Standardized diagram improves diagnostic accuracy and variability in the identification of testicular cells prepared for testicular sperm extraction and intracytoplasmic sperm injection

Objective To examine the accuracy and variability of identification of human germ cells and validate the previously reported diagram referable to identifying human testicular cells, which was made to improve the identification. Methods Eighty-seven testicular cells obtained from azoospermic patients were stained with MitoTracker, and observed under phase contrast, fluorescent, and differential interference microscopy. The recorded image and movie data of phase contrast microscopy were assessed by 10 reviewers comprising embryologists and reproductive physicians 2 times, once without the diagram and 1 year later with use of it. True cell type identifications were determined as referenced by morphologic characteristics and MitoTracker staining. Variability between reviewers was assessed using multirater κ statistics, and changes of the concordance rates to the reference were examined. Results Multirater κ coefficients changed from 0.14 to 0.49 overall, from 0.10 to 0.34 for sperm-like cells, and from 0.044 to 0.46 in round-shaped cells before and after using the diagram, which represents a change from fair to substantial agreement overall for round-shaped cells and to moderate agreement for sperm-like cells. The concordance rates to the reference before and after the use of the diagram also significantly improved from 28.4% to 59.1% overall, from 38.9% to 54.6% for sperm-like cells, and from 19.4% to 59.1% for round-shaped cells, respectively. Conclusion Identification of human germ cells by embryologists and reproductive physicians was not uniform or satisfactory. However, the diagram significantly improved identification such that it may be useful as an efficient checklist for the identification of germ cells.