In the treatment of male infertility by intra-cytoplasmic injection of spermatozoa (ICSI) extracted from testicular tissue (TESE), the high incidence of negative TESE outcome calls for non-invasive prognostic methods. Literature suggests that seminal haploid germ cell detection could be one. For this purpose, a multi-parametric stringent flow cytometric method was applied to 50 TESE patients for the quantification of ejaculated germ cells. Cells from 50 ejaculates were identified and quantified as spermatozoa (HC, highly condensed), round spermatids (1N), primary spermatocytes (SPC) (4N) or diploid cells (2N, including somatic and non-testicular cells) by their DNA and mitochondria staining and laser scatter characteristics, and compared with testicular biopsy histology and TESE outcome. Whereas 96% of patients displayed a diploid peak in the distribution histograms, the HC, 1N and 4N peaks were absent from the majority of samples. In 13 ejaculates, either a HC or 1N or 4N peak, or a combination of these, was discernible. Although seminal germ cell numbers bore no overall association with elongated spermatids (ES) in histology or spermatozoa retrieval in TESE outcome, 4N cells per ejaculate were correlated with the percentage of tubule sections showing SPC as the most advanced germ cells. The incidence of HC peaks was higher in patients showing some ES in histology or sperm retrieval than in the sperm-negative groups. In groups with suspected obstruction showing nearly full spermatogenesis and maximal sperm retrieval, there was no incidence of a HC peak. Germ cell peaks were associated with germ cell degeneration noted in testicular histology. In conclusion, seminal germ cells cannot provide good prognosis for TESE, although their presence could indicate the spermatogenic activity in the testis.