Abstract Study question Can epigenetic profiling of seminal plasma be used to predict successful testicular sperm retrieval for men with non-obstructive azoospermia (NOA)? Summary answer Epigenetic screening of cell-free seminal RNA identified gene imbalances in NOA men, with specific correlation to those who failed to yield spermatozoa at testicular biopsy. What is known already Although the chance of a successful microdissection testicular sperm extraction (micro-TESE) in men with NOA can be up to 60%, the procedure may still fail to yield spermatozoa. Several factors have been proposed to predict a successful retrieval, including FSH, inhibin B, genetics, and histopathology. Although histopathology would be the most reliable of these to predict successful micro-TESE, it is equally invasive to perform. Indeed, cell-free RNA extracted from testicular biopsy specimens has been shown to be differentially expressed in infertile men according to the origin of their azoospermia, whether obstructive or nonobstructive, and in relation to a normozoospermic control. Study design, size, duration Over a 2-year period, we identified men in whom no spermatozoa were identified despite extensive semen analyses conducted by multiple embryologists. These patients, who were negative for Y microdeletion, subsequently underwent micro-TESE. For consenting men, we performed epigenetic analyses on their seminal plasma by RNAseq. Significant differentially expressed gene (DEG) profiles were then assessed and compared according to whether surgical sperm retrieval successfully yielded spermatozoa (+TESE) or not (-TESE). Participants/materials, setting, methods RNA was isolated from the ejaculates for RNAseq using a commercially available spin column kit. RNA isolates were sequenced by Illumina HiSeq at 2x150bp. An absolute log2fold change of > 1 and a P-value of < 0.05 was considered significant. DEG profiles were compared within, as well as between, the +TESE and -TESE cohorts in comparison to a donor control. Main results and the role of chance All 12 men (37.3±6yrs) had normal peripheral karyotypes. Six (38.0±7yrs) underwent successful testicular sperm retrievals, defining the +TESE cohort. These men exclusively shared 10 significantly imbalanced genes involved in processes such as spermatogenesis (n = 4), sperm function (n = 2), and testis development (n = 1). For the 6 men (36.6±5 yrs) who underwent testicular sperm retrievals that failed to yield spermatozoa (-TESE), we identified 16 significantly imbalanced genes, exclusively shared by these patients. These genes are mainly involved in spermatogenesis (n = 9), sperm maturation (n = 1), and cell cycle regulation (n = 4). We then compared the DEG profiles between the +TESE and -TESE cohorts and identified 8 imbalanced genes that were shared among all 12 NOA men. Of interest, TPTE2 was partially (67%) expressed in patients from the +TESE group, while IGSF11-AS1 was underexpressed in all men from the -TESE group. Both of these genes are implicated in spermatogenic defects and are normally highly expressed in the testis. Interestingly, we identified a gene (NA) that was solely and specifically underexpressed in all men from the -TESE group, yet simultaneously overexpressed in all men from the +TESE group. NA, which is well known for its role in sialic acid metabolism, is also present on the sperm acrosome. Limitations, reasons for caution Using non-invasive RNAseq on the seminal plasma of NOA men, we were able to identify DEGs according to whether spermatozoa were successfully retrieved or had failed retrieval with micro-TESE. Although intriguing, these are preliminary results that should be further validated in a larger study cohort. Wider implications of the findings RNAseq identified genes shared within the same prognostic cohort. Moreover, differential expression of some specific genes predicted micro-TESE outcome. This epigenetic assessment, carried out on the ejaculate, can therefore be used as a non-invasive biomarker tool to predict loss of spermatogenesis in NOA men, sparing them from unnecessary surgery. Trial registration number N/A