Archives of pathology & laboratory medicine最新文献

Context.—: Artificial intelligence algorithms hold the potential to fundamentally change many aspects of society. Application of these tools, including the publicly available ChatGPT, has demonstrated impressive domain-specific knowledge in many areas, including medicine.

Objectives.—: To understand the level of pathology domain-specific knowledge for ChatGPT using different underlying large language models, GPT-3.5 and the updated GPT-4.

Design.—: An international group of pathologists (n = 15) was recruited to generate pathology-specific questions at a similar level to those that could be seen on licensing (board) examinations. The questions (n = 15) were answered by GPT-3.5, GPT-4, and a staff pathologist who recently passed their Canadian pathology licensing exams. Participants were instructed to score answers on a 5-point scale and to predict which answer was written by ChatGPT.

Results.—: GPT-3.5 performed at a similar level to the staff pathologist, while GPT-4 outperformed both. The overall score for both GPT-3.5 and GPT-4 was within the range of meeting expectations for a trainee writing licensing examinations. In all but one question, the reviewers were able to correctly identify the answers generated by GPT-3.5.

Conclusions.—: By demonstrating the ability of ChatGPT to answer pathology-specific questions at a level similar to (GPT-3.5) or exceeding (GPT-4) a trained pathologist, this study highlights the potential of large language models to be transformative in this space. In the future, more advanced iterations of these algorithms with increased domain-specific knowledge may have the potential to assist pathologists and enhance pathology resident training.

Context.—: Biomarker reporting has increasingly become a key component of pathology reporting, providing diagnostic, prognostic, and actionable therapeutic data for patient care.

Objective.—: To expand and improve the College of American Pathologists (CAP) biomarker protocols.

Design.—: We surveyed CAP members to better understand the limitations they experienced when reporting cancer biomarker results. A Biomarker Workgroup reviewed the survey results and developed a strategy to improve and standardize biomarker reporting. Drafts of new and revised biomarker protocols were reviewed in both print and electronic template formats during interactive webinars presented to the CAP House of Delegates. Feedback was collected, and appropriate revisions were made to finalize the protocols.

Results.—: The first phase of the CAP Biomarker Workgroup saw the development of (1) a new stand-alone general Immunohistochemistry Biomarker Protocol that includes reporting for ER (estrogen receptor), PR (progesterone receptor), Ki-67, HER2 (human epidermal growth factor receptor 2), PD-L1 (programmed death ligand-1), and mismatch repair; (2) a new Head and Neck Biomarker Protocol that updates the prior 2017 paper-only version into an electronic template, adding new diagnostic and theranostic markers; (3) a major revision to the Lung Biomarker Protocol to streamline it and add in pan-cancer markers; and (4) a revision to the Colon and Rectum Biomarker Protocol to add HER2 reporting.

Conclusions.—: We have taken a multipronged approach to improving biomarker reporting in the CAP cancer protocols. We continue to review current biomarker reporting protocols to reduce and eliminate unnecessary methodologic details and update with new markers as needed. The biomarker templates will serve as standardized modular units that can be inserted into cancer-reporting protocols.

Context.—: Existing targeted cystic fibrosis screening assays miss important pathogenic CFTR variants in the ethnically diverse US population.

Objective.—: To evaluate the analytic performance of a multiplex polymerase chain reaction (PCR)/capillary electrophoresis (CE) CFTR assay panel that simultaneously interrogates primary pathogenic variants of different ethnic/ancestral groups.

Design.—: Performance characteristic assessment and variant coverage comparison of the panel with a focus on ethnicity-specific CFTR variants were performed. Sample DNA was primarily from whole blood or cell lines. Detection of CFTR carriers was compared across several commercially available CFTR kits and recommended variant sets based on panel content.

Results.—: The panel interrogated 65 pathogenic CFTR variants representing 92% coverage from a recent genomic sequencing survey of the US population, including 4 variants with top 5 frequency in African or Asian populations not reflected in other targeted panels. In simulation studies, the panel represented 95% of carriers across the global population, resulting in a 6.9% to 19.0% higher carrier detection rate compared with 10 targeted panels or variant sets. Precision and sensitivity/specificity were 100% concordant. Multisite sample-level genotyping accuracy was 99.2%. Across PCR and CE instruments, sample-level genotyping accuracy was 97.1%, with greater than 99% agreement for all variant-level metrics.

Conclusions.—: The CFTR assay achieves 92% or higher coverage of CFTR variants in diverse populations and provides improved pan-ethnic coverage of minority subgroups of the US populace. The assay can be completed within 5 hours from DNA sample to genotype, and performance data exceed acceptance criteria for analytic metrics. This assay panel content may help address gaps in ancestry-specific CFTR genotypes while providing a streamlined procedure with rapidly generated results.

Context.—: Transfusion medicine can be a challenging subject to teach to pathology residents while also ensuring that all the Accreditation Council for Graduate Medical Education's (ACGME's) milestones are met.

Objective.—: To explore how one major academic residency program has structured its transfusion medicine rotation.

Design.—: The residents on the pathology rotation have very defined roles for their day-to-day responsibilities. Many new resources have been developed during the past 3 years to improve the residents' educational experience on their transfusion medicine rotation. A daily patient list is used to direct the residents' educational and service responsibilities. They also have numerous resources to help with independent study and reading during their rotation.

Results.—: The implementation of several new resources has greatly improved the residents' educational experience and has improved the overall evaluation of the rotation by the residents. Many of the ACGME milestones can be met by the structure of this rotation.

Conclusions.—: With the proper structure and resources, transfusion medicine can be effectively taught to all pathology residents while also meeting the ACGME milestones requirements.

Context.—: Machine learning applications in the pathology clinical domain are emerging rapidly. As decision support systems continue to mature, laboratories will increasingly need guidance to evaluate their performance in clinical practice. Currently there are no formal guidelines to assist pathology laboratories in verification and/or validation of such systems. These recommendations are being proposed for the evaluation of machine learning systems in the clinical practice of pathology.

Objective.—: To propose recommendations for performance evaluation of in vitro diagnostic tests on patient samples that incorporate machine learning as part of the preanalytical, analytical, or postanalytical phases of the laboratory workflow. Topics described include considerations for machine learning model evaluation including risk assessment, predeployment requirements, data sourcing and curation, verification and validation, change control management, human-computer interaction, practitioner training, and competency evaluation.

Data sources.—: An expert panel performed a review of the literature, Clinical and Laboratory Standards Institute guidance, and laboratory and government regulatory frameworks.

Conclusions.—: Review of the literature and existing documents enabled the development of proposed recommendations. This white paper pertains to performance evaluation of machine learning systems intended to be implemented for clinical patient testing. Further studies with real-world clinical data are encouraged to support these proposed recommendations. Performance evaluation of machine learning models is critical to verification and/or validation of in vitro diagnostic tests using machine learning intended for clinical practice.

Context.—: The Nottingham Grading System (NGS) developed by Elston and Ellis is used to grade invasive breast cancer (IBC). Glandular (acinar)/tubule formation is a component of NGS.

Objective.—: To investigate the ability of pathologists to identify individual structures that should be classified as glandular (acinar)/tubule formation.

Design.—: A total of 58 hematoxylin-eosin photographic images of IBC with 1 structure circled were classified as tubules (41 cases) or nontubules (17 cases) by Professor Ellis. Images were sent as a PowerPoint (Microsoft) file to breast pathologists, who were provided with the World Health Organization definition of a tubule and asked to determine if a circled structure represented a tubule.

Results.—: Among 35 pathologists, the κ statistic for assessing agreement in evaluating the 58 images was 0.324 (95% CI, 0.314-0.335). The median concordance rate between a participating pathologist and Professor Ellis was 94.1% for evaluating 17 nontubule cases and 53.7% for 41 tubule cases. A total of 41% of the tubule cases were classified correctly by less than 50% of pathologists. Structures classified as tubules by Professor Ellis but often not recognized as tubules by pathologists included glands with complex architecture, mucinous carcinoma, and the "inverted tubule" pattern of micropapillary carcinoma. A total of 80% of participants reported that they did not have clarity on what represented a tubule.

Conclusions.—: We identified structures that should be included as tubules but that were not readily identified by pathologists. Greater concordance for identification of tubules might be obtained by providing more detailed images and descriptions of the types of structures included as tubules.

Context.—: Social media has become widely adopted by pathologists and other physicians for professional purposes. While engagement has likely increased over time, there remain few concrete data regarding attitudes toward its use.

Objective.—: To assess pathologists' use of and attitudes toward social media over time.

Design.—: We created a survey regarding personal and professional use of social media and circulated it via multiple channels in December 2017 and again in February 2022. Results of the 2 surveys were compared for statistically significant differences.

Results.—: The 2017 survey was completed by 97 participants, and the 2022 survey by 305 participants. Respondents were predominantly female and academics, included pathologists in all age categories and all time-in-practice length. In both surveys, Twitter (now X) was the most popular platform for professional use and Facebook was the most popular for personal use. Professional barriers to social media use remained consistent between the 2 surveys, including the amount of time required. Education was seen as the main benefit of social media use in both surveys, while other benefits such as networking and increasing professional visibility were endorsed significantly less often in the second survey. While the second survey received more than 3 times as many responses as the first, several aspects of social media use (mainly demographics) remained similar during the timeframe, while other aspects (such as usage and perceived values) decreased.

Conclusions.—: Pathologists continue to find social media valuable. Barriers remain, though overall pathologists of all ages and practice settings appear receptive to using social media to further educational and other opportunities.

Context.—: Carbapenem-resistant Enterobacterales are disseminated worldwide and associated with infections with high rates of morbidity and mortality. Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) is a useful tool for identification of pathogens directly from blood cultures in clinical microbiology laboratories. Furthermore, it has been applied for the detection of carbapenemase production, by evaluating carbapenem hydrolysis.

Objective.—: To determine meropenem hydrolysis to detect carbapenemase production directly from positive blood cultures, using logRQ to establish a quantitative measure of hydrolysis.

Design.—: We evaluated 100 Enterobacterales from positive blood cultures, with 81 carrying a carbapenemase gene (blaKPC, blaGES, blaNDM-1, blaIMP, blaVIM, and blaOXA-48-like), as determined by real-time multiplex polymerase chain reaction with high-resolution melting (HRM-qPCR). Bacterial proteins extracted from positive blood culture bottles were incubated in a meropenem solution (2-4 hours) followed by centrifugation for MALDI-TOF MS analysis. The intensity of peaks of the hydrolyzed and nonhydrolyzed forms were used to calculate the logRQ value.

Results.—: Overall, sensitivity was 86.8% and specificity, 89.5%. Of note, sensitivity varied depending on enzyme type. For blaKPC-positive isolates, sensitivity was 97.9%, while it reduced significantly for blaNDM-1 and blaOXA-48-like isolates: 62.5% (10 of 16) and 66.7% (6 of 9), respectively. Indeed, logRQ was higher in blaKPC-positive isolates (0.37-1.97) than in blaNDM-1 (-1.37 to 0.83) and blaOXA-48-like isolates (-1.08 to 1.79).

Conclusions.—: This is an inexpensive and rapid test to identify carbapenemase activity directly from blood culture bottles, which contributes to early adequate antimicrobial therapy and implementation of infection control measures.