Clinical chemistry and laboratory medicine最新文献

Objectives: Serum protein electrophoresis is one of the keystone investigations for screening for monoclonal proteins and assessing the serum protein profile. The newly released Helena V8ultra capillary electrophoresis system has been evaluated.

Methods: In total, 164 serum samples were assessed on the Helena V8 UltraCE system and compared with the Sebia Capillarys™ instrument. Abnormalities suggestive of monoclonal proteins were confirmed by immunosubtraction. Imprecision studies intervals were determined. Special attention was paid to complement C3 polymorphism.

Results: The imprecision of the Helena V8ultra was inferior or equal from 0.73 % (albumin) to 4.32 % (alpha₂ globulin). The mean bias of Helena V8ultraCE vs. Sebia Capillarys was about -0.225 % for albumin; -0.791 % for alpha₁ globulins; 1.353 % for alpha₂ globulins; -2.317 % for beta globulins; 0.066 % for gamma globulins. Among the 6 samples with monoclonal proteins confirmed by immunofixation, all were seen on both methods, with only 1 discordant result. The theoretical plate number (albumin fraction) was 2,371 ± 718 (Helena V8UltraCE) vs. 445 ± 115 (Sebia Capillarys 2). The high resolution also allows to distinguish Complement C3 phenotypes. CV values for the various fractions were low, ranging from 0.47 % (albumin) to 0.89 % (gamma globulins). IgG, IgA and IgM M-proteins could be detected in the electropherogram with an excellent sensitivity (<0.1 g/L).

Conclusions: Our evaluation confirms the good analytical performance of the Helena V8 analyzer as a suitable alternative to the Sebia Capillarys instrument. The high resolution allows detailed analysis of individual protein fractions which is an excellent basis for studying microheterogeneity.

Objectives: The performance of a novel urine particle analyzer, fluidlab 2 (Anvajo GmbH, Dresden, Germany), was evaluated against phase-contrast visual microscopy according to the most recent EFLM European Urinalysis Guideline.

Methods: The fluidlab 2 device combines digital holographic microscopy with neural network-based object detection for particle classification. Its compact benchtop design is suitable for bedside use, reducing turnaround times. The analytical performance (imprecision, linearity, LoQ) was evaluated according to the 2023 EFLM Urinalysis Guideline. Method comparison involved the analysis of 450 urine samples, assessing RBC, WBC, and SEC counts against visual microscopy using Passing-Bablok regression and Spearman's correlation. Bland-Altman plots were used to evaluate the agreement with clinical performance standards, while weighted Cohen's kappa was used to measure diagnostic agreement on an ordinal scale.

Results: By applying Dahlberg's procedure, a desirable relative coefficient of variation R(CV) ≤2.0 was obtained for RBC and WBC. Linearity of up to 7 × 10⁶/L and 6 × 10⁶/L was achieved. The estimated LoQ at CV=30 % reached 20 × 10⁶/L for RBC and 5 × 10⁶/L for WBC. Spearman's correlation coefficient against visual microscopy was 0.86, 0.92 and 0.94 for RBC, WBC and SEC, respectively. Agreement with visual microscopy (Cohen's weighted kappa) was 0.92 for RBC, 0.93 for WBC, 0.96 for SEC, 0.86 for casts, 0.82 for non-SEC, 0.33 for crystals and 0.51 for bacterial counts.

Conclusions: Fluidlab 2 provides desirable imprecision for RBC and WBC, and meets the criteria for linearity and LoQ. Cohen's weighted kappa coefficients show an optimal comparison to visual microscopy for RBC, WBC and SEC and a minimum comparison for casts and non-SEC. This evaluation demonstrated promising results for the use of the fluidlab 2 analyzer in a clinical setting to detect kidney-related diseases based on urine particle analysis.

Agentic artificial intelligence (AI) systems are distinguished by their ability to invoke multiple tools, compose command chains, and combine chain-of-thought reasoning with deep research to execute complex tasks and take actions. This represents a major evolution beyond machine learning and large language models (LLM), toward systems capable of planning, executing, and coordinating complex workflows. In contrast to traditional LLMs, which primarily focus on generating and classifying information, agentic AI introduces elements of autonomy, reasoning, and orchestration, while digital twins extend this concept to dynamic virtual representations of patients and laboratory processes, capable of continuous learning, simulation and adaptation. This transition has profound implications for laboratory medicine, a field characterized by high data complexity, multi-omics integration, and stringent operational demands. At the same time, laboratories face growing expectations regarding efficiency, resource stewardship, and value-based healthcare delivery. This article explores both the opportunities and limitations of agentic AI in laboratory medicine, highlighting its potential to move beyond static automation toward autonomous, outcome-driven innovation. It also examines the ethical, interpretability, and governance considerations that must accompany its implementation.

Objectives: The Kappa Index has proven its diagnostic value for multiple sclerosis (MS), while its prognostic potential remains to be fully explored. The objective of this study is thus to investigate the value of the Kappa Index at disease onset in predicting disease activity and high-efficacy therapy (HET) initiation.

Methods: We enrolled MS patients with available Kappa Index values at disease onset and a follow-up of at least two years. Primary outcome was the time to loss of NEDA3 (no evidence of disease activity-3) defined as the absence of relapses, MRI activity, and disability progression. Secondary outcome was the time to HET initiation.

Results: Of 120 enrolled patients (36 M, 84 F, mean age: 35 ± 11 years), NEDA3 loss occurred in 89 (74 %) by the end of the follow-up period. A total of 98 (82 %) initiated a moderate efficacy therapy (MET); of these, 34 (28 %) transitioned to a HET during follow-up. Kappa Index values above the maximally selected log-rank statistic-derived cut-off of 38 were independent risk factors for NEDA3 loss (HR 1.75, 95 % CI: 1.09-2.80, p=0.021) and HET initiation (3.25, 95 % CI: 1.54-6.87, p=0.002) and also independently predicted HET following MET failure (2.54, 95 % CI: 1.17-5.51, p=0.018).

Conclusions: Elevated Kappa Index values at diagnosis predict disease activity, MET failure and HET initiation and may be a valuable adjunctive tool in identifying patients in need of prompt HET initiation.

Objectives: Compare Elecsys (Roche) and Simoa (Quanterix) immunoassays for serum glial fibrillary acidic protein (GFAP) using our reference database and Z scores, and evaluate their prognostic value for progression independent of relapse activity (PIRA) in multiple sclerosis (MS).

Methods: Platform correlation was assessed in 612 samples from healthy controls (n=188; median [interquartile range, IQR] age 45.1 [36.4-61.7] years) and people with MS (n=424; 45.3 [35.2-53.9] years). Elecsys values were converted to Z scores via Passing-Bablok-derived regression and validated in fingolimod (n=414), and B-cell depleting therapy (BCDT; n=353) cohorts. Z scores and hazard ratios (HRs) for time-to-PIRA were compared using Cox regression.

Results: GFAP_Simoa and GFAP_Elecsys measurements were correlated (r=0.94), with Elecsys values ∼54 % lower (GFAP_Elecsys, ng/L=2.847 [95 % confidence interval, CI: 1.335 - 4.98] + 0.457 [0.434 - 0.478] * GFAP_Simoa, ng/L). In univariable Cox models, GFAP_Simoa and GFAP_Elecsys Z scores were associated with time-to-PIRA in both validation cohorts. In multivariable Cox models, higher GFAP_Simoa Z scores were associated with shorter time-to-PIRA in fingolimod cohort (HR: 1.27 [95 % CI 1.08 - 1.50], p=0.0031) and trended toward significance in BCDT (1.18 [0.99 - 1.41, p=0.0693). In contrast, GFAP_Elecsys Z scores were associated with time-to-PIRA in both cohorts (fingolimod: 1.27 [1.09 - 1.48], p=0.0023; BCDT: (1.19 [1.00 - 1.40], p=0.0487).

Conclusions: Serum GFAP measured by Elecsys shows a comparable association with time-to-PIRA as Simoa, and GFAP_Simoa Z scores can be successfully bridged to GFAP_Elecsys Z scores, supporting Elecsys`s potential for clinical implementation.

Objectives: Free triiodothyronine (FT3) and free thyroxine (FT4) are important diagnostic markers for assessing thyroid function. However, their accurate quantification remains challenging due to low serum concentrations. Significant variability exists among current assay methods for measuring FT3 and FT4. This study aims to establish a candidate Reference Measurement Procedure (cRMP) for simultaneous quantification of serum FT3 and FT4 based on isotope-dilution liquid chromatography-tandem mass spectrometry (ID-LC-MS/MS) technology.

Methods: A convenient and reusable equilibrium dialysis (ED) device was utilized to separate free thyroid hormones from their protein-bound counterparts in serum. Key dialysis parameters, including temperature, pH, membrane type, and duration, were optimized to ensure consistent and reliable performance. The dialysate containing FT3 and FT4 was directly quantified by ID-LC-MS/MS. The method underwent systematic validation, comparative analysis with existing assays, and a comprehensive uncertainty assessment.

Results: The developed cRMP demonstrated limits of quantification (LoQ) of 1.54 pmol/L for FT3 and 3.22 pmol/L for FT4, and with an imprecision of less than 3 %. No interference from endogenous analogs was observed, and the method showed good consistency in interlaboratory comparison. In contrast, chemiluminescent immunoassay results exhibited poor agreement with and the cRMP.

Conclusions: This study developed a highly precise, accurate, specific, and sensitive ID-LC-MS/MS-based cRMP for the simultaneous measurement of FT3 and FT4 in human serum. This method provides a reliable tool for standardizing routine thyroid function tests.

Objectives: Prognostication in patients with severe acute brain injury (SABI) in the Intensive Care Unit (ICU) is complex, often requiring early decisions on life-sustaining therapy. Serum biomarkers such as neurofilament light (NFL), glial fibrillary acidic protein (GFAP) and neuron specific enolase (NSE) may aid in predicting outcomes. This study investigated the association between biomarker concentrations and hospital mortality in ICU patients with SABI.

Methods: In this prospective multi-center study, 103 adult ICU patients with SABI were enrolled within 24 h of injury. Serum concentrations (NFL, GFAP, NSE) were measured at 24 h, 72 h, 7 days, and 14 days post-injury. Biomarker concentrations between hospital survivors and non-survivors were compared using a linear mixed-effects model and ROC curves.

Results: GFAP concentrations peaked on day 1, while NSE and NFL concentrations peaked on day 14. Higher concentrations were observed in non-survivors for all biomarkers on day 3, a 1,4 fold increase for NSE (p=0.034) and approximately 2-fold increase for GFAP (p=0.041) and NFL (p=0.018). Distinct temporal patterns were observed for all biomarkers, although biomarker trajectories did not significantly differ between survivors and non-survivors. Prognostic accuracy was moderate when biomarkers were combined with established mortality predictors (AUC 0.749) on day 3.

Conclusions: NSE, GFAP and NFL show distinct temporal trajectories in SABI patients. Concentrations were higher in non-survivors during the early phase of admission. While individual biomarkers showed limited prognostic accuracy, in combination with other predictors mortality prediction improved. Serum biomarkers may still be valuable in a multimodal prognostication framework for SABI patients.