Assessment of a large language model's utility in helping pathology professionals answer general knowledge pathology questions.

OBJECTIVES: To assess the utility and performance of the large language model ChatGPT 4.0 regarding accuracy, completeness, and its potential as a time-saving tool for pathologists and laboratory directors. METHODS: A deidentified database of questions previously sent to pathology residents from health care providers was used as a source of general knowledge-type pathology questions. These questions were submitted to the large language model and the responses graded by subject matter experts in different pathology subspecialties. The grading criteria assessed accuracy, completeness, and the potential time savings for helping the pathologist craft the response. RESULTS: Overall, respondents thought that most of the answers would take less than 5 minutes of additional work to be used (85%). Accuracy and completeness for the 61 questions was high, with 98% of responses being at least "completely or mostly accurate" and 82% of responses "containing all relevant information." Of the respondents, 97% stated that the response would have "zero or near-zero potential for medical harm," and all thought the tool had potential to save time in constructing answers to health care providers' queries. Performance was similar in both Anatomic Pathology (AP) and Clinical Pathology (CP), with the only exception being some relevant information was excluded in 46% of AP answers vs only 10% in CP (P < .01). CONCLUSIONS: ChatGPT version 4.0 gave responses that were predominantly accurate and complete for general knowledge-type pathology questions. With further research and when reviewed by a pathologist or laboratorian, this could facilitate its use as a pathologist's aid in answering questions from health care providers.

Hook Effect in Semiquantitative SARS-CoV-2 Antispike Total Antibody Assay.

OBJECTIVES: The hook effect is a common preanalytical error that results in falsely decreased analyte concentrations in immunoassays. We present here an example in a semiquantitative SARS-CoV-2 antispike total antibody assay and report the incidence of this error at our institution. METHODS: All specimens with initial results within the reportable range of the assay were diluted. Results with higher results upon dilution were determined to have the hook effect. In a subset of specimens, these results were also confirmed as elevated on an alternative SARS-CoV-2 antibody assay. RESULTS: Over 1 month, 12 (9.1%) of 132 results were within the analytical measuring range of the assay. Of these, 11 showed the hook effect and required dilution to obtain accurate results. These represented 8.3% of our total testing volume. CONCLUSIONS: The hook effect was detected in a semiquantitative SARS-CoV-2 antispike total antibody assay at a high incidence. This error results in observed concentrations much lower than is accurate. Laboratories should be aware of this issue and consider manually diluting specimens within the reportable range of the assay to detect this issue.

Elevated Lactate Dehydrogenase Concentrations in Plasma Compared to Serum.

OBJECTIVE: To evaluate the difference in lactate dehydrogenase (LDH) concentrations in plasma vs serum specimens in our patient population. MATERIALS AND METHODS: We measured LDH in 110 paired plasma and serum specimens over a 2-week period. Hemolytic indices were performed on each specimen. These paired specimens were drawn in a single setting and stored under the same conditions. For the last 14 paired specimens, cell counts were performed on the plasma/serum. RESULTS: Plasma LDH was on average 22% higher than serum LDH. There was no difference in the hemolytic indices between the plasma and the serum specimens. In the last 14 specimens, cell counts revealed increased platelets in the plasma specimens compared to the serum specimens. CONCLUSION: We propose switching back to using serum for LDH testing because there was unpredictable elevation in plasma LDH concentrations. These elevations in LDH levels may be linked to the platelets present in plasma and that may lyse or become activated with storage at refrigerated temperature.

Target specific serologic analysis of COVID-19 convalescent plasma.

This study compared the performance of four serology assays for Coronavirus Disease 2019 (COVID-19) and investigated whether COVID-19 disease history correlates with assay performance. Samples were tested at Northshore using the Elecsys Anti-SARS-CoV-2 (Roche Diagnostics), Access SARS-CoV-2 IgG anti-RBD (Beckman Coulter), and LIAISON SARS-CoV-2 S1/S2 IgG (DiaSorin) as well as at Genalyte using Maverick Multi-Antigen Serology Panel. The study included one hundred clinical samples collected before December 2019 and ninety-seven samples collected from convalescent plasma donors originally diagnosed with COVID-19 by PCR. COVID-19 disease history was self-reported by the plasma donors. There was no difference in specificity between the assays tested. Clinical sensitivity of these four tests was 98% (Genalyte), 96% (Roche), 92% (DiaSorin), and 87% (Beckman). The only statistically significant differences in clinical sensitivity was between the Beckman assay and both Genalyte and Roche assays. Convalescent plasma donor characteristics and disease symptoms did not correlate with false negative results from the Beckman and DiaSorin assays. All four tests showed high specificity (100%) and varying sensitivities (89-98%). No correlations between disease history and serology results were observed. The Genalyte Multiplex assay showed as good or better sensitivity to three other previously validated assays with FDA Emergency Use Authorizations.

Detection of Falsely Elevated Point-of-Care Potassium Results Due to Hemolysis Using Predictive Analytics.

OBJECTIVES: Preanalytical factors, such as hemolysis, affect many components of a test panel. Machine learning can be used to recognize these patterns, alerting clinicians and laboratories to potentially erroneous results. In particular, machine learning might identify which cases of elevated potassium from a point-of-care (POC) basic metabolic panel are likely erroneous. METHODS: Plasma potassium concentrations were compared between POC and core laboratory basic metabolic panels to identify falsely elevated POC results. A logistic regression model was created using these labels and the other analytes on the POC panel. RESULTS: This model has high predictive power in classifying POC potassium as falsely elevated or not (area under the curve of 0.995 when applied to the test data set). A rule-in and rule-out approach further improves the model's applicability with a positive predictive value of around 90% and a negative predictive value near 100%. CONCLUSIONS: Machine learning has the potential to detect laboratory errors based on the recognition of patterns in commonly requested multianalyte panels. This could be used to alert providers at the POC that a result is suspicious or used to monitor the quality of POC results.

Clinical Impact of Rapid Point-of-Care PCR Influenza Testing in an Urgent Care Setting: a Single-Center Study.

Seasonal influenza virus causes significant morbidity and mortality each year. Point-of-care (POC) testing using rapid influenza diagnostic tests (RIDTs), immunoassays that detect viral antigens, are often used for diagnosis by physician offices and urgent care centers. These tests are rapid but lack sensitivity, which is estimated to be 50 to 70%. Testing by PCR is highly sensitive and specific, but historically these assays have been performed in centralized clinical laboratories necessitating specimen transport and increasing the time to result. Recently, Clinical Laboratory Improvement Amendments (CLIA)-waived, POC PCR influenza assays have been developed with >95% sensitivity and specificity compared to centralized PCR assays. To determine the clinical impact of a POC PCR test for influenza, we compared antimicrobial prescribing patterns of one urgent care location using the Cobas LIAT Influenza A/B assay (LIAT assay; Roche Diagnostics, Indianapolis, IN) to other urgent care centers in our health system using traditional RIDT, with negative specimens being reflexed to PCR. Antiviral prescribing was lower in patients with a negative LIAT PCR result (2.3%) than in patients with a negative RIDT result (25.3%; P < 0.005). Antivirals were prescribed more often in patients that tested positive by LIAT PCR (82.4%) than in those testing positive by either RIDT or reflex PCR (69.9%; P < 0.05). Antibacterial prescriptions for patients testing negative by LIAT PCR were higher (44.5%) than for those testing negative by RIDT (37.7%), although the difference was not statistically significant. In conclusion, having results from a PCR POC test during the clinic visit improved antiviral prescribing practices compared to having rapid results from an RIDT.

Variable Potassium Concentrations: Which Is Right and Which Is Wrong?

Reverse pseudohyperkalemia is a term used to describe in vitro, falsely elevated potassium concentrations in plasma specimens that occur in association with extreme leukocytosis and are commonly associated with hematologic malignant neoplasms. Tumor lysis syndrome is an in vivo lysis of tumor cells that leads to elevated levels of potassium, uric acid, phosphate, and lactate dehydrogenase, as well as decreased calcium concentrations. Herein, we report a case of a 66-year-old Caucasian man with stage IV mantle-cell lymphoma who has elevated levels of potassium, uric acid, and phosphorus, as well as a white blood cell (WBC) count greater than 100,000 cells per mm3. The patient initially was diagnosed as having tumor lysis syndrome. His subsequent potassium concentrations in whole blood remained elevated even after hemodialysis; however, his serum potassium concentrations were decreased. The patient then was diagnosed accurately as having reverse pseudohyperkalemia, and accurate potassium measurements were obtained via serum specimens.

Structural basis of ubiquitin recognition by translesion synthesis DNA polymerase ι.

Cells have evolved mutagenic bypass mechanisms that prevent stalling of the replication machinery at DNA lesions. This process, translesion DNA synthesis (TLS), involves switching from high-fidelity DNA polymerases to specialized DNA polymerases that replicate through a variety of DNA lesions. In eukaryotes, polymerase switching during TLS is regulated by the DNA damage-triggered monoubiquitylation of PCNA. How the switch operates is unknown, but all TLS polymerases of the so-called Y-family possess PCNA and ubiquitin-binding domains that are important for their function. To gain insight into the structural mechanisms underlying the regulation of TLS by ubiquitylation, we have probed the interaction of ubiquitin with a conserved ubiquitin-binding motif (UBM2) of Y-family polymerase Polι. Using NMR spectroscopy, we have determined the structure of a complex of human Polι UBM2 and ubiquitin, revealing a novel ubiquitin recognition fold consisting of two α-helices separated by a central trans-proline residue conserved in all UBMs. We show that, different from the majority of ubiquitin complexes characterized to date, ubiquitin residue Ile44 only plays a modest role in the association of ubiquitin with Polι UBM2. Instead, binding of UBM2 is centered on the recognition of Leu8 in ubiquitin, which is essential for the interaction.

Molecular basis for the association of human E4B U box ubiquitin ligase with E2-conjugating enzymes UbcH5c and Ubc4.

Human E4B, also called UFD2a, is a U box-containing protein that functions as an E3 ubiquitin ligase and an E4 polyubiquitin chain elongation factor. E4B is thought to participate in the proteasomal degradation of misfolded or damaged proteins through association with chaperones. The U box domain is an anchor site for E2 ubiquitin-conjugating enzymes, but little is known of the binding mechanism. Using X-ray crystallography and NMR spectroscopy, we determined the structures of E4B U box free and bound to UbcH5c and Ubc4 E2s. Whereas previously characterized U box domains are homodimeric, we show that E4B U box is a monomer stabilized by a network of hydrogen bonds identified from scalar coupling measurements. These structural studies, complemented by calorimetry- and NMR-based binding assays, suggest an allosteric regulation of UbcH5c and Ubc4 by E4B U box and provide a molecular basis to understand how the ubiquitylation machinery involving E4B assembles.

Melting, freezing, sublimation, and phase coexistence in sodium chloride nanocrystals.

Calorimetry measurements, performed by multicollision induced dissociation, have been used to probe the melting of a number of (NaCl)nNa+ clusters with n=22-37. The clusters anneal at 225-325 K and melt at 750-850 K. (NaCl)22Na+ and (NaCl)37Na+, which can adopt geometries that are perfect fragments of the bulk lattice melt at around 850 K. The other clusters, which (except for n=31) must have defects, melt at temperatures which are up to 100 K lower than the perfect nanocrystals. The internal energy distributions become bimodal near the melting temperature. This is the signature of slow dynamic phase coexistence where clusters spontaneously jump back and forth between the solid and liquid states with an average period that is longer than required for thermal equilibration. The jump frequency must be between 10(4) and 10(7) s(-1) for the bimodal distribution to be observable in our experiments. The (NaCl)nNa+ clusters can dissociate by an unusual thermally activated process where melting and freezing raise the internal energy to generate hot solid clusters that can sublime before they cool to the ambient temperature.

Gallium cluster "magic melters".

Calorimetry measurements (using a method based on multicollision induced dissociation) have been performed for unsupported gallium clusters, Gan+ (n = 30-50 and 55). Melting transitions have been identified from spikes in the heat capacities recorded as a function of temperature. There are enormous fluctuations in the melting temperatures and the heats of fusion with cluster size. Clusters with n = 31, 33, 37, and 45-47 are "magic melters" with particularly well-defined melting transitions. There is a strong correlation between the heats of fusion, entropies of fusion, and the stabilities of the clusters. However, these quantities are not strongly correlated with the melting temperatures.

Hot and solid gallium clusters: too small to melt.

A novel multicollision induced dissociation scheme is employed to determine the energy content for mass-selected gallium cluster ions as a function of their temperature. Measurements were performed for Ga(+)(n) (n=17 39, and 40) over a 90-720 K temperature range. For Ga+39 and Ga+40 a broad maximum in the heat capacity-a signature of a melting transition for a small cluster-occurs at around 550 K. Thus small gallium clusters melt at substantially above the 302.9 K melting point of bulk gallium, in conflict with expectations that they will remain liquid to below 150 K. No melting transition is observed for Ga+17.