Journal of Clinical Microbiology最新文献

{"title":"Impact of packed red blood cells versus fresh whole human blood on the growth of nutritionally variant streptococci and <i>Streptococcus anginosus</i> group during verification of the BACT/ALERT VIRTUO blood culture system.","authors":"Heather Glassman, Jasmine Ahmed-Bentley, Rebecca Buckman, Mark Carbungco, Colleen Fletcher, Kim Sy, Joyce Wong, Charlene Porter, Claudine Desruisseaux, Valery Lavergne, Suefay Liu, Anthony Lieu, Marthe K Charles","doi":"10.1128/jcm.01987-24","DOIUrl":"https://doi.org/10.1128/jcm.01987-24","url":null,"abstract":"","PeriodicalId":15511,"journal":{"name":"Journal of Clinical Microbiology","volume":" ","pages":"e0198724"},"PeriodicalIF":6.1,"publicationDate":"2025-03-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143752958","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Improvement of the diagnosis of intestinal protozoa using a multiplex qPCR strategy compared to classical microscopy: a prospective study on 3,500 stool samples over 3 years.","authors":"Florence Robert-Gangneux, Xavier Duval, Clément Cazala, Sorya Belaz, Anne Dupuis, Hélène Guegan, Brice Autier, Jean-Pierre Gangneux","doi":"10.1128/jcm.01610-24","DOIUrl":"https://doi.org/10.1128/jcm.01610-24","url":null,"abstract":"<p><p>Commercial multiplex real-time PCR (qPCR assays) are now widely used for the diagnosis of intestinal protozoan infections, but few prospective studies evaluated their performances on large patient cohorts. We extracted data from our information system from 1 January 2021 to 15 March 2024 and included all stool samples analyzed in routine. Parasites were searched using a multiplex PCR (AllPlex Gastrointestinal Panel assay, Seegene) and microscopic examination with two concentration methods. Acid-fast staining was performed when <i>Cryptosporidium</i> detection was specifically requested. In total, 3,495 stools were analyzed from 2,127 patients. <i>Giardia intestinalis</i>, <i>Cryptosporidium</i> spp., <i>Entamoeba histolytica</i>, <i>Dientamoeba fragilis,</i> and <i>Blastocystis</i> spp. were found by multiplex qPCR in 45 (1.28%), 30 (0.85%), 9 (0.25%), 310 (8.86%), and 673 (19.25%) samples, respectively, alone or in combination (<i>n</i> = 909). In the vast majority of cases, PCR detected a protozoan on the first stool sample. Microscopy was positive for <i>G. intestinalis</i>, <i>Cryptosporidium</i> spp., <i>Entamoeba histolytica/dispar</i>, <i>D. fragilis,</i> and <i>Blastocystis</i> spp. in 25 (0.7%), 8 (0.23%), 24 (0.68%), 22 (0.63%), and 229 (6.55%), samples, respectively, alone or in combination (<i>n</i> = 286 samples). No samples were PCR-/Microscopy+ for <i>G</i>. <i>intestinalis</i>, <i>Cryptosporidium</i> spp., and <i>E. histolytica</i>, while <i>D. fragilis</i> and <i>Blastocystis</i> spp. were detected only with microscopy in 6 and 20 samples, respectively. Microscopy allowed the detection of parasites not targeted by the multiplex panel (5 <i>Cystoisospora belli,</i> 331 samples with non-pathogenic protozoa, and 68 samples with helminths). Overall, the multiplex PCR proved more efficient to detect protozoan parasites, but a microscopic technique should be performed when infection with <i>C. belli</i> (HIV-infected patients) or helminths is suspected (migrants and travelers).IMPORTANCEIn the era of increasing use of multiplex PCR panels for the diagnosis of intestinal protozoan infections, it is important to form an opinion on the positioning of those assays within a lab workflow. This study analyzes routine results obtained prospectively by microscopy and a commercial multiplex PCR over 3 years, and shows that the assay meets the expectations of a clinical laboratory for the detection of protozoan parasites of medical interest. It is recalled that <i>Cystoisospora belli</i> is not targeted by the multiplex assay, and that microscopy still remains necessary to detect helminths.</p>","PeriodicalId":15511,"journal":{"name":"Journal of Clinical Microbiology","volume":" ","pages":"e0161024"},"PeriodicalIF":6.1,"publicationDate":"2025-03-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143752960","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Rapid prediction of carbapenemases in <i>Pseudomonas aeruginosa</i> by imipenem/relebactam and MALDI-TOF MS.","authors":"Ana Candela, María Fernández-Billón, Pablo Aja-Macaya, Lucía González-Pinto, Pablo Arturo Fraile-Ribot, Esther Viedma, Isaac Alonso-García, Tania Blanco-Martín, Roberto Estévez-Alfaya, Ana Fernández-González, Alejandro Beceiro, Carla López-Causapé, Marina Oviaño, Germán Bou, Antonio Oliver, Jorge Arca-Suárez","doi":"10.1128/jcm.01105-24","DOIUrl":"https://doi.org/10.1128/jcm.01105-24","url":null,"abstract":"<p><p><i>Pseudomonas aeruginosa</i> is a major nosocomial pathogen commonly involved in multidrug-resistant (MDR) infections that are very difficult to treat. Imipenem/relebactam is a new carbapenem/β-lactamase inhibitor combination with robust activity against <i>P. aeruginosa</i>. However, resistance is increasingly reported, and rapid detection is, therefore, crucial so that appropriate treatments can be prescribed. We have developed a rapid MALDI TOF-MS-based method that can accurately predict the presence of carbapenemases in <i>P. aeruginosa</i> using imipenem/relebactam. The method was developed using a retrospective and a prospective collection of 419 <i>P</i>. <i>aeruginosa</i> isolates (including recombinant isolates and WGS-characterized clinical strains) encompassing the most important β-lactam resistance mechanisms. The MALDI TOF-MS method is based on the detection of the hydrolysis of imipenem in the presence or absence of relebactam, measuring modifications in the mass spectra of imipenem after incubation with bacteria. The method was evaluated against a retrospective collection and then validated against 250 prospectively collected clinical isolates, showing a 98% (246/250) agreement between the phenotype and the MALDI-TOF MS hydrolysis result and a 100% accordance with the β-lactam resistance genotype. Some errors in detecting GES-producing isolates and in detecting different mutational resistance mechanisms associated with imipenem/relebactam resistance (MICs ranging from 4 to 8 mg/L) were observed. All results were obtained within 1 hour, positioning the MALDI-TOF-based test as a rapid and easy-to-perform method for detection of carbapenemases (except GES enzymes) in <i>P. aeruginosa</i>. Besides, implementation of the method in routine laboratory screening would facilitate the correct use of imipenem/relebactam to treat <i>P. aeruginosa</i> infections.IMPORTANCEWhile several rapid diagnostic methods have been developed for the detection of ESBLs and carbapenemases to improve treatment decision-making in Enterobacterales, there is a lack of approaches to rapidly identify resistance mechanisms and predict β-lactam susceptibility in <i>Pseudomonas aeruginosa</i>. Taking advantage of the mechanism of action and the high efficacy of the newly developed β-lactam/β-lactamase inhibitor combination imipenem/relebactam against <i>P. aeruginosa</i>, we developed a WGS-guided, MALDI-TOF-based algorithm that accurately predicts the presence of carbapenemase enzymes in this bacterium and aids in forecasting the imipenem/relebactam susceptibility profile. The implementation of this method in routine laboratory testing would provide significant support in the rapid decision-making for the use of imipenem/relebactam in severe <i>P. aeruginosa</i> infections.</p>","PeriodicalId":15511,"journal":{"name":"Journal of Clinical Microbiology","volume":" ","pages":"e0110524"},"PeriodicalIF":6.1,"publicationDate":"2025-03-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143700645","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Evaluation of the cobas MTB and MTB-RIF/INH assay in clinical samples for the detection of <i>Mycobacterium tuberculosis</i> in respiratory specimens.","authors":"Ulrich Eigner, Jasmin Köffer, Ulrike Betz, Janina Koglin, Elvira Richter","doi":"10.1128/jcm.01959-24","DOIUrl":"https://doi.org/10.1128/jcm.01959-24","url":null,"abstract":"<p><p>The aim of this study was to evaluate the performance of the automated cobas MTB-Real-Time PCR assay for the rapid direct detection of <i>Mycobacterium tuberculosis</i> complex (MTBC) in clinical specimens and the ability of the cobas MTB-RIF/INH assay to correctly detect drug resistance to rifampin (RIF) and isoniazid (INH). The PCR assays were set up on the automated Cobas 6800 system, and the results were compared to liquid culture using BACTEC mycobacteria growth indicator tubes 960 TB system as the gold standard and line probe assays or sequencing results. A total of 500 N-acetyl-L-cysteine/sodium hydroxide (NALC-NaOH)-processed sputum samples were tested with the respective methods. The performance of MTBC detection in pulmonary specimens showed 91.8% sensitivity and 99.3% specificity in comparison to culture. The sensitivity for acid-fast bacteria (AFB) smear-positive specimens and for AFB smear-negative specimens was 100% and 85.1%, respectively. Due to the low prevalence of tuberculosis (TB) resistance in Germany, a collection of resistant TB strains with a wide variety of mutations was analyzed. The cobas MTB-RIF/INH assay detected 19 out of 21 INH-resistant and 22 out of 24 RIF-resistant TB strains. In conclusion, the cobas MTB and the cobas MTB-RIF/INH assays implemented on the automated cobas6800 instrument are reliable and versatile tools for the detection of MTB and RIF/INH resistance.</p><p><strong>Importance: </strong>Our manuscript addresses the WHO recommendation for the use of \"moderate-complexity automated NAATs for detection of TB and resistance to rifampicin and isoniazid\" as a part of the WHO End TB Strategy. Rapid detection of tuberculosis (TB) patients is essential to preventing TB transmission and finally reducing TB burden. In this study, we present data on the sensitivity and specificity of the novel cobas MTB assay for TB detection in a low-incidence country, demonstrating highly promising results. Additionally, by analyzing TB strains with various mutations conferring resistance to INH and/or RMP, we assess the opportunities and limitations of the cobas MTB-RIF/INH assay in reliably detecting drug resistance in sputum specimens, thereby facilitating the early onset of appropriate treatment.</p>","PeriodicalId":15511,"journal":{"name":"Journal of Clinical Microbiology","volume":" ","pages":"e0195924"},"PeriodicalIF":6.1,"publicationDate":"2025-03-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143700628","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Hematology thin smears perform equally to parasitology thick and thin blood smears for the diagnosis of <i>Plasmodium</i> and <i>Babesia</i> infections in a low prevalence setting.","authors":"Janmesh Patel, Jill Schuett, Derrick J Chen","doi":"10.1128/jcm.01601-24","DOIUrl":"https://doi.org/10.1128/jcm.01601-24","url":null,"abstract":"<p><p>Malaria and babesiosis are significant parasitic infections, requiring timely diagnosis to avoid severe complications. This study retrospectively compared hematology thin smears (HS) with parasitology thick and thin blood smears (PS), the gold standard for diagnosis, to evaluate HS's performance in detecting <i>Plasmodium</i> and <i>Babesia</i> infections. Of 529 cases with paired HS and PS testing, HS demonstrated 93.3% sensitivity and 99.8% specificity, with 97.7% positive and 99.4% negative predictive values. When only considering new diagnoses, HS and PS were 100% concordant. No significant difference was found in percent parasitemia between HS and PS, highlighting HS as a reliable diagnostic tool in settings where PS or other diagnostic modalities may not be readily available.IMPORTANCEThis study demonstrates that hematology thin smears-often available in laboratories that may not have other means of diagnosing blood parasite infections such as parasitology thick and thin smears, rapid diagnostics tests, or polymerase chain reaction-are an accurate and reliable way to diagnose <i>Plasmodium</i> and <i>Babesia</i> infections in a low prevalence setting.</p>","PeriodicalId":15511,"journal":{"name":"Journal of Clinical Microbiology","volume":" ","pages":"e0160124"},"PeriodicalIF":6.1,"publicationDate":"2025-03-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143700633","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Correction for Lamberink et al., \"Multicenter validation of a galactomannan chemiluminescence immunoassay for the diagnosis of pulmonary aspergillosis on serum of patients with hematological disease\".","authors":"Hanne Lamberink, Sammy Huygens, Robina Aerts, Katrien Lagrou, Karin van Dijk, Diana Langerak, Ine Moors, Jerina Boelens, Marijke Reynders, Johan Maertens, Alexander Schauwvlieghe, Mireille van Westreenen, Ga-Lai M Chong, Paul E Verweij, Jochem B Buil, Bart J A Rijnders","doi":"10.1128/jcm.00394-25","DOIUrl":"https://doi.org/10.1128/jcm.00394-25","url":null,"abstract":"","PeriodicalId":15511,"journal":{"name":"Journal of Clinical Microbiology","volume":" ","pages":"e0039425"},"PeriodicalIF":6.1,"publicationDate":"2025-03-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143700553","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Impact of media brand on cefiderocol disk diffusion results.","authors":"M G DeMarco, A M Field, L E Donohue, H L Cox, T S Kidd, K E Barry, A J Mathers","doi":"10.1128/jcm.01648-24","DOIUrl":"https://doi.org/10.1128/jcm.01648-24","url":null,"abstract":"&lt;p&gt;&lt;p&gt;Cefiderocol is a siderophore cephalosporin that utilizes iron transport systems to cross cell membranes. This unique strategy complicates antimicrobial susceptibility testing (AST) due to variable iron content in media. While guidance for using iron-depleted media exists for broth microdilution (BMD), disk diffusion (DD) with commercial media is a common AST method in the clinical laboratory. We investigated cefiderocol DD result variability using multiple Mueller-Hinton agar (MHA) brands. DD results using Remel (Thermo Fisher Scientific, San Diego, CA), Hardy (Hardy Diagnostics, Springboro, OH), and BBL (Becton Dickinson, East Rutherford, NJ) MHA were compared to those of BMD using iron-depleted cation-adjusted Mueller-Hinton broth. BMD reproducibility, BMD trailing endpoints, and DD intra- and inter-brand variability in zones of inhibition were investigated. Forty-seven multidrug-resistant clinical isolates and three Antibiotic Resistance Bank isolates composed of &lt;i&gt;Pseudomonas aeruginosa&lt;/i&gt;, carbapenemase (CP-) and non-carbapenemase-producing (non-CP-) carbapenem-resistant Enterobacterales (CRE), &lt;i&gt;Acinetobacter baumannii&lt;/i&gt; complex, &lt;i&gt;Stenotrophomonas maltophilia&lt;/i&gt;, and &lt;i&gt;Burkholderia cepacia&lt;/i&gt; complex were tested. Categorical agreement (CA) ≥ 90% was only demonstrated using CLSI breakpoints with BBL agar. Intra- and inter-brand variability in DD were highest for &lt;i&gt;P. aeruginosa&lt;/i&gt; and CRE, with 25% (5/20) and 16.7% (3/18) exhibiting discrepant AST interpretations, respectively. Isolates not susceptible to cefiderocol via BMD were commonly associated with AST interpretation errors and lower CA. Using commercial MHA for DD resulted in frequent AST interpretation discrepancies, particularly for isolates that were not susceptible to cefiderocol by BMD. Methods and quality control may need to be revisited to ensure the reliability of DD for cefiderocol AST.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Importance: &lt;/strong&gt;The novel mechanism of action of cefiderocol overcomes a variety of resistance mechanisms associated with gram-negative bacteria and positions the agent as an attractive option for treating infections involving multidrug-resistant pathogens. The availability of accurate, timely antimicrobial susceptibility testing methods for cefiderocol in clinical microbiology laboratories is critical as cefiderocol-resistant isolates have been described and may contribute to treatment failure. Iron-depleted broth microdilution testing may not be feasible for use in many clinical laboratories. While disk diffusion is an appealing, practical method to implement, our data demonstrate reproducibility issues across agar brands, most notably for organisms that do not test susceptible to cefiderocol when using broth microdilution. Discrepancy errors and misclassifications of resistant isolates as susceptible, and susceptible isolates as resistant, may mislead clinicians and compromise the treatment efficacy. More work is needed to standardize practical yet repro","PeriodicalId":15511,"journal":{"name":"Journal of Clinical Microbiology","volume":" ","pages":"e0164824"},"PeriodicalIF":6.1,"publicationDate":"2025-03-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143700640","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"An overview of the laboratory diagnosis of <i>Pneumocystis jirovecii</i> pneumonia.","authors":"Alexis Jaramillo Cartagena, Osaretin Emmanuel Asowata, Dianna Ng, N Esther Babady","doi":"10.1128/jcm.00361-24","DOIUrl":"10.1128/jcm.00361-24","url":null,"abstract":"<p><p><i>Pneumocystis jirovecii</i> (<i>P. jirovecii</i>) is a fungal pathogen associated with significant morbidity in immunocompromised patients, including both HIV- and non-HIV-infected patients. The nonspecific clinical and radiological presentation makes clinical diagnostic challenging, emphasizing the need for accurate laboratory diagnostic tests. However, <i>P. jirovecii</i> does not grow in routine culture media, which presents diagnostic challenges in the laboratory as well. Recent publications from the European Organization for Research and Treatment of Cancer and the Mycoses Study Group Education and Research Consortium continue to rely on direct detection of <i>P. jirovecii</i> organisms in tissues and respiratory samples to define proven <i>P. jirovecii</i> pneumonia (PCP) even as the sensitivity of these methods are lower. Novel, standardized methods are needed to improve the clinical and laboratory diagnosis and management of PCP. This minireview provides an overview of current diagnostic tests for PCP and emerging applications that aim at filling existing diagnostic gaps and providing more accurate and less invasive diagnoses for this significant disease.</p>","PeriodicalId":15511,"journal":{"name":"Journal of Clinical Microbiology","volume":" ","pages":"e0036124"},"PeriodicalIF":6.1,"publicationDate":"2025-03-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11898755/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143080214","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Comparison of the Filamentous Fungi Library v4.0 MALDI Biotyper Platform vs MSI-2 performance for identifying filamentous fungi from liquid cultures.","authors":"María F Gonzalez-Lara, Carla M Román-Montes, Paulette Díaz-Lomelí, Winston Hernández-Ceballos, Lizeth Morales-Camilo, Axel Cervantes-Sánchez, Andrea Cordero-Rangel, Jazmín Tejeda-Olán, Alexandro Bonifaz-Trujillo, Alfredo Ponce-de-León, Areli Martínez-Gamboa","doi":"10.1128/jcm.01371-24","DOIUrl":"10.1128/jcm.01371-24","url":null,"abstract":"<p><p>Correct, rapid, and reliable filamentous fungi identification is crucial for timely diagnosis and therapy. We compared the performance of the FFLv4.0 MALDI Biotyper and MSI-2 to identify filamentous fungi from clinical isolates. We analyzed 307 clinical isolates of <i>Aspergillus</i> spp.<i>, Fusarium</i> spp.<i>,</i> and <i>Mucorales</i> and compared them to sequencing as the reference standard. The overall identification rates to genus (<i>Mucorales</i>), section (<i>Aspergillus</i>), and species complex (<i>Fusarium</i>) level were 96% (296/307) for FFLv4.0 and 78.5% (241/307) for MSI-2. By each genus, correct species identification was achieved by FFLv4.0 and MSI-2 as follows: 72.4% (165/228) and 55.3% (126/228) for <i>Aspergillus</i> species, 17.6% (6/34) and 38.2% (13/34) for <i>Fusarium</i> species, and 88.9% (40/45) and 55.5% (25/45) for the <i>Mucorales</i>. The rates of non-identification by FFLv4.0 and MSI-2, respectively, were 4% (9/228) and 18% (41/228) for <i>Aspergillus</i> spp., 0% and 17.6% (6/34) for <i>Fusarium</i> spp. and 4.4% (2/45) and 42.2% (19/45%) for the <i>Mucorales</i>. Misidentification rates by FFLv4.0 and MSI-2, respectively, were 16.2% (37/228) and 6.1% (14/228) for <i>Aspergillus</i> species, 67.6% (23/34) and 14.7% (5/34) for <i>Fusarium</i> spp., and 0% and 2.2% (1/45) for the <i>Mucorales</i>. The FFLv4.0 MALDI Biotyper outperformed MSI-2 in identifying filamentous fungi from liquid culture spectra.</p>","PeriodicalId":15511,"journal":{"name":"Journal of Clinical Microbiology","volume":" ","pages":"e0137124"},"PeriodicalIF":6.1,"publicationDate":"2025-03-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11898572/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143065888","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A novel framework for the automated characterization of Gram-stained blood culture slides using a large-scale vision transformer.","authors":"Jack McMahon, Naofumi Tomita, Elizabeth S Tatishev, Adrienne A Workman, Cristina R Costales, Niaz Banaei, Isabella W Martin, Saeed Hassanpour","doi":"10.1128/jcm.01514-24","DOIUrl":"10.1128/jcm.01514-24","url":null,"abstract":"<p><p>This study introduces a new framework for the artificial intelligence-based characterization of Gram-stained whole-slide images (WSIs). As a test for the diagnosis of bloodstream infections, Gram stains provide critical early data to inform patient treatment in conjunction with data from rapid molecular tests. In this work, we developed a novel transformer-based model for Gram-stained WSI classification, which is more scalable to large data sets than previous convolutional neural network-based methods as it does not require patch-level manual annotations. We also introduce a large Gram stain data set from Dartmouth-Hitchcock Medical Center (Lebanon, New Hampshire, USA) to evaluate our model, exploring the classification of five major categories of Gram-stained WSIs: gram-positive cocci in clusters, gram-positive cocci in pairs/chains, gram-positive rods, gram-negative rods, and slides with no bacteria. Our model achieves a classification accuracy of 0.858 (95% CI: 0.805, 0.905) and an area under the receiver operating characteristic curve (AUC) of 0.952 (95% CI: 0.922, 0.976) using fivefold nested cross-validation on our 475-slide data set, demonstrating the potential of large-scale transformer models for Gram stain classification. Results were measured against the final clinical laboratory Gram stain report after growth of organism in culture. We further demonstrate the generalizability of our trained model by applying it without additional fine-tuning on a second 27-slide external data set from Stanford Health (Palo Alto, California, USA) where it achieves a binary classification accuracy of 0.926 (95% CI: 0.885, 0.960) and an AUC of 0.8651 (95% CI: 0.6337, 0.9917) while distinguishing gram-positive from gram-negative bacteria.</p><p><strong>Importance: </strong>This study introduces a scalable transformer-based deep learning model for automating Gram-stained whole-slide image classification. It surpasses previous methods by eliminating the need for manual annotations and demonstrates high accuracy and generalizability across multiple data sets, enhancing the speed and reliability of Gram stain analysis.</p>","PeriodicalId":15511,"journal":{"name":"Journal of Clinical Microbiology","volume":" ","pages":"e0151424"},"PeriodicalIF":6.1,"publicationDate":"2025-03-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11898657/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143483329","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}