Fraunhofer-Gesellschaft

Publica

Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

A comprehensive way to rate sputum quality in clinical trials

 
: Pedersen, Frauke; Zissler, U.; Watz, Henrik; Rabe, Klaus F.; Hohlfeld, Jens M.; Holz, Olaf

American Journal of Respiratory and Critical Care Medicine 197 (2018), Art. A4769
ISSN: 1073-449X
ISSN: 0003-0805
ISSN: 1535-4970
American Thoracic Society (ATS International Conference) <2018, San Diego/Calif.>
English
Abstract
Fraunhofer ITEM ()

Abstract
Introduction: The analysis of induced sputum is widely used in clinical trials to assess airway inflammation. Quality of sputum cell samples is variable and currently defined by the level of contamination with squamous cells (SQ%). This definition does not consider the quality of relevant sputum cells. Here we suggest a novel quality score and evaluated how it is related to differential cell count inter-observer variability and agreement. Methods: Thirty sputum cytospin samples, with a broad range of quality, from three DZL sites were analyzed by nine evaluators using standard techniques. Slide quality was rated by all evaluators on a 5-point scale (0, 0.5, 1, 1.5, 2; low-high). The slide quality score included a rating of cell morphology, level of cellular debris and SQ%. Inter-evaluator variability (SD), evaluator accuracy and intra class-correlation coefficients (ICC) between evaluator and overall mean cell counts (as reference) were computed. To assess the relationship between these parameters and slide quality the dataset was split into three quality levels based on the mean slide score (level C: <0.75; level B: ≥0.75,<1.25; level A: ≥1.25). Results: The overall mean (range) slide quality score was 0.9 (0.0; 1.6), with a mean variability across evaluators of 0.3. Mean SQ% was 18.5 (1; 85)%. For 7 of 30 slides individual evaluator quality scores were 0, resulting in 1 to 5 missing differential cell count results. The correlation between slide quality and SQ% was significant for each evaluator, but with a wide range of r-values (-0.39; -0.69). The 17 slides from quality level A and B had a maximum SQ% of 22%, while 6 slides from the low quality level C had below 22% squamous cells. The variability (SD) for alveolar macrophages (AM) and neutrophil granulocytes (NG) counts across evaluators was significantly decreased with increasing slide quality (ANOVA: p<0.02; p<0.01). Slide quality had only a minor effect on the accuracy of AM counts, but there was a trend for an increased accuracy for NG with increasing quality score. The ICC for AM and NG was always >0.9 and increased for all evaluators, if the slides from the low quality level C were excluded. Conclusion We propose to include sputum leukocyte cell integrity, the amount of cellular debris and SQ% into a slide quality score. Excluding samples based on this score reduced differential cell count inter-evaluator variability and improved the agreement of evaluators with the overall mean differential cell count.

: http://publica.fraunhofer.de/documents/N-530967.html