Explore the vast range of titles available.

Background Chronic wounds affect millions of people and cost billions of dollars in the United States each year. These wounds harbor polymicrobial biofilm communities, which can be difficult to elucidate using culturing methods. Clinical molecular microbiological methods are increasingly being employed to investigate the microbiota of chronic infections, including wounds, as part of standard patient care. However, molecular testing is more sensitive than culturing, which results in markedly different results being reported to clinicians. This study compares the results of aerobic culturing and molecular testing (culture-free 16S ribosomal DNA sequencing), and it examines the relative abundance score that is generated by the molecular test and the usefulness of the relative abundance score in predicting the likelihood that the same organism would be detected by culture. Methods Parallel samples from 51 chronic wounds were studied using aerobic culturing and 16S DNA sequencing for the identification of bacteria. Results One hundred forty-five (145) unique genera were identified using molecular methods, and 68 of these genera were aerotolerant. Fourteen (14) unique genera were identified using aerobic culture methods. One-third (31/92) of the cultures were determined to be < 1% of the relative abundance of the wound microbiota using molecular testing. At the genus level, molecular testing identified 85% (78/92) of the bacteria that were identified by culture. Conversely, culturing detected 15.7% (78/497) of the aerotolerant bacteria and detected 54.9% of the collective aerotolerant relative abundance of the samples. Aerotolerant bacterial genera (and individual species including Staphylococcus aureus , Pseudomonas aeruginosa , and Enterococcus faecalis ) with higher relative abundance scores were more likely to be detected by culture as demonstrated with regression modeling. Conclusion Discordance between molecular and culture testing is often observed. However, culture-free 16S ribosomal DNA sequencing and its relative abundance score can provide clinicians with insight into which bacteria are most abundant in a sample and which are most likely to be detected by culture.

Pseudomonas aeruginosa is a persistent and difficult-to-treat pathogen in many patients, especially those with Cystic Fibrosis (CF). Herein, we describe a longitudinal analysis of a series of multidrug resistant (MDR) P. aeruginosa isolates recovered in a 17-month period, from a young female CF patient who underwent double lung transplantation. Our goal was to understand the genetic basis of the observed resistance phenotypes, establish the genomic population diversity, and define the nature of sequence evolution over time. Twenty-two sequential P. aeruginosa isolates were obtained within a 17-month period, before and after a double-lung transplant. At the end of the study period, antimicrobial susceptibility testing, whole genome sequencing (WGS), phylogenetic analyses and RNAseq were performed in order to understand the genetic basis of the observed resistance phenotypes, establish the genomic population diversity, and define the nature of sequence changes over time. The majority of isolates were resistant to almost all tested antibiotics. A phylogenetic reconstruction revealed 3 major clades representing a genotypically and phenotypically heterogeneous population. The pattern of mutation accumulation and variation of gene expression suggested that a group of closely related strains was present in the patient prior to transplantation and continued to change throughout the course of treatment. A trend toward accumulation of mutations over time was observed. Different mutations in the DNA mismatch repair gene mutL consistent with a hypermutator phenotype were observed in two clades. RNAseq performed on 12 representative isolates revealed substantial differences in the expression of genes associated with antibiotic resistance and virulence traits. The overwhelming current practice in the clinical laboratories setting relies on obtaining a pure culture and reporting the antibiogram from a few isolated colonies to inform therapy decisions. Our analyses revealed significant underlying genomic heterogeneity and unpredictable evolutionary patterns that were independent of prior antibiotic treatment, highlighting the need for comprehensive sampling and population-level analysis when gathering microbiological data in the context of CF P. aeruginosa chronic infection. Our findings challenge the applicability of antimicrobial stewardship programs based on single-isolate resistance profiles for the selection of antibiotic regimens in chronic infections such as CF.

Modern artificial intelligence (AI) and machine learning (ML) methods are now capable of completing tasks with performance characteristics that are comparable to those of expert human operators. As a result, many areas throughout healthcare are incorporating these technologies, including in vitro diagnostics and, more broadly, laboratory medicine. However, there are limited literature reviews of the landscape, likely future, and challenges of the application of AI/ML in laboratory medicine. In this review, we begin with a brief introduction to AI and its subfield of ML. The ensuing sections describe ML systems that are currently in clinical laboratory practice or are being proposed for such use in recent literature, ML systems that use laboratory data outside the clinical laboratory, challenges to the adoption of ML, and future opportunities for ML in laboratory medicine. AI and ML have and will continue to influence the practice and scope of laboratory medicine dramatically. This has been made possible by advancements in modern computing and the widespread digitization of health information. These technologies are being rapidly developed and described, but in comparison, their implementation thus far has been modest. To spur the implementation of reliable and sophisticated ML-based technologies, we need to establish best practices further and improve our information system and communication infrastructure. The participation of the clinical laboratory community is essential to ensure that laboratory data are sufficiently available and incorporated conscientiously into robust, safe, and clinically effective ML-supported clinical diagnostics.

* Context.--Machine learning (ML) allows for the analysis of massive quantities of high-dimensional clinical laboratory data, thereby revealing complex patterns and trends. Thus, ML can potentially improve the efficiency of clinical data interpretation and the practice of laboratory medicine. However, the risks of generating biased or unrepresentative models, which can lead to misleading clinical conclusions or overestimation of the model performance, should be recognized. (Arch Pathol Lab Med. 2023;147:826-836; doi: 10.5858/ arpa.2021-0635-RA)

Machine learning (ML) allows for the analysis of massive quantities of high-dimensional clinical laboratory data, thereby revealing complex patterns and trends. Thus, ML can potentially improve the efficiency of clinical data interpretation and the practice of laboratory medicine. However, the risks of generating biased or unrepresentative models, which can lead to misleading clinical conclusions or overestimation of the model performance, should be recognized. To discuss the major components for creating ML models, including data collection, data preprocessing, model development, and model evaluation. We also highlight many of the challenges and pitfalls in developing ML models, which could result in misleading clinical impressions or inaccurate model performance, and provide suggestions and guidance on how to circumvent these challenges. The references for this review were identified through searches of the PubMed database, US Food and Drug Administration white papers and guidelines, conference abstracts, and online preprints. With the growing interest in developing and implementing ML models in clinical practice, laboratorians and clinicians need to be educated in order to collect sufficiently large and high-quality data, properly report the data set characteristics, and combine data from multiple institutions with proper normalization. They will also need to assess the reasons for missing values, determine the inclusion or exclusion of outliers, and evaluate the completeness of a data set. In addition, they require the necessary knowledge to select a suitable ML model for a specific clinical question and accurately evaluate the performance of the ML model, based on objective criteria. Domain-specific knowledge is critical in the entire workflow of developing ML models.

Clinical diagnostics of chronic polymicrobial infections, such as those found in chronic wounds, represent a diagnostic challenge for both culture and molecular methods. In the current retrospective study, the results of aerobic bacterial cultures and culture-free bacterial identification using DNA analyses were compared. A total of 168 chronic wounds were studied. The majority of bacteria identified with culture testing were also identified with molecular testing, but the majority of bacteria identified with the molecular testing were not identified with culture testing. Seventeen (17) different bacterial taxa were identified with culture, and 338 different bacterial taxa were identified with molecular testing. This study demonstrates the increased sensitivity that molecular microbial identification can have over culture methodologies, and previous studies suggest that molecular bacterial identification can improve the clinical outcomes of patients with chronic wounds.

Abstract Background Urine culture images collected using bacteriology automation are currently interpreted by technologists during routine standard-of-care workflows. Machine learning may be able to improve the harmonization of and assist with these interpretations. Methods A deep learning model, BacterioSight, was developed, trained, and tested on standard BD-Kiestra images of routine blood agar urine cultures from 2 different medical centers. Results BacterioSight displayed performance on par with standard-of-care–trained technologist interpretations. BacterioSight accuracy ranged from 97% when compared to standard-of-care (single technologist) and reached 100% when compared to a consensus reached by a group of technologists (gold standard in this study). Variability in image interpretation by trained technologists was identified and annotation “fuzziness” was quantified and found to correlate with reduced confidence in BacterioSight interpretation. Intra-testing (training and testing performed within the same institution) performed well giving Area Under the Curve (AUC) ≥0.98 for negative and positive plates, whereas, cross-testing on images (trained on one institution’s images and tested on images from another institution) showed decreased performance with AUC ≥0.90 for negative and positive plates. Conclusions Our study provides a roadmap on how BacterioSight or similar deep learning prototypes may be implemented to screen for microbial growth, flag difficult cases for multi-personnel review, or auto-verify a subset of cultures with high confidence. In addition, our results highlight image interpretation variability by trained technologist within an institution and globally across institutions. We propose a model in which deep learning can enhance patient care by identifying inherent sample annotation variability and improving personnel training.

The measurement of cytokines in clinical laboratories is becoming an increasingly routine part of immune monitoring when administering biologic and cell-based immunotherapies and also for clinical assessment of inflammatory conditions. While a number of commercial assays and platforms are available for cytokine measurement, there is currently little standardization among these analytical methods. To characterize the variability and comparability among cytokine testing platforms that are commonly used in clinical laboratories. We analyzed data for 4 cytokines (interleukin [IL]-1, IL-6, IL-8, and tumor necrosis factor-alpha [TNF-α]) from 6 College of American Pathologists cytokine surveys administered from 2015 to 2018. Analyses interrogated variability between testing methods and variability within each laboratory across the mailings. Significant variability was noted across methods with analysis of IL-1 showing the least variability and IL-6, IL-8, and TNF-α varying between methods to a greater extent. Intralab variability was also significant with TNF-α measurements again showing the greatest variability. This retrospective analysis of College of American Pathologists proficiency testing data for cytokine measurement is the largest method comparison to date, and this study provides a description of the variation of cytokine measurement across methods, across laboratories, and within laboratories. Serial monitoring of cytokines should preferentially be performed by the same method within the same laboratory.

[...]the authors trained a commercially available cloud-based image analysis software, WebMicroscope (https://webmicroscope.com/imageanalysis/), to perform a two-step analysis pipeline. Comments Image analysis software is on the verge of becoming mainstream in clinical microbiology as a growing number of laboratories are converting to digital plate reading of bacterial cultures as a component of total laboratory automation. Performance evaluation of the CellaVision DM96 system: WBC differentials by automated digital image analysis supported by an artificial neural network.