Clinical chemistry最新文献

Background: Digital polymerase chain reaction (dPCR) is widely recognized for its high sensitivity in detecting low-frequency variants; however, conventional 2-color systems have limited multiplex capacity. Expanding this capability is essential for simultaneous detection of multiple driver mutations in cancer-related genes. KRAS and GNAS are key driver genes in the early development of pancreatic cancer and its precursor lesions, and mutations in these genes are often present at low abundance in clinical samples.

Methods: Two 6-color dPCR assays were developed using a droplet-based platform. PlexScreen-dPCR is a multicolored drop-off assay designed to screen for mutations in KRAS codons 12/13 and 61 and GNAS codon 201, without specifying individual variants. PlexID-dPCR employs variant-specific probes to distinguish among 14 predefined KRAS and GNAS mutations in a single reaction. The assays were validated using synthetic DNA, cell lines, 23 tissue samples, and 12 duodenal fluid samples. Customized primer/probe sets with 6 fluorophores were employed in a 6-color droplet dPCR system, and the limits of detection (LOD) were determined.

Results: PlexScreen-dPCR, applied in contrived samples, demonstrated LODs as low as 0.03% to 0.06%, enabling high-sensitivity detection of low-abundance mutations. PlexID-dPCR accurately identified all 14 variants in a single well. Both assays showed complete concordance with conventional methods, exhibiting a strong correlation for variant allele frequency quantification.

Conclusions: These 6-color dPCR assays offer scalable solutions for improved throughput detection of KRAS and GNAS mutations. Their compatibility with commercially available platforms and streamlined workflow support their integration into clinical practice. Further optimization can enhance cluster interpretation in high-plex settings and facilitate expansion toward broader genomic targets.

Background: Urinalysis is a standard clinical test that includes the microscopic examination of urinary sediment to identify formed elements. Manual evaluation by laboratory technicians is time-intensive and subject to human error. Automated analysis using digital microscopy images presents a potential alternative. This study evaluates the integration of a deep learning approach to automatically classify urinary sediment images in the clinical laboratory, including independent prospective validation of its performance.

Methods: An annotated data set comprising 13 classes of urinary sediment elements was created from a database of Sysmex UD-10 digital microscope images. An EfficientNet-based model was trained and tested across three experimental scenarios to evaluate the effects of data collection strategies on performance. Uncertainty calibration was examined. The model's robustness and interpretability were examined using gradient-weighted class activation mapping (Grad-CAM) to visualize influential image regions and t-distributed stochastic neighbor embedding (t-SNE) to analyze learned feature embeddings. Lastly, a graphical user interface was developed for a prospective evaluation in the laboratory.

Results: The model achieved approximately 97% overall accuracy on the test set. Experiments revealed sensitivity to data set variability, suggesting that performance may improve by integrating additional training examples. Confidence scores aligned with accuracy, and interpretability analyses showed that the model focused on relevant image regions and learned embeddings demonstrated clear class separation. In the prospective evaluation, top 1 and top 3 accuracies decreased to approximately 78% and 92%, respectively.

Conclusions: Our results indicate that a lightweight deep learning model can achieve high performance in classifying urine particles. Analysis of discrepancies between retrospective and prospective evaluations provides important insights toward reliable clinical application.

Background: Neurofilament light chain (Nf-L) is a key early biomarker for axonal damage and neurodegeneration, increasingly used in clinical practice for diagnosis, prognosis, and treatment monitoring. To ensure reliable clinical implementation, standardized measurement procedures and calibrators traceable to the International System of Units (SI) are needed. Although a few mass-spectrometry methods for Nf-L quantification in cerebrospinal fluid CSF) and plasma have been described, none currently use properly defined SI-traceable calibrators and existing harmonization efforts rely solely on immunoassays. This study presents a validated immunoprecipitation (IP)-LC-MS/MS assay using an SI-traceable calibrator and compares it with Lumipulse and Simoa platforms to assess agreement and bias.

Methods: A new IP-LC-MS/MS method was developed based on SI-traceable calibrator quantification and analytically validated using International Council for Harmonisation (ICH) guidelines. Analysis of 69 CSF samples by MS assay and Lumipulse (Fujirebio®) was performed as well as head-to-head comparisons between MS, Simoa (Quanterix®) and Lumipulse assays on 12 CSF pools.

Results: The method relying on 3 peptides was validated analytically. Significant results (P < 0.05) were obtained between amyloid positive to negative group when using the LC-MS/MS assay. MS and Lumipulse results were correlated. Head-to-head comparison of the 3 methods showed great correlation (r2 > 0.98) but systematic bias was identified between all techniques.

Conclusion: A new IP-LC-MS/MS method using a SI-traceable calibrator was developed, and analytically and clinically validated. Comparison between available immunoassays resulted in great correlation but biases were identified reinforcing the need of standardization for Nf-L measurement.