Advances in laboratory medicine最新文献

Introduction: Data science is an umbrella term encompassing a set of tools and processes that make it possible to extract new information from structured or unstructured databases. This scientific discipline is gaining relevance in healthcare. In the clinical laboratory, the multiple applications of data science include the development of algorithms for obtaining population-based reference intervals or biological variation (BV) estimates. These algorithms contribute to overcoming the drawbacks of traditional or direct methods.

Content: A review was performed of the state-of-the-art in algorithm-based methods for obtaining BV estimates using Real-World Data (RWD) in the field of data science.

Summary: A description is provided of the structure of the algorithms currently available for obtaining BV estimates based on the scientific evidence available. An overview is provided of the advantages and drawbacks of direct methods.

Outlook: The use of RWD to obtain BV estimates is a novel discipline with a considerable potential for improving our understanding of BV.

Introduction: Circulating cell-free DNA (cfDNA) consists of extracellular DNA fragments that circulate in the bloodstream and derived from apoptotic cells such as hematopoietic cells or placental trophoblast cells during pregnancy.

Contents: cfDNA screening has been included in prenatal screening programs for the detection of chromosomal abnormalities. Unlike other invasive techniques, such as amniocentesis or chorionic villus sampling, cfDNA screening only requires a maternal plasma test. The use of advanced technologies for cfDNA testing, including DNA sequencing and SNP arrays, enables the detection of pregnancies at risk for trisomy 21, 18 or 13.

Summary: This test has demonstrated a high accuracy and reliability, with detection rates exceeding 99 % for trisomy 21, and a very low rate of false-positive and false-negative results. In some countries, cfDNA screening has already been integrated in combined or universal prenatal screening programs.

Outlook: As new technologies emerge and become widely available, more accurate prenatal tests will be developed for other genetic abnormalities.

Objectives: Lipemia poses a significant preanalytical problem for complete blood count (CBC) measurement due to limited and non-standardized methods for recognition and removal. We aimed to verify the optical hemoglobin (Hb-O) measurements on the Sysmex XN-1000 hematology analyzer (HA) as a possible reliable method for managing lipemic CBC samples.

Methods: Ninety CBC samples with varying Hb concentrations were gradually spiked with a lipid emulsion. Measurements were repeated and Hb-O concentrations were recorded. Spiked CBC samples were centrifuged (400 g/10 min). Plasma was carefully removed, and Hb concentration was measured. The values obtained from the lipemic samples were adjusted according to the measurements in the plasma. The removed plasma was substituted with the analyzer's diluent, and measurements were repeated. Triglyceride concentrations were measured in lipemic plasma samples.

Results: Hb-O showed statistically insignificant and acceptable bias compared to the initial Hb measurement according to the strictest acceptability criteria (-0.4 %, 95 % CI: -1.2-0.3, p=0.2447). The observed bias did not correlate with the degree of lipemia (rho=-0.072, 95 % CI: -0.295 to 0.157, p=0.537). Hemoglobin measured in samples with lipemic plasma replaced by analyzer diluent exhibited minimal, albeit statistically significant, bias (-1.1 %, 95 % CI: -2.0 to (-0.1), p=0.025). The observed bias negatively correlated with the degree of lipemia (rho=-0.369, 95 % CI: -0.550 to (-0.155), p=0.001). The highest unacceptable bias was found in the recalculated hemoglobin values based on the measured plasma hemoglobin (-3.5 %, 95 % CI: -4.1 to (-2.9), p<0.0001).

Conclusions: Hb-O measurement is the most reliable measure of lipemia removal in CBC samples on the Sysmex XN-1000 HA.

Objectives: To evaluate seven bioinformatics platforms for automated AI-based genomic variant prioritization and classification.

Methods: An evaluation was performed of 24 genetic variants that explained the phenotype of 20 patients. FASTQ files were simultaneously uploaded on the following bioinformatics platforms: Emedgene, eVai, Varsome Clinical, CentoCloud, QIAGEN Clinical Insight (QCI) Interpret, SeqOne and Franklin. Automated variant prioritization and classification was performed using patient phenotypes. Phenotypes were entered onto the different platforms using HPO terms. The classification of reference was established based on the criteria of the American College of Medical Genetics and Genomics (ACMG) and the Association of Molecular Pathology and ACMG/ClinGen guidelines.

Results: SeqOne demonstrated the highest performance in variant prioritization and ranked 19 of 24 variants in the Top 1; four in the Top 5, and one in the Top 15, followed by CentoCloud and Franklin. QCI Interpret did not prioritize six variants and failed to detect one. Emedgene did not prioritize one and failed to detect one. Finally, Varsome Clinical did not prioritize four variants. Franklin classified correctly 75 % of variants, followed by Varsome Clinical (67 %) and QCI Interpret (63 %).

Conclusions: SeqOne, CentoCloud, and Franklin had the highest performance in automated variant prioritization, as they prioritized all variants. In relation to automated classification, Franklin showed a higher concordance with the reference and a lower number of discordances with clinical implications. In conclusion, Franklin emerges as the platform with the best overall performance. Anyway, further studies are needed to confirm these results.

Objectives: To prospectively examine the ability of some glycemic variability metrics from continuous glucose monitoring (CGM) to predict the development of diabetes in a non-diabetic population.

Methods: A total of 497 non-diabetic patients from the AEGIS study were included. Participants used a CGM system (iPro2^®) over a six-day period. The following parameters were analyzed: standard deviation (SD), coefficient of variation (CV) and mean amplitude of glucose excursion (MAGE). Six-years follow-up was performed. ROC curves were constructed to determine the predictive value of glycemic variability metrics. Sensitivity and specificity were calculated.

Results: Of the 497 participants, 16 women (4.9 %) and 9 men (5.2 %) developed diabetes. Initial HbA_1c and fasting glucose levels were significantly higher in the participants who ultimately developed diabetes. Glycemic variability metrics were also significantly higher in these subjects (SD: 18 vs. 13 mg/dL; CV: 17 vs. 14 %; MAGE: 36 vs. 27 mg/dL; p<0.001 in all cases). SD showed the highest AUC (0.81), with a sensitivity of 80 % and a specificity of 72 % for a cut-off of 14.9 mg/dL. AUCs were higher in men for all metrics.

Conclusions: The metrics obtained by MCG, especially SD, are effective predictors of progression to type 2 diabetes in a non-diabetic population. These findings suggest that glycemic variability is useful for the early identification of subjects at a higher risk of developing diabetes.