Better Performance of Flow-FISH in Comparison to qPCR As a Diagnostic Test for Telomere Diseases

Excessive telomere erosion is the molecular etiology of a group of disorders (dyskeratosis congenita, aplastic anemia, idiopathic pulmonary fibrosis) collectively called telomeropathies. Telomere length measurement is an essential diagnostic test for these diseases. The most commonly used methods are terminal restriction fragment (TRF) analysis by Southern blotting (the gold-standard method), flow cytometry combined with fluorescence in situ hybridization (flow-FISH), and quantitative PCR (qPCR). Although the clinical use of these methods has been reported, their utility and characteristics have not been widely compared. Measurement techniques and coefficients of variations often differ among diagnostic services. Here, we directly compared the accuracy, reproducibility, sensitivity, and specificity of flow-FISH and qPCR in comparison to TRF to measure peripheral blood leukocyte’s telomere length in healthy individuals and patients with telomeropathies. TRF analyses and flow-FISH showed good correlation in the analysis of samples from healthy subjects (R²=0.60; p<0.0001) and patients (R²=0.51; p<0.0001). Bland-Altman analyses also displayed a very good agreement between these methods for both healthy individuals (bias±SD = 0.17±1.03; limits of agreement ranging from 2.24 to -1.88) and patients (bias±SD = 0.0±1.21; limits of agreement ranging from 2.41 to -2.41). In contrast, the comparison between TRF and qPCR yielded modest correlation for the analysis of samples of healthy individuals (R²=0.35; p<0.0001) and low correlation for patients (R²=0.20; p=0.001). Bland-Altman analysis indicated poor agreement between the two methods for both patients and controls. The differences averages were very different from zero and standard deviation was wide. For patients, the bias±SD was 0.78±1.34 with limits of agreement ranging from 3.47 to -1.90, and for controls, the bias±SD was 1.15±1.49 with limits of agreement ranging from 4.14 to -1.84. Finally, qPCR and flow-FISH also modestly correlated in the analysis of healthy individual samples (R²=0.33; p<0.0001) and did not correlate in the comparison of patients’ samples (R²=0.1, p=0.08). Bland-Altman analysis corroborate this finding. For controls, the bias±SD were very similar to the one found by comparison between qPCR and TRF analysis (-0.6±1.27; limits of agreement ranging from 1.94 to -3.16). For patients, bias ± SD were -1.15 ± 1.65 with limits of agreement ranging from 2.15 to -4.45, which evidenced a poor agreement between flow-FISH and qPCR in these samples. Intra-assay coefficient of variation (CV) was 10.8±7.1% for flow-FISH and 9.5±7.4% for qPCR (p=0.35). The inter-assay CV was lower for flow-FISH (9.6±7.6%) in comparison to qPCR (16±19.5%; p=0.02). Flow-FISH and qPCR were sensitive (both 100%) and specific (93% and 89%, respectively) to distinguish very short telomeres. However, qPCR sensitivity (40%) and specificity (63%) to detect telomere length below tenth percentile were lower in comparison to flow-FISH (80% sensitivity and 85% specificity). Taken together, these findings indicate that, in the clinical setting, flow-FISH is more accurate and reproducible in the measurement of human leukocyte’s telomere length in comparison to qPCR. Quantitative PCR exhibited low accuracy in the analysis of samples of patients with short telomeres. In conclusion, flow-FISH appears to be a more appropriate method for diagnostic purposes. Studies that compare methodologies are helpful in the selection of standard methods and to narrow the differences among laboratories.