments, not from human patients or a specific country of origin.
- Microbial Detection Limit (MDL) - Sterile Body Fluid (SBF):
- Sample Size: 264 paired vials (plastic vs. glass). (3 organisms x 3 lots x 4 SBF volumes x 2 inoculum levels x 1 instrument x 2 SBF types) + (5 organisms x 3 lots x 4 SBF volumes x 2 inoculum levels x 1 instrument x 1 SBF type).
- Data Provenance: In vitro analytical data.
- Delayed Entry in blood (Mycobacteria):
- Sample Size: Not explicitly stated as a total paired set number, but involved 8 strains of mycobacterium. Total number of vials for each delay condition (e.g., 72/72 for no delay, 70/70 for 12h @ 20-25C, etc.) is provided in Table 7. Each test used 3 lots and 1, 3, or 5 mL human blood.
- Data Provenance: In vitro analytical data using human blood matrix.
- Percent Recovery (Detection) - Blood (10-100 CFU/vial):
- Sample Size: 207 paired sets (plastic vs. glass). (23 organisms x 3 lots x 3 blood volumes x 1 instrument).
- Data Provenance: In vitro analytical data using blood matrix.
- Percent Recovery (Detection) - SBF (10-100 CFU/vial):
- Sample Size: 132 paired sets (plastic vs. glass). (3 organisms x 3 lots x 4 SBF volumes x 1 inoculum level x 1 instrument x 2 SBF types) + (5 organisms x 3 lots x 4 SBF volumes x 1 inoculum level x 1 instrument x 1 SBF type).
- Data Provenance: In vitro analytical data using SBF matrix.
- False Positive Rate - Blood:
- Sample Size: 144 paired sets initially, 143 after contamination exclusion. (8 vials x 3 blood volumes x 3 lots x 2 instruments).
- Data Provenance: In vitro analytical data using fresh blood.
- False Positive Rate - SBF:
- Sample Size: 432 paired sets initially, 425/424 after contamination exclusions. (8 vials x 3 SBF volumes x 3 lots x 2 instruments x 3 SBF types).
- Data Provenance: In vitro analytical data using sterile body fluid.
- False Negative Rate - Blood:
- Sample Size: 619 combined paired sets from Recovery and MDL studies.
- Data Provenance: In vitro analytical data.
- False Negative Rate - SBF:
- Sample Size: 396 combined paired sets from Recovery and MDL studies.
- Data Provenance: In vitro analytical data.
- Instrument Compatibility - Blood:
- Sample Size: 54 paired sets for FX and 9050 each, 53 paired sets for FX-40 and 9240 each (total 214-216 paired sets for blood). (6 organisms x 3 blood volumes x 1 inoculum x 3 lots / 4 instruments).
- Data Provenance: In vitro analytical data.
- Instrument Compatibility - SBF:
- Sample Size: 48 paired sets per instrument (total 192 paired sets for SBF). (4 organisms x 4 SBF volumes x 1 inoculum x 3 lots / 4 instruments).
- Data Provenance: In vitro analytical data.
All data described is retrospective in the sense that it was collected as part of a pre-market submission process, likely after the plastic vial formulation was developed. It is not prospective clinical data from patient studies. The data provenance is internal laboratory studies, not patient data from a specific country.
3. Number of Experts Used to Establish the Ground Truth for the Test Set and Qualifications of Those Experts
The ground truth for these analytical studies is based on controlled, seeded experiments where the presence and identity of the microorganisms are known.
- The "ground truth" is the known inoculum (CFU/vial) of specific ATCC strains or clinical isolates of mycobacteria, yeast, and fungi.
- Detection by instruments (BD BACTEC fluorescent series instruments) is validated against the known presence/absence of these seeded organisms and further confirmed by subculture for discordant results (e.g., false positives/negatives), which is a standard microbiological method.
- No human experts (like radiologists) are mentioned or typically involved in establishing ground truth for in vitro diagnostic (IVD) culture media performance studies. The ground truth is intrinsically tied to the experimental design – what was precisely inoculated and subsequently confirmed by laboratory methods.
4. Adjudication Method for the Test Set
Adjudication, in the context of IVD performance, would typically involve resolving discrepancies between a device's result and a reference method/truth.
- For these studies, the primary comparison is the Time-to-Detection (TTD) and Percent Recovery of the plastic vial versus the glass predicate vial against the known seeded inoculum.
- For false positive/negative rates, discrepancies between instrument reads and definitive culture (subculture) results are identified. The report explicitly mentions how these were defined:
- False Positive: "instrument positive but subculture negative."
- False Negative: "instrument-negative at the end of protocol yet contains viable organisms upon subculturing onto appropriate culture media."
- There's no mention of a multi-observer or consensus-based adjudication process as one would see in image interpretation studies. The "adjudication" is inherent in the design of comparing instrument readouts to known inputs and standard microbiological subculture confirmation.
5. If a Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study Was Done
No, a Multi-Reader Multi-Case (MRMC) comparative effectiveness study was not done.
- MRMC studies are typically performed for medical imaging devices where human readers interpret images, and the AI's impact on their performance is evaluated.
- This submission concerns an in vitro diagnostic culture medium, which is an automated system for microbial growth detection. The "reader" is the BD BACTEC fluorescent series instrument, not a human.
- Therefore, the concept of "human readers improve with AI vs without AI assistance" is not applicable here. The comparison is between two different types of culture vials (plastic vs. glass), both processed by the same automated instruments.
6. If a Standalone (i.e., algorithm only without human-in-the-loop performance) Was Done
Yes, in essence, standalone performance was evaluated.
- The device being assessed is the BD BACTEC Myco/F Lytic Culture Vial (plastic), which operates when placed into the BD BACTEC fluorescent series instruments.
- The performance metrics (TTD, Percent Recovery, False Positive/Negative rates) are generated by the instrument's automated detection algorithm based on the changes in fluorescence caused by microbial growth in the vial.
- Humans are involved in inoculating the vials and performing subcultures for confirmation, but the "performance" described is the direct, automated output of the integrated vial-instrument system. It is not an "AI" in the sense of an image interpretation algorithm, but an automated detection system where the organism's metabolism directly causes a measurable signal change detected by the instrument's programming (analogous to an algorithm).
7. The Type of Ground Truth Used
The ground truth used primarily in these analytical studies is an expert-determined, controlled, seeded panel.
- This involves known strains of mycobacteria, yeast, and fungi (e.g., ATCC strains and clinical isolates) at specified inoculum levels (CFU/vial).
- For discordant results (e.g., instrument negative but suspected positive), subculture onto appropriate culture media plates was used to confirm the presence of viable organisms, acting as a definitive microbiological reference method for the seeded content.
- It does not utilize pathology reports (which are for tissue diagnosis) or outcomes data (which would be clinical outcomes in patients).
8. The Sample Size for the Training Set
The document does not explicitly mention a "training set" or "training data" in the conventional machine learning sense.
- This is an IVD device validation, not an AI/ML software validation.
- The comparison is between a new device (plastic vial) and a predicate device (glass vial), demonstrating substantial equivalence based on a series of analytical performance studies.
- The "training" of the instrument's underlying detection algorithm (which interprets the fluorescent signals) would have occurred during its initial development and prior regulatory submissions (e.g., for K970333, K970512 for the predicate device). The current submission focuses on demonstrating that the change in vial material (plastic vs. glass) does not adversely affect this established performance.
9. How the Ground Truth for the Training Set Was Established
As noted above, there is no explicitly defined "training set" for the vial in the context of this submission. The ground truth for the instrument's core detection capabilities would have been established historically during its development. This would also involve known, quantifiable microbial cultures and their characteristic growth patterns and metabolic activity in the media over time, likely confirmed by standard agar plate cultures or other gold standard microbiological methods.
§ 866.2560 Microbial growth monitor.
(a)
Identification. A microbial growth monitor is a device intended for medical purposes that measures the concentration of bacteria suspended in a liquid medium by measuring changes in light scattering properties, optical density, electrical impedance, or by making direct bacterial counts. The device aids in the diagnosis of disease caused by pathogenic microorganisms.(b)
Classification. Class I. With the exception of automated blood culturing system devices that are used in testing for bacteria, fungi, and other microorganisms in blood and other normally sterile body fluids, this device is exempt from the premarket notification procedures in subpart E of part 807 of this chapter.