Review Reports - Multicenter Analytical Performance Evaluation of the BD Phoenix NMIC-461 Panel for Carbapenemase Classification and Antimicrobial Susceptibility Testing of Enterobacterales, Pseudomonas aeruginosa, and Acinetobacter spp.

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Manuscript Title: Performance evaluation of the BD Phoenix NMIC-461 Panel for Carbapenemase Classification and Antimicrobial Susceptibility Testing of Enterobacterales, Pseudomonas aeruginosa, and Acinetobacter spp.

This multicenter study evaluates the BD Phoenix NMIC-461 panel for carbapenemase detection, classification, and antimicrobial susceptibility testing in Gram-negative pathogens. The work is clinically relevant and timely, with a large isolate set. However, methodological clarity, data presentation, and interpretation particularly regarding “positive unclassified” results require substantial revision.

Title

Consider specifying “multicenter analytical performance evaluation” rather Performance evaluation.

Acinetobacter spp. correct it.

Lines 28, 29, 30: Sensitivity/specificity values are difficult to follow due to long compound sentences.

Lines 36–41: Conclusions are repetitive (“In conclusion” used twice) and mix AST and carbapenemase findings without clear separation.

Keywords: Inconsistent punctuation: “Acinetobacter spp.;” uses a non-English comma., correct it.

Consider adding “BD Phoenix” and “NMIC-461” as keywords.

Introduction: “CRE and non-fermenting drug-resistant bacteria … remains prevalent” should be “remain prevalent.”

Line 49: Correct it throughout your manuscript “Acinetobacter spp.

Lines 53: CHINET data are cited via website only; a formal reference is required.

Lines 63 to 70: Carbapenemase classification could be condensed.

Lines 83 to 84: Novelty relative to existing evaluations in China is not clearly articulated.

Lines 85–87: Study objective would benefit from a clearer hypothesis.

Methods

Major concern: The Methods section is largely missing or insufficiently described in the provided manuscript.

No dedicated Materials and Methods section describing:

Isolate selection criteria

Inclusion/exclusion rules

AST conditions (media, incubation, breakpoints)

Statistical analysis

CLSI references (e.g., M52, breakpoints) are cited later but not introduced systematically.

PCR methods lack primer sequences, cycling conditions, and quality control measures.

This omission represents a major methodological flaw.

Lines 126: Rationale for excluding “positive unclassified” strains

Lines 156: “positive unclassified” strains represent 34.8% of CPOs this is a major limitation but is framed too mildly.

“Enterobacteriales” is used instead of “Enterobacterales” (Lines 129–130).

Table 1: mCIM is treated as gold standard without justification.

Insufficient comparison with prior BD Phoenix CPO panel studies.

Limitations (selection bias, regional strain distribution, exclusion strategy) are inadequately acknowledged.

Author Response

Title

Consider specifying “multicenter analytical performance evaluation” rather Performance evaluation.

Response: Thanks for your suggestion. We have corrected “Performance evaluation” to “Multicenter analytical performance evaluation” in the revised manuscript. (lines 2)

Acinetobacter spp. correct it.

Response: Thanks for your suggestion. I have corrected all instances of “Acinetobacter spp.” to “Acinetobacter spp.”.

Lines 28, 29, 30: Sensitivity/specificity values are difficult to follow due to long compound sentences.

Response: Thanks for your suggestion. I have revised the expression to make it potentially easier to understand. (lines 28-31)

Lines 36–41: Conclusions are repetitive (“In conclusion” used twice) and mix AST and carbapenemase findings without clear separation.

Response: Thanks for your suggestion. I have revised the expression to make it potentially easier to understand. (lines 41-46)

Keywords: Inconsistent punctuation: “Acinetobacter spp.;” uses a non-English comma., correct it.

Response: Thanks for your suggestion. We have revised the punctuation mark. (lines 51)

Consider adding “BD Phoenix” and “NMIC-461” as keywords.

Response: Thanks for your suggestion. We have added these two terms to the keywords. (lines 52)

Introduction: “CRE and non-fermenting drug-resistant bacteria … remains prevalent” should be “remain prevalent.”

Response: Thanks for your suggestion. We have corrected “remains” to “remain”. (lines 57)

Line 49: Correct it throughout your manuscript “Acinetobacter spp.

Response: Thanks for your suggestion. I have corrected all instances of “Acinetobacter spp.” to “Acinetobacter spp.”.

Lines 53: CHINET data are cited via website only; a formal reference is required.

Response: Thanks for your suggestion. The data displayed on the CHINET website is more comprehensive than that in the references; therefore, we chose to cite the website data.

Lines 63 to 70: Carbapenemase classification could be condensed.

Response: Thanks for your suggestion. We have streamlined the wording in this section. (lines 71-79)

Lines 83 to 84: Novelty relative to existing evaluations in China is not clearly articulated.

Response: Thanks for your suggestion. Since the NMIC-461 panel was only introduced to China in 2024, there are currently no corresponding evaluation studies available in China.

Lines 85–87: Study objective would benefit from a clearer hypothesis.

Response: Thanks for your suggestion. We have revised the wording to make the research hypothesis clearer. (lines 95-99)

Methods

Major concern: The Methods section is largely missing or insufficiently described in the provided manuscript.

No dedicated Materials and Methods section describing:

Isolate selection criteria

Inclusion/exclusion rules

AST conditions (media, incubation, breakpoints)

Statistical analysis

CLSI references (e.g., M52, breakpoints) are cited later but not introduced systematically.

PCR methods lack primer sequences, cycling conditions, and quality control measures.

This omission represents a major methodological flaw.

Response: Thanks for your suggestion. Considering the journal's formatting requirements, the Methods section in our manuscript has been placed between the Discussion and the References. (lines 286-331)

Lines 126: Rationale for excluding “positive unclassified” strains

Response: Thanks for your suggestion. We did not exclude the positive unclassified isolates; instead, we performed calculations for both scenarios—with and without exclusion—to present a more comprehensive analysis of all data results.

Lines 156: “positive unclassified” strains represent 34.8% of CPOs this is a major limitation but is framed too mildly.

Response: Thanks for your suggestion. In the Discussion, we have dedicated a paragraph to discussing the occurrence of positive unclassified isolates in Klebsiella pneumoniae and Acinetobacter baumannii, aiming to explain the possible reasons for this phenomenon.

“Enterobacteriales” is used instead of “Enterobacterales” (Lines 129–130).

Response: Thanks for your suggestion. We have corrected “enterobacteriales” to “enterobacterales”. (lines 142-143)

Table 1: mCIM is treated as gold standard without justification.

Response: Thanks for your suggestion. The mCIM is still the internationally recommended gold standard for detecting carbapenemase production in Enterobacterales and Pseudomonas aeruginosa; therefore, this method was retained in our study.

Insufficient comparison with prior BD Phoenix CPO panel studies.

Limitations (selection bias, regional strain distribution, exclusion strategy) are inadequately acknowledged.

Response: Thanks for your suggestion. We have added the corresponding description in the Discussion. (lines 268-269)

Reviewer 2 Report

Comments and Suggestions for Authors

This study evaluates the BD Phoenix NMIC-461 Panel that combines MIC testing for ten antimicrobial agents with detection and classification of common carbapenemases. The authors tested over 700 isolates of Enterobacterales, Pseudomonas aeruginosa, and Acinetobacter spp, and compared the Phoenix NMIC-461 panel to standard broth microdilution (BMD), the mCIM method for phenotypic detection of carbapenemase production, and PCR/Sanger sequencing for detection of carbapenemase genes. Several parts of the manuscript need clarification.

The Abstract does not adequately convey the limitations of the CPO assay. The assay correctly classified only 8% of the Ambler A-producing Enterobacterales and incorrectly reported the presence of a carbapenemase in 35% of P. aeruginosa isolates. The magnitude of these limitations needs to be clearly stated in the Abstract.

It would help if the Introduction included a brief description of the CPO detection system. For example, the FDA [https://www.accessdata.fda.gov/cdrh_docs/pdf18/K181665.pdf] describes it as: “CPO detect includes set concentrations of meropenem, doripenem, temocillin, and cloxacillin alone and in combination with various chelators and beta- lactamase inhibitors used in a decision tree.”

Lines 65-67: Unlike avibactam, clavulanic acid and tazobactam have very limited activity against KPC enzymes.

Lines 140-142: Which species accounted for the most of the 61 Enterobacterales isolates harboring an Ambler B enzyme? Lade et al (PMID 37508311), using the NMIC 500 panel, observed that Enterobacter and Citrobacter accounted for the majority of the Ambler B-positive Enterobacterales in a study from Korea.

Table 2: There were no true-positive or false-positive Ambler D results for Enterobacterales and P. aeruginosa; for these organisms, PPV = 0/(0+0) which is undefined, not zero.

Table 2: It would be useful to have a supplemental table showing the MIC results for the 19 P. aeruginosa isolates that had false-positive classifications for Ambler A, B, or D.

Table 3: Why were so few P. aeruginosa isolates tested by BMD? There were 67 P. aeruginosa isolates in the CPO comparison, but only 23 BMD results for ceftazidime, 24 for imipenem, and 33 for meropenem. Also, for drug-organism combinations that do not have defined breakpoints, eg P. aeruginosa vs ceftriaxone, it would be better to report “not tested” rather than 0.

Lines 278-279: “Using the standard broth microdilution (BMD) method as the control method.” This is a sentence fragment. Please clarify.

Lines 287-289: Was CLSI M100 used for breakpoint interpretation?

Figure 1 is probably not necessary; the data could be displayed more compactly in a table.

Lines 296-299: The last portion of this sentence, “indeterminate if zone diameter of 16-18 mm or ≥ 19 mm and the presence of pinpoint colonies within the zone ，as well as the presence or absence of a carbapenemase could not be confirmed” is confusing. Please clarify.

Author Response

Response: Thanks for your suggestion. We have supplemented the abstract to explicitly mention the high proportion of positive unclassified isolates and the low specificity for Pseudomonas aeruginosa. (lines 41-46)

Response: Thanks for your suggestion. Given that this is an academic paper, we did not provide an extensive description of the testing product, and we have instead uploaded the NMIC-461 instructions as supplementary material.

Lines 65-67: Unlike avibactam, clavulanic acid and tazobactam have very limited activity against KPC enzymes.

Response: Thanks for your suggestion. To avoid ambiguity, we have removed the references to clavulanic acid and tazobactam from the revised manuscript. (lines 75)

Response: Thanks for your suggestion. Among the 61 Enterobacterales isolates producing Ambler class B enzymes enrolled in our study, 23 were Escherichia coli and 18 were Klebsiella pneumoniae. I carefully read the literature you recommended, and the number of different species enrolled in Kim et al.'s study differs considerably from ours. This may also be related to differences in isolation rates between different countries.

Table 2: There were no true-positive or false-positive Ambler D results for Enterobacterales and P. aeruginosa; for these organisms, PPV = 0/(0+0) which is undefined, not zero.

Response: Thanks for your suggestion. After careful verification, these values are indeed zero. For example, the PPV for Enterobacterales producing Ambler class D enzymes should be calculated as 0/ (0+1), and the sensitivity is undefined.

Table 2: It would be useful to have a supplemental table showing the MIC results for the 19 P. aeruginosa isolates that had false-positive classifications for Ambler A, B, or D.

Response: Thanks for your suggestion. We have compiled the raw experimental data for all 714 enrolled isolates into a table, which is provided as Supplemental Table S9.

Response: Thanks for your suggestion. We have changed the corresponding entries in Table 3 to "not tested" as per your suggestion.

Lines 278-279: “Using the standard broth microdilution (BMD) method as the control method.” This is a sentence fragment. Please clarify.

Response: Thanks for your suggestion. We have revised that sentence. (lines 295-296)

Lines 287-289: Was CLSI M100 used for breakpoint interpretation?

Response: Thanks for your suggestion. Yes, we referred to the CLSI M100 2024 edition.

Figure 1 is probably not necessary; the data could be displayed more compactly in a table.

Response: Thanks for your suggestion. Figure 1 is primarily designed to illustrate the enrollment numbers for each bacterial species. The bar chart format allows for a clear visual comparison of which species were enrolled in greater numbers. Additionally, we have enlarged the font in Figure 1 to ensure readability for the audience.

Response: Thanks for your suggestion. We have revised this paragraph to make it easier for readers to understand. (lines 316-318)

Reviewer 3 Report

Comments and Suggestions for Authors

This manuscript describes a multicenter evaluation of the BD Phoenix NMIC-461 panel for (i) carbapenemase detection/classification and (ii) AST versus broth microdilution (BMD) across 714 Gram-negative clinical isolates from 3 tertiary hospitals in China. The dataset is sizable and the use of reference methods (BMD, mCIM, PCR/Sanger sequencing) is appropriate. However, the manuscript currently overstates “classification” performance by presenting “improved” sensitivity after excluding “positive unclassified” results; yet, these are part of the real-world output and are frequent. Performance in P. aeruginosa is particularly problematic due to very low PPV/specificity for CPO detection, which has major implications for clinical workflow and infection control decisions. Overall, I recommend major revisions to improve the manuscript according to the below points.

Major:

Readers can’t easily see how Phoenix results match / fail to match reference results, especially when the instrument reports “positive but unclassified.” However, “unclassified” is still a real output in practice and affects how useful the test is. It would be helpful for the manuscript to provide a table that shows what the test said vs what was truly present (e.g., a confusion matrix). One way would be, for each isolate, to record the reference result (no carbapenemase or class A/B/D) alongside the Phoenix result (A, B, D, positive-unclassified, or negative), then summarize all isolates in a single table that counts how often each Phoenix call matches each reference category. This table would make it immediately clear if errors come from missed positives, false positives, wrong class calls, or frequent “positive but unclassified” results, and it highlights any differences between Enterobacterales, P. aeruginosa, and Acinetobacter.
The way the paper describes performance in P. aeruginosa needs to be more cautious, as the test produces many false “carbapenemase-positive” results in this species, which could trigger unnecessary confirmatory testing and infection-control actions; please show where these false positives occur (e.g., by linking them to MIC levels or resistance patterns) and make clear that a positive result in P. aeruginosa should not be treated as a reliable “rule-in” without confirmation.
The reference method used to define which carbapenemases are present is limited to a specific set of genes, so some apparent false positives or “unclassified” results might actually be caused by other enzymes or resistance mechanisms that were not tested; please state this limitation clearly and, if possible, do broader testing (e.g., expanded sequencing) for the discordant isolates.
The number of isolates included in each antibiotic analysis changes because some tests are excluded (e.g., growth/validity issues), but the manuscript does not make this easy to track; please add a clear summary in the main text showing, for each drug (and ideally each organism group), how many isolates were tested and how many were excluded so readers can judge whether exclusions could affect the results.
The manuscript acknowledges statistical support from Becton, Dickinson and Company; please explicitly state whether the manufacturer provided panels, software support, funding, or had any role in study design, analysis, or manuscript preparation, so readers can interpret the findings appropriately.

Minor:

It will be helpful to expand the Introduction section. The introduction would be clearer with a brief summary of prior evaluations of similar BD Phoenix panels, and one sentence clarifying the timeline (isolates from 2020–2022 vs panel introduction in 2024).
Figure 1: The fonts are very small and inconsistent. The figure would benefit from larger font size to improve the readability. The legend should also explain how the organisms are grouped.

Comments on the Quality of English Language

There are several typos; I recommend review by a native speaker

Author Response

Readers can’t easily see how Phoenix results match / fail to match reference results, especially when the instrument reports “positive but unclassified.” However, “unclassified” is still a real output in practice and affects how useful the test is. It would be helpful for the manuscript to provide a table that shows what the test said vs what was truly present (e.g., a confusion matrix). One way would be, for each isolate, to record the reference result (no carbapenemase or class A/B/D) alongside the Phoenix result (A, B, D, positive-unclassified, or negative), then summarize all isolates in a single table that counts how often each Phoenix call matches each reference category. This table would make it immediately clear if errors come from missed positives, false positives, wrong class calls, or frequent “positive but unclassified” results, and it highlights any differences between Enterobacterales, P. aeruginosa, and Acinetobacter.

Response: Thanks for your suggestion. We have compiled the raw experimental data for all 714 enrolled isolates into a table, which is provided as Supplemental Table S9.

The way the paper describes performance in P. aeruginosa needs to be more cautious, as the test produces many false “carbapenemase-positive” results in this species, which could trigger unnecessary confirmatory testing and infection-control actions; please show where these false positives occur (e.g., by linking them to MIC levels or resistance patterns) and make clear that a positive result in P. aeruginosa should not be treated as a reliable “rule-in” without confirmation.

Response: Thanks for your suggestion. In both the abstract and discussion sections of the manuscript, we have included a description of the low specificity observed for Pseudomonas aeruginosa and recommend that its positive results be confirmed using an alternative method. (lines 41-46, and 264-267)

The reference method used to define which carbapenemases are present is limited to a specific set of genes, so some apparent false positives or “unclassified” results might actually be caused by other enzymes or resistance mechanisms that were not tested; please state this limitation clearly and, if possible, do broader testing (e.g., expanded sequencing) for the discordant isolates.

Response: Thanks for your suggestion. We have already described this in the revised manuscript. (lines 264-266)

The number of isolates included in each antibiotic analysis changes because some tests are excluded (e.g., growth/validity issues), but the manuscript does not make this easy to track; please add a clear summary in the main text showing, for each drug (and ideally each organism group), how many isolates were tested and how many were excluded so readers can judge whether exclusions could affect the results.

Response: Thanks for your suggestion. Including the enrollment numbers for each section in the main text of the manuscript would make it overly redundant. We have therefore presented the number of isolates included in the analysis for each section in Supplemental File S3.

The manuscript acknowledges statistical support from Becton, Dickinson and Company; please explicitly state whether the manufacturer provided panels, software support, funding, or had any role in study design, analysis, or manuscript preparation, so readers can interpret the findings appropriately.

Response: Thanks for your suggestion. The test panels used in this study were also provided by BD Company, and this has been acknowledged in the Acknowledgements section. (lines 339-340)

It will be helpful to expand the Introduction section. The introduction would be clearer with a brief summary of prior evaluations of similar BD Phoenix panels, and one sentence clarifying the timeline (isolates from 2020–2022 vs panel introduction in 2024).

Response: Thanks for your suggestion. In the discussion, we have cited several relevant evaluation studies on the BD Phoenix panel. Additionally, a sentence describing and citing previous literature has been included in the introduction.

Figure 1: The fonts are very small and inconsistent. The figure would benefit from larger font size to improve the readability. The legend should also explain how the organisms are grouped.

Response: Thanks for your suggestion. The font size in Figure 1 has been increased to improve readability.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The manuscript is greatly improved. It is acceptable now.

Author Response

Thank you again for your kind words.

Reviewer 2 Report

Comments and Suggestions for Authors

Several sections of the revised manuscript have been clarified.

It would be clearer if "among Enterobacterales" was added in the Abstract on line 41 ["unclassified isolates among Enterobacterales and low specificity for P. aeruginosa"].

There also appears to be some text missing at the end of line 41 (see below).

"encountered Gram-negative bacterial isolates. Regarding carbapenemase detection, the 38 panel demonstrates high sensitivity but only moderate specificity in classifying car-39 bapenemase-producing organisms (CPO), with a relatively high proportion of positive 40 unclassified isolates and low specificity for P. aeruginosa. It is recommended that positive 41 Overall, the implementation of NMIC-461 testing holds promise for significantly reducing 42 turnaround time in both carbapenemase detection and classification." 43

Author Response

Thanks for your comments.

We have added "among Enterobacterales" in the abstract on line 41, and deleted "It is recommended that positive".

Thank you again for your kind words.

Reviewer 3 Report

Comments and Suggestions for Authors

The authors have addressed all my comments satisfactorily. From my point of view it is now eligible for publication.

Comments on the Quality of English Language

There are several typos; I recommend review by a native speaker

Author Response

Thanks for your comments.

We have thoroughly revised the manuscript to correct spelling and grammatical errors.

Thank you again for your kind words.