Abstract Study question Can a genetic assessment of cell-free DNA, isolated from seminal plasma of nonobstructive azoospermic (NOA) men, predict successful sperm extraction at testicular biopsy? Summary answer Seminal cell-free DNA (cfDNA) isolated from the ejaculates of NOA men with failed testicular sperm retrieval outcomes concurrently and consistently displayed mutations on three genes. What is known already The most puzzling form of azoospermia is the testicular type, where the chances of successful microdissection testicular sperm extraction (micro-TESE) are unpredictable. Therefore, developing a method that could identify or predict individuals with absent spermatogenesis would be most useful. While histopathology is considered the most reliable method for predicting successful micro-TESE, it is invasive and often inconclusive. Recently, cfDNA testing has been gaining momentum as a noninvasive method for prenatal screening and even with the potential for eventually replacing PGT-A. However, the prognostic value of using cfDNA to predict the presence/absence of spermatozoa in men undergoing micro-TESE is relatively unknown. Study design, size, duration Over a period of 8 months, we recruited azoospermic men with normal peripheral karyotypes and for whom an extensive semen analysis did not identify spermatozoa. For consenting patients (n = 12), we performed genetic assessments on cfDNA from their ejaculates (IRB 1006101085). These patients subsequently underwent micro-TESE, and their cfDNA mutation profiles were compared according to whether spermatozoa were successfully retrieved ((+)Sperm, n = 5), or not ((-)Sperm, n = 7). Participants/materials, setting, methods Using a commercially available kit, circulating cfDNA was eluted from the ejaculates of consenting NOA men, as well as a normozoospermic control (n = 2). The isolated cfDNA specimens subsequently underwent whole exome sequencing (WES) on an Illumina HiSeq at 2x150bp. Gene mutations were detected by CLC Genomic Server 9.0 and compared within, as well as between, the +Sperm and -Sperm cohorts in comparison to the control group. Main results and the role of chance All 12 men (35.5±4yrs) had normal peripheral karyotypes and tested negative for Y-microdeletions. They had no history of genetic diseases and were off any supplements, medications, and/or hormones before or during the study. Genetic assessment on cfDNA from the +Sperm cohort (n = 5; 38.3±2yrs) identified 8 genes (ADAMTS7, ATP9B, CBFA2T3, FAM20C, FCGBP, PEBP4, PHF2, RPS9) that were concurrently mutated among all NOA men from this group. Although ADAMTS7 plays a potential role in sperm-egg fusion, the remaining genes are unrelated to sperm production. Whole exome sequencing of cfDNA from the -Sperm cohort (n = 7; 33.5±4yrs) consistently detected mutations on 3 genes, including one involved in spermatogenesis (CFAP54), one essential for oxidative regulation of spermatozoa (GSTT4), and one that encodes testis antigens that are normally expressed during spermatogenesis (BAGE2). Moreover, we concurrently and consistently identified pathogenic mutations on an additional gene in the -Sperm cohort. This gene (MUC2) is involved in the epithelial coating of mucus membrane-containing organs, and its expression has been correlated with spermatogenesis. Interestingly, these genes were all unaffected in men with successful micro-TESE outcomes. Limitations, reasons for caution Although we detected candidate gene mutations that correlate with micro-TESE outcome, these are preliminary findings that need further validation in a larger study cohort. Furthermore, while cfDNA analysis has shown great promise, it is not specific enough on its own and should be corroborated by epigenetic testing. Wider implications of the findings Genetic screening of seminal cfDNA provides a noninvasive alternative to histopathology and can be used to predict the absence of spermatogenesis, sparing these men from unnecessary surgery and cost. This novel biomarker approach lays the groundwork for precision medicine in male infertility testing. Trial registration number n/a