Intracytoplasmic spermatid injection in non-obstructive azoospermia: a review of current evidence and future directions

Abstract

Non-obstructive azoospermia (NOA) represents one of the most severe forms of male infertility, with spermatozoa retrieval only possible in about half of the cases. Spermatids can be identified in up to 30% of men with negative spermatozoa during testicular sperm extraction (TESE) procedure. Intracytoplasmic spermatid injection has been proposed for TESE-negative patients desiring biological parenthood. However, significant limitations jeopardize the clinical implementation of this technique. Key challenges include the accurate identification of viable spermatids and the need for artificial oocyte activation. These limitations contribute to lower fertilization, pregnancy, and live birth rates compared to conventional intracytoplasmic sperm injection using mature sperm. Additionally, the absence of rigorous randomized controlled trials and the predominance of low-quality and outdated underpinning studies have significantly jeopardized its clinical implementation. These concerns have led the Practice Committee of the American Society for Reproductive Medicine to consider the technique experimental. Recent advances in microscopy, improved tissue processing, and refined activation techniques have enhanced outcomes. Furthermore, while initial safety concerns about epigenetic modifications persist, follow-up studies of spermatid injection offspring have shown normal development. This review comprehensively explores current evidence regarding intracytoplasmic spermatid injection techniques, focusing on their utilization, efficacy, and safety in men with NOA. We also evaluate spermatid identification and retrieval methods, ethical considerations, technical limitations, and emerging technologies that could enhance outcomes.

INTRODUCTION

Approximately 1% of all men and 10%–15% of infertile men have azoospermia.1,2 Non-obstructive azoospermia (NOA), which constitutes two-thirds of azoospermia cases, can arise from various forms of testicular dysfunction and exhibit diverse histopathological patterns.2,3 No known remedy exists that can effectively reinstate spermatogenesis in most NOA patients, except in cases of secondary testicular failure, when medical therapy can be attempted.4 Consequently, the sole approach of management without a donor’s involvement is to extract spermatozoa directly from the testis for intracytoplasmic sperm injection (ICSI).5

Despite the advances in surgical sperm retrieval techniques, there remains a considerable number of men with NOA whose spermatozoa cannot be found, even after repeated surgery.6 With the absence of testicular spermatozoa retrieved, spermatids have been considered as the alternative for these men to obtain their biological child.7 Round spermatids are the first haploid germ cells in the spermatogenesis process, which later will undergo maturation into spermatozoa.8 Current evidence showed that round spermatid injection (ROSI) resulted in a low pregnancy rate and a low live birth rate.9 Such low success rate is also coupled with concerns about safety and potential risks to the offspring, limiting this approach from regular practice.

The present article aims to overview the available literature regarding the utilization, efficacy, and safety of spermatid injection techniques in men with NOA. Additionally, the review will explore the techniques employed in identifying and retrieving spermatids. Furthermore, the present review will highlight ethical issues and limitations jeopardizing the use of spermatid for NOA. We also explore future avenues for research, technologies, or therapies that have the potential to enhance the efficacy of intracytoplasmic spermatid injection in males diagnosed with NOA.

An extensive search on the PubMed database was carried out by two independent authors (DP and MS) using the keywords “spermatid” and “non-obstructive azoospermia” to identify relevant articles published up to November 2024. Research addressing the role of spermatids in male infertility or NOA was included in this review. The initial screening involved reviewing titles and abstracts of articles retrieved from PubMed, followed by a detailed evaluation of the full texts of selected abstracts. Additionally, a manual search of the references from the selected articles was performed to identify further relevant articles. Among 104 articles reviewed, 46 studies were chosen based on relevance including 32 clinical studies investigating the outcome of intracytoplasmic spermatid injection in infertile men with NOA7,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40 (Table 1).

Table 1.

Summary of clinical studies reporting outcomes of intracytoplasmic spermatid injection

Study	Study design	Type of spermatids	Source of spermatids	Fresh/frozen-thawed	Activation method	Injected oocyte (n)	Fertilized oocyte (n)	Fertilization rate (%)	Number of transferred embryos	Number of pregnancy	Total number of live offspring
Vanderzwalmen et al.10 1995	Case report	Round	Testis	Fresh	NR	1	1	100	1	NR	NR
Hannay11 1995	Case report	Round	NR	NR	NR	NR	46	NR	NR	4	0
Fishel et al.12 1995	Case report	Elongated	Testis	Fresh	NR	10	1	10	1	1	NR
Tesarik et al.13 1995	Case series	Round	Ejaculates	Fresh	Vigorous ooplasm aspiration	39	14	36	NR	2	1
Tesarik and Mendoza14 1996	Clinical trial	Elongated, round	Ejaculates	Fresh	Vigorous ooplasm aspiration	86	37	43	NR	NR	NR
Chen et al.15 1996	Case report	Round	Testis	Fresh	Vigorous ooplasm aspiration	13	4	31	2	0	NA
Vanderzwalmen et al.16 1997	Clinical trial	Elongated, elongating, round	Testis	Fresh	Vigorous ooplasm aspiration and ionophore	296	80	27	39	5	3
Antinori et al.17 1997	Clinical trial	Elongated, round	Testis	Fresh	Vigorous ooplasm aspiration	258	146	57	111	5	NR
Antinori et al.18 1997	Clinical trial	Round	Testis	Frozen-thawed	Vigorous ooplasm aspiration	15	7	47	6	1	NR
Amer et al.19 1997	Retrospective cohort	Round, elongated	Testis, ejaculate	Fresh	NR	544	129	24	53	2	NA
Yamanaka et al.20 1997	Clinical trial	Round	Testis	Fresh	None	49	34	69	24	0	NA
Kahraman et al.21 1998	Clinical trial	Elongated, round	Testis	Fresh	Vigorous ooplasm aspiration	230	72	31	43	3	2
Barak et al.22 1998	Clinical trial and case report	Elongated, round	Ejaculates, testis	Fresh	None	174	59	34	4	1	1
Bernabeu et al.23 1998	Clinical trial	Elongated, round	Ejaculates, testis	Fresh	NR	76	10	13	10	1	1
Sofikitis et al.24 1998	Clinical trial	Elongating	Testis	Fresh	NR	73	52	71	41	2	2
Ghazzawi et al.25 1999	Clinical trial	Round	Testis	Fresh	Vigorous ooplasm aspiration	574	126	22	40	0	NA
Al-Hasani et al.26 1999	Clinical trial	Elongated, round	Testis	Frozen-thawed	None	67	19	28	14	2	NR
Gianaroli et al.27 1999	Case report	Round	Testis	Frozen-thawed	Vigorous ooplasm aspiration	5	2	40	2	1	1
Balaban et al.28 2000	Case–control	Round	Testis	Fresh	Vigorous ooplasm aspiration	356	200	56	NA	NA	NA
Tesarik et al.29 2000	Case report	Elongated	Testis	Fresh	NR	6	5	83	5	1	2
Levran et al.30 2000	Retrospective study	Round	Testis	Fresh, frozen-thawed	Vigorous ooplasm aspiration	178	81	46	48	0	NA
Vicdan et al.31 2001	Case–control	Round	Testis	Fresh	Vigorous ooplasm aspiration	60	17	29	5	0	NA
Borges et al.32 2002	Case–control	Round, elongated	Testis	Fresh	NR	107	31	29	23	2	NR
Urman et al.33 2002	Clinical trial	Round	Testis	Fresh	Vigorous ooplasm aspiration	1021	414	41	16	0	NA
Sousa et al.34 2002	Retrospective cohort	Elongated, round	Testis	Fresh, frozen-thawed	None	515	186	36	NR	10	NR
Sousa et al.35 2002	Case–control	Elongated, elongating, round	Testis	Fresh	None	107	29	27	NR	NR	NR
Khalili et al.36 2002	Clinical trial	Elongated, round	Testis	Fresh	None	57	18	32	14	1	NR
Ulug et al.37 2003	Retrospective cohort	Round	Testis	Fresh	Vigorous ooplasm aspiration	36	15	42	10	0	NA
Goswami et al.7 2015	Case series	Round	Testis	Fresh	Calcium injection ± ionomycin	25	10	40	6	NR	NR
Tanaka et al.38 2015	Clinical trial	Round	Testis	Fresh, frozen-thawed (embryo)	Electric current	734	437	60	208	30	12
Tanaka et al.39 2018	Clinical trial	Round	Testis	Fresh, frozen-thawed (embryo)	Electric current	14 324	8132	57	3882	138	90
Cheng et al.40 2025	Retrospective cohort	Round	Testis	Fresh	Calcium ionophore	NR	NR	51.0	NR	2	3

NR: not reported; NA: not applicable

HUMAN SPERMATOGENESIS

Spermatogenesis is the biological process by which primordial diploid germ cells (PGCs) differentiate into mature haploid spermatozoa.41 The maturation process of human sperm, from spermatogonial stem cells to fully developed ejaculated sperm, lasts from 42 days to 76 days.42 Upon arrival at the gonad, the PGCs develop into gonocytes, which will differentiate into spermatogonia after birth.41 Spermatogonia have the characteristics of well-defined nuclear envelopes and granular chromatin with one or more distinctive nucleoli.43

There are 3 subtypes of spermatogonia based on their heterochromatin content, which are Ad (darker), Ap (lighter/pale), and B (dense clump heterochromatin at nuclear periphery). Both spermatogonia (Ad and Ap), also called A1 and A2, are regarded as the stem gonocytes. The former refers to the reserve stem gonocytes, whereas the later denotes the active ones that will proliferate for self-renewal or develop into spermatogonia B.44 Spermatogonia B increases in number and undergoes mitosis to generate preleptotene spermatocytes, signaling the end of the proliferative phase.45,46 The preleptotene spermatocytes are transported across the blood–testis barrier to become leptotene spermatocytes and advance to 2 cycles of meiosis.45 Following meiosis I, two secondary spermatocytes are produced, each possessing a haploid chromosome number. In meiosis II, a single secondary spermatocyte undergoes division into two haploid spermatids, which remain connected through cytoplasmic bridges.47,48 The spermatids possess haploid nuclei yet may exhibit functional diploidy owing to the ease with which a gene product synthesized in one cell can diffuse into the cytoplasm of its adjacent cells.49

Based on the morphology, cytoplasm, and tail size, spermatids can be classified into three categories: round, elongating, and elongated spermatids (Figure 1).50 Round spermatids are the initial cells that possess a haploid set of chromosomes.51 They are sized 6.5–8 µm and may exhibit a developing acrosomal structure as a bright spot on one side of the cell due to nucleus transfer to a peripheral location during the acrosomal phase.50 Nevertheless, this characteristic may not be observed in a round spermatid in the Golgi phase.50 Subsequently, round spermatids undergo elongation, signifying the commencement of the maturation phase.50

Figure 1.

Schematic figures of spermatids in different stages. Created in BioRender. Savira, M. (2025) https://BioRender.com/ce4brua.

Elongating and elongated spermatids are differentiated based on the tail length encased in cytoplasm.50 The elongating spermatid exhibits a relatively oval shape, with the nucleus positioned at the cell periphery and a protruding tail, signifying the shift from rounded to elongated spermatid.52 In contrast, elongated spermatids will develop into late elongated spermatids, characterized by a reduction in cytoplasmic volume and a mature nuclear morphology.52 It will then become mature spermatozoa with a functional flagellum. This maturation process from round spermatids to spermatozoa is called spermiogenesis, which involves cytoplasmic and nuclear changes without further cell meiosis.8

INDICATIONS OF INTRACYTOPLASMIC SPERMATIDS INJECTION

In cases of failed retrieval of spermatozoa in NOA patients, employing spermatids for ICSI may be the only solution for these patients to conceive biological offspring. Additionally, spermatids can be cryopreserved for future use.7,18,26,27,30,34,38,39 This approach not only addresses severe forms of male infertility but may also provide opportunities for fertility preservation in pediatric cancer patients undergoing gonadotoxic chemotherapy.53

SPERMATID RETRIEVAL TECHNIQUES

A spermatid can be found in either ejaculate or testicular tissue.16,54 A study found that round spermatids could be detected in ejaculates in 83.7% and 22% of TESE-positive and TESE-negative cases, respectively.55 Similarly, Timm et al.56 reported the discovery of spermatids in 11 out of 17 (64.7%) semen samples of patients with NOA. They stained the centrifuged seminal pellets with the panoptic method and used immersion light microscopy with 1250× magnification to identify spermatids. This indicates that careful assessment may improve the chances of finding viable germ cells for fertility treatment in NOA patients.

Unlike Timm et al.56 who used a staining method that would interfere with spermatid viability, Tesarik and Mendoza14 adopted morphological identification of ejaculate spermatids according to their morphology. They reported the preferential isolation of living spermatids from the 70% fraction of Percoll used during the density gradient preparation of semen. The ejaculate-isolated spermatids were used to perform intracytoplasmic injection of oocytes, reporting a successful one-live birth.13,14 Other authors adopted the method developed by Tesarik and Mendoza,14 with or without slight modifications, and reported live birth after ICSI of round/elongated spermatid obtained from the ejaculate.22,23

However, the use of ejaculated spermatids is discouraged if the possibility of obtaining testicular spermatozoa has not been explored. In such cases, a testicular biopsy or sperm extraction to retrieve testicular spermatozoa would be worthwhile, as the success rate of ICSI with spermatozoa is superior to that with spermatids.54 On the contrary, if testicular spermatozoa or ejaculated spermatids cannot be obtained, spermatids may be extracted from testicular tissue.

While it remains uncertain whether ejaculated or testicular spermatid extraction will be the most effective for intracytoplasmic spermatid injection, a study by Fishel et al.54 suggests that elongated spermatids from the testicles had a greater fertilization rate (38% vs 18%) than those obtained from the ejaculate. On the other hand, fertilization rates for ejaculate round spermatids were higher than those coming from testicular round spermatids (33% vs 22%).54 In a case where testicular spermatids are required, there are several surgical techniques that can be considered, including testicular sperm aspiration (TESA), conventional testicular sperm extraction (cTESE), and microdissection testicular sperm extraction (micro-TESE).

Testicular sperm aspiration

TESA involves blind percutaneous aspiration of testicular tissue using an 18- to 22-gauge needle to obtain sperm by creating negative pressure from the syringe plunger retraction.57 In case no sperm is retrieved after single aspiration, TESA can be done systematically throughout the whole testis on both sides.58 Despite being a simple procedure, TESA has an overall low surgical sperm retrieval rate (SRR) at around 10%–30% in NOA patients.59,60 For spermatids, there is a limited research that specifically mentioned the success rate for spermatid retrieval using TESA. Borges et al.32 employed TESA and percutaneous epididymal sperm aspirations (PESA) to retrieve spermatozoa and spermatids in azoospermic males. They successfully acquired round spermatids and elongated spermatids in 14 ICSI cycles where no spermatozoa were available.32 Nevertheless, the recent guidelines are against using TESA for NOA due to the low successful sperm retrieval rate.61

Conventional testicular sperm extraction

cTESE procedure is an open surgical technique using single or multiple incisions on the testes to obtain biopsies of testicular parenchyma for the purpose of searching and isolate sperm, without the use of optical magnification.57 It has an SRR of around 46% in NOA patients.62 Multiple biopsies are often performed in cTESE as they can improve the precision of diagnosing absolute testicular failure and augment the quantity of retrieved sperm cells.63 Vicdan et al.31 reported the utilization of cTESE to retrieve spermatids, revealing that among 18 men with NOA, round spermatids were identified in seven patients, spermatozoa in nine patients, and no spermatids or spermatozoa in the other two patients. In NOA patients in whom round spermatids were found, fertilization was attempted using ROSI. Nevertheless, none of them successfully achieved pregnancy afterward.

Microdissection testicular sperm extraction

The micro-TESE methodology was initially documented by Schlegel64 in 1999. Using an operating microscope with a magnification range of 15×–25× during TESE enables the visualization of more dilated seminiferous tubules with higher odds of containing spermatozoa. From a laboratory standpoint, micro-TESE is considered an optimal approach as it restricts the quantity of tissue that necessitates scrutiny by the andrologist to identify sperm to the most prolific sperm-producing tissue. At proficient centers, the standard duration for locating sperm in isolated and dispersed testicular tissue samples is merely 3–5 min.65 Additionally, micro-TESE results in fewer short-term and long-term complications compared to cTESE due to better preservation of testicular vascularization and minimal testicular tissue extraction,66 although some research found that serum testosterone may be significantly reduced after micro-TESE.67,68

The use of micro-TESE has been reported to yield an SRR of 43%–63% in patients with NOA.69 Tanaka et al.38 reported successful spermatid retrieval with a repeated micro-TESE in 10.4% of 730 NOA cases, in which the initial micro-TESE conducted in other institutions failed to find spermatozoa. They also reported fourteen successful live births achieved by intracytoplasmic spermatid injections. This outcome was facilitated by appropriate tissue processing and accurate spermatid identification.38 Data comparing the results of micro-TESE versus cTESE in retrieving spermatids is lacking. However, a meta-analysis found that successful sperm retrieval is 1.5 times more likely with micro-TESE as compared to cTESE.70

TECHNIQUES FOR INTRACYTOPLASMIC SPERMATID INJECTION

In general, intracytoplasmic spermatid injection works better only when spermatids present at the more mature form, such as elongating or elongated stage.21 Unfavorable outcome of immature spermatid has been attributed to unrecognized apoptosis,71 multiple centriole (normal situation in spermatids),72 and/or decreased capability of oocyte activation.73

Round spermatid injection

ROSI procedures start with identifying round spermatids from either ejaculates or testicular biopsy samples, injecting a spermatid into oocytes, and post-injection oocyte activation.40 Accurate identification of round spermatids is crucial for improving clinical outcomes of ROSI. However, unlike the elongating or elongated spermatids that are relatively easy to recognize, round spermatids identification is more challenging.74,75 Laboratory error can lead to mistaken round cells other than round spermatids for injection, which results in a failed ROSI.

Cytological features such as round shape sized 6–8 μm in diameter with smooth outline, thin rim of cytoplasm around the nucleus, acrosome granule and vesicles, are initially used to identify round spermatids.14,38,75 The acrosome vesicle was the most important distinctive feature for the recognition of round spermatids. However, a study found that it was only seen in less than 10% of round spermatids during identification.38 Moreover, the acrosome granule can oftentimes be confused as a vacuole.76 As a result, differentiating round spermatids from other round cells with similar morphology, such as spermatocytes, spermatogonia, lymphocytes, and somatic cells, is challenging. Furthermore, human round spermatids do not exhibit a distinct chromatin granule under Hoffman modulation contrast microscopy like in mice.77

Another way is by looking at integrity during aspiration as proposed by Tesarik and Mendoza.14 Unlike somatic cells, aspirated round spermatids retain structural integrity during pipette aspiration. However, this method has been critiqued, as such descriptions align more closely with somatic cell characteristics.75 It is lymphocytes that possess a resilient plasma membrane, rendering them impervious to disruption even with vigorous pipetting. Moreover, in contrast to spermatogonia, the cytoplasm of round spermatids can be easily separated from the nucleus by pipetting.38 Confocal scanning laser microscopy and phase-contrast optics on an inverted microscope can be reliable alternative modalities for round spermatid identification.20,76 An inverted microscope equipped with phase-contrast optics could provide a clear image, allowing reliable recognition of round spermatids in cell suspensions smeared at the glass bottom of the dish.76

To separate round spermatids from other cells, the gradient separation technique is the standard seminal plasma processing technique that was used previously to aid the separation,14 and the technique can be further optimized for more efficient yield.78 As another option, Hayama et al.79 used a flow cytometry technique to isolate rat and mouse round spermatids while maintaining their viability and DNA integrity, using them to perform ROSI, and obtaining healthy offspring.

Several other techniques allow more accurate identification of round spermatids with high certainty. However, these techniques interfere with cell viability and hence can be used to validate clinically applicable techniques, rather than being used directly to identify round spermatids before ROSI. While immunohistochemical staining can assist in distinguishing round spermatids from other cells with similar morphologies, the standard approach to identifying round cells is using staining, particularly immunofluorescence staining with spermatid-specific markers.80,81 Some spermatid-specific markers include protamine-1 (PRM-1), protamine-2 (PRM-2), and cAMP responsive element modulator (CREM).81,82 To complement, alternative approaches are the use of quantitative or digital reverse transcriptase-polymerase chain reaction, and fluorescence in situ hybridization can also be employed to identify spermatids based on the stage-specific gene expression during spermiogenesis.38,78,83,84 Single-cell sequencing technology has emerged as a new method for accurately identifying human germline cells, including spermatids, with high sensitivity and specificity.85 Computer-aided identification is another simple and promising technique yet under development.38 Methods for identifying round spermatids are summarized in Table 2 and Figure 2.

Table 2.

Methods to identify spermatid

Method	Description	Advantages	Limitations
Methods preserving spermatid viability
Conventional Hoffman modulation contrast microscopy	Detects optical gradients and converts them into variations of light intensity to identify the shape of spermatid	Simple and widely available	Relatively difficult and requires expertise to differentiate from other cells
Inverted phase-contrast microscopy	Uses an inverted microscope with phase-contrast optics to identify spermatids based on size, shape, and nuclear features	Simple and widely available	Relatively difficult and requires expertise to differentiate from other cells
Confocal scanning laser microscopy	Producing longitudinal optical sections in the layer of spermatogonia, spermatocytes, and spermatids	Provides detailed cellular information	Expensive and not accessible to all laboratories
Gradient separation	Separates cells based on density through centrifugation in a gradient medium	Relatively simple, effective for isolating viable sperm based on motility and morphology	May not differentiate between all cell types effectively and requires optimization
FCM	Analyzes and sorts cells based on light scattering and fluorescence properties	Allows rapid analysis and assessment of multiple parameters simultaneously	Requires specialized equipment and expertise; potential for sample contamination
Computer-aided identification	Emerging technology for automated spermatid detection	Potential for high-throughput screening	Still in development and not yet widely available
Methods sacrificing spermatid viability
Immunochemical imaging	Visualizing spermatid using a fluorescently labeled antibody	Rapid identification	Requires specialized equipment and expertise
FISH	Confirms haploid status of selected cells	Accurate in identifying round spermatids	Cannot be used for clinical procedures
Single-cell sequencing technology	Analyzes gene expression and mutations at the single-cell level using RNA sequencing	Provides detailed insights into cellular heterogeneity and gene expression profiles	High cost; requires advanced bioinformatics for data analysis; and limited to specific applications

FCM: flow cytometry; FISH: fluorescence in situ hybridization

Figure 2.

Methods facilitating round spermatid identification. Created in BioRender. Elsuity MA. (2025) https://BioRender.com/w19y792.

Upon finding the round spermatid, the subsequent step is to inject it into the oocyte and initiate artificial oocyte activation (AOA), which is needed for fertilization and embryogenesis.86 AOA is required in ROSI due to the lack of oocyte-activating capacity in round spermatids.73 Oocyte activation involves a series of increases in the cytosolic-free Ca²²⁺ concentration, thereby inducing calcium oscillation.87,88 A study found that human spermatids could activate 64% of oocytes, with only 35% of the oocytes exhibiting type A calcium oscillation, in contrast to spermatozoa, which activated 100% of oocytes.73 Several methods of AOA are available,86 including (1) mechanical activation (intended to manually disrupt the plasma membrane, thereby increasing the oocyte Ca²⁺ load during injection) using vigorous cytoplasm aspiration;13,14,15,16,17,18 (2) chemical activation with calcium ionophore or ionomycin treatment that promotes intracellular calcium release and facilitates the influx of extracellular calcium ions;7,40,89 and (3) electrical activation using electrostimulation that induces calcium influx through the formation of pores in the plasma membrane.38,39 Unfortunately, there has not been any formal comparative study that evaluates the effect of each type of AOA method based on the Ca²²⁺ oscillation. However, vigorous oocyte aspiration was primarily used in the early studies of ROSI, whereas electrical stimulation and chemical activation are more commonly utilized in recent studies. Some authors considered that a mechanical approach with vigorous oocyte activation may not be sufficient for AOA.75,90 After the successful AOA process, the oocytes are subjected to the standard ICSI procedure, encompassing fertilization monitoring, embryo culture, and embryo transfer.

Round spermatid nuclei injection (ROSNI)

ROSNI and ROSI are sometimes interchangeable in the literature.91 However, it is essential to differentiate them due to the material injected into the oocyte during ICSI. While ROSI uses the entire round spermatid, ROSNI only uses the nuclei, which might increase the risk of genetic defects in offspring.88 Using immature sperm in the ROSNI procedure poses novel technical obstacles and unresolved genetic apprehensions. Initially, the ROSNI technique has been proposed as a potential solution to address certain limitations of ROSI, as indicated by findings from animal experimentation.92,93,94 These include a smaller diameter of micropipettes utilized for ROSNI that minimize the likelihood of oocyte harm.92,93,94 Concurrently, Yavas et al.95 also found that using an injection pipette with the smallest inner diameter could enhance the fertilization rate and embryo development in ICSI. Another drawback of ROSI addressed by ROSNI is the substantial cytoplasm surrounding the spermatid nucleus during ROSI, which may hinder the conversion into the male pronucleus.92,93,94 Nonetheless, comparable successful outcomes of ROSNI were not observed in human trials.

Yamanaka et al.20 reported that forty-nine mature oocytes were successfully injected with nuclei and then cultured for 72 h. From there, twenty-four embryos were transferred to nine women. However, no pregnancy was achieved. Hannay11 also reported that all round spermatid-derived pregnancies achieved by ROSNI in 4 couples ended in spontaneous abortion. As per the recommendation of Practice Committee of the American Society for Reproductive Medicine, it is suggested that ROSNI be regarded as an experimental procedure, and its application should be limited to the clinical trial setting.91

The clinical significance of utilizing the entire round spermatid instead of solely utilizing nuclei was underscored by Tesarik and Mendoza.14 There are three reasons why using complete spermatids is more advantageous. First, the oocyte activating factors are predominantly found within the cytoplasmic cells of spermatids.96 Second, the isolation of nuclei from spermatids could result in the depletion of centrosomal components that is essential for ensuring proper alignment and arrangement of mitotic spindles during the early stages of embryonic development.97 Lastly, the process of isolating nuclei in spermatids has the potential to cause harm to the nuclear envelope, resulting in the partial loss or damage of intranuclear materials.14

Elongated spermatid injection (ELSI)

ELSI is considered one of the best options for assisted conception using immature spermatozoa in males with NOA, with a higher success rate compared to ROSI.98 Elongated spermatids represent the most advanced stage of spermatid development, right before spermatozoa development.69 This approach is based on the notion that the histone–protamine transition has already commenced in this stage of meiosis.99 The presence of protamines may confer a protective effect on the sperm chromatin against maturation factors present in the oocyte.100 This rationale has been considered in the context of utilizing human-elongated spermatids.

Elongating and elongated spermatids could be obtained from round spermatids using in vitro maturation. During the experiment by Cremades et al.,101 37 round spermatids were cocultured in monolayer Vero cells for a maximum of 5 days. Out of these, almost half matured into elongating spermatids, and about one-fifth matured into elongated spermatids. Primary cultures of three round spermatids with flagella in the same patient resulted in the creation of one additional mature spermatozoon. Tanaka et al.102 in their study with Vero cells, also confirmed the success in resuming spermiogenesis in around 50% of the 1850 round spermatids from men with azoospermia.

This achievement is highly constrained in the context of in vitro experimentation. Consequently, cultivating samples to obtain haploid round spermatids could have broader clinical utility in other individuals.103 However, it is still crucial to develop a novel approach for men with pure Sertoli cell-only disorders and for those who have spermatogenesis up to the spermatocyte stage but do not achieve cell maturation during in vitro culture.

OUTCOMES OF INTRACYTOPLASMIC SPERMATID INJECTION

The initial successful application of the ROSI technique was reported by Ogura and Yanagimachi,93 who asserted the capability of fertilization with round spermatid using hamster round spermatid in 1993. Subsequently, a study in human subjects was reported by Vanderzwalmen et al.10 in 1995, which resulted in the successful development of two polar bodies in the injected oocyte. The first documented instances of successful ICSI births with spermatids were achieved by Fishel et al.12 and Tesarik et al.13 who utilized round spermatids obtained from the ejaculate and elongated spermatids retrieved from the testis, respectively.

After the initial phase, several studies were conducted to evaluate the outcome of intracytoplasmic spermatid injection in the mid-nineties. Subsequent clinical studies found that clinical pregnancy rates and live births remained limited, as shown in Table 1. However, there seemed to be significant differences in the outcomes of early studies compared to the more recent work. In 2015, Tanaka et al.38 reported the successful birth of 14 infants with ROSI. In 2018, their group documented 90 successful live births from 721 men who underwent ROSI.39 This progress is attributed to advancements in microscopy, improved tissue preparation methods utilizing enzymatic preparation with DNase and collagenase, and refined oocyte activation techniques involving electric current.38,39,75

As for overall success, a recent meta-analysis of 22 studies that comprised 1099 couples and 4218 embryo transfers found that the mean fertilization rate following ROSI was 38.7% (95% confidence interval [CI]: 31.5%–46.3%) and a much lower mean pregnancy rate of 3.7% (95% CI: 3.2%–4.4%).9 The live birth rate was relatively low, as only 4.3% of embryo transfers resulted in live births. The study found that the pregnancy rate per couple was 13.4% (95% CI: 6.8%–19.1%), and the live birth rate per couple was 8.1% (95% CI: 6.1%–14.4%).9 The most recent work on the success of ROSI was published by Cheng et al.40 from Taiwan, China, in 2025. They found no significant differences in the fertilization rates (56.3% vs 51.0%) and implantation rate (10.3% vs 4.0%) in the ICSI compared to ROSI groups after adjusting for the use of anastrozole, although the live birth rate was still significantly lower in the ROSI group (8.3%) than that in the ICSI group (30.8%).40 In total, there have been at least 100 live births through the ROSI technique, mainly those that were reported by Tanaka et al.39 in the 2-year follow-up results of 90 ROSI babies with the addition of 3 live births from Cheng et al.40

In regard to the utilization of fresh and frozen-thawed spermatids, several important considerations emerged. Cryopreserved spermatids usually serve as a safeguard, as there is no assurance of complete retrieval of spermatids during micro-TESE. However, early reports predominantly utilized fresh round spermatids as cryopreservation techniques might not have been adequately developed, with only limited reports using frozen-thawed spermatids. Included among these were two successful trials employing vitrification as the cryopreservation method. Antinori et al.18 reported using frozen-thawed spermatids 3 months post-retrieval, achieving a viability rate of 70%, which resulted in one successful ongoing clinical pregnancy. Gianaroli et al.27 reported another case, a successful live birth rate utilizing frozen-thawed spermatids in 1999. From then on, the advancement of the cryopreservation method has led to an improvement in the quality of frozen-thawed sperm.

A study by Levran et al.30 found that cleavage and fertilization rates following ROSI with fresh spermatids were comparable to those with frozen-thawed spermatids. Differently, Ogonuki et al.,104 in their animal study, found the benefit of frozen-thawed round spermatids in providing some endogenous oocyte activation as opposed to fresh round spermatids. Tanaka et al.39 also observed a superior pregnancy rate and a superior live birth rate in frozen-thawed embryos from ROSI compared to fresh ones, with 15.8% versus 5.4% and 10.6% versus 3.1%, respectively. These findings suggest that frozen-thawed ROSI embryos may be more effective than fresh embryos, particularly if it is coupled with enhanced uterine endometrial preparation, the selection of superior-quality spermatids, and the availability of genetic testing before embryo transfer. In terms of cryopreservation methods, vitrification is recognized to outperform conventional freezing techniques (mean±standard deviation [s.d.]: 72.9 ± 15.4 vs 58.2 ± 17.7, P< 0.05) in terms of survival rates for round spermatids in mice.105 However, the slow freezing technique seemed to result in superior outcomes to vitrification for human spermatids.106

When comparing ELSI to ROSI, Kahraman et al.21 reported a case series of 20 patients. They found a fertilization rate of 71% in three cases where elongating and elongated spermatids were used, compared to just 25.6% in situations where round spermatids were used.21 During follow-up, twin healthy babies from a single pregnancy were successfully born in a patient who underwent ELSI. Conversely, round spermatids resulted in one biochemical pregnancy and eventually led to an early pregnancy loss. Therefore, the ELSI technique is preferred whenever elongating or elongated spermatids can be obtained.

Nevertheless, continued refinement and development of strategies for intracytoplasmic spermatid injection are still needed. It was determined that a lack of mature centrosomes, abnormal demethylation, and epigenetic alterations are the fundamental factors contributing to unsatisfactory outcomes. Hence, rectifying epigenetic mistakes can enhance embryonic growth to its maximum potential.75

ETHICAL AND SAFETY ISSUES ON THE USE OF SPERMATIDS IN ASSISTED REPRODUCTION

A preliminary investigation indicated that in the spermatids of men with NOA undergoing TESE–ICSI, altered chromatin condensation was commonly found.107 It further raised the concern on the impact of spermatids on the genetics and any anomalies in offspring. In an animal study, spermatids of infertile mice had a higher rate of aneuploidy.108 Therefore, it was suggested that infertile males with independent genetic mutations may have a high risk of aneuploidy in their spermatids. Additionally, the genetic safety issues related to concurrently applying multiple oocyte activation techniques have not been well studied.109

Other concerns have also been raised about the high rates of DNA fragmentation observed in round spermatids retrieved from testicular biopsies of azoospermic patients.110 Moreover, ROSI is often associated with the possibility of epigenetic modification due to the incomplete transition of histone nuclear proteins to protamines, resulting in the low implantation rates observed in ROSI embryos.75 Wang et al.111 found that ROSI embryos at the pronuclear stages had a reprogramming defect, primarily due to misexpression of a cohort of minor zygotic genome activation. However, serious epigenetic abnormalities in ROSI embryos may be naturally eliminated, as no confirmed abnormalities have been reported in children born through ROSI. A study on embryonic day 11.5 mouse fetuses and placentas found that the transcriptome and DNA methylation are similar in ROSI, IVF, and ICSI group.112 Even if there was partial DNA methylation, obtaining normal mice offspring from ROSI had been reported, albeit with very low success rates.113 Similarly, another study demonstrated that ART-associated epigenetic variation at birth as depicted in DNA methylation abnormalities in neonatal blood cells, tended to resolve by adulthood, suggesting no significant long-term effects on growth or health.114 This finding provides a hopeful perspective but underscores the importance of further research to track the long-term implications of embryonic methylation abnormalities.

Antinori et al.17 reported that chromosomal analysis showed that all fetuses in four ongoing pregnancies from spermatids had a normal karyotype without any chromosomal abnormalities. Tanaka et al.39 reported that only 3.3% of 90 ROSI babies observed congenital malformations, including ventricular septal defects, cleft lip, and omphalocele. This rate seems comparable with the prevalence of congenital abnormalities reported in the normal populations, which was 6.96% in Japan based on a nationwide real-world database and approximately 1.84% in India based on a meta-analysis of publications that included 52 hospital-based and three community-based investigations.115,116 Additionally, the cognitive development, body weight, weight gain, and height of the ROSI newborns were reported to be comparable to that of babies conceived through natural pregnancy upon follow-up up to 2 years old.39

LIMITATIONS

Many limitations jeopardize the clinical application of intracytoplasmic spermatid injection. None of the 32 primary clinical studies we identified in the literature were designed as a fully randomized clinical trial, and more than half were carried out before 2000. In a meta-analysis assessing the quality of 22 cohort ROSI studies, Hanson et al.9 in 2021 reported only 7/22 (32%) studies to be of high quality. The low quality of most reported studies casts doubt upon the spermatid injection technique for treatment of infertile men with NOA. Despite the emergence of many laboratory techniques facilitating the identification of spermatids, especially round one, the difficult morphological identification remains the standard method, with the other methods being expensive, underestablished, or interfering with spermatid viability (Table 2). The extremely low live birth rate of 84 live birth deliveries out of 1838 transfer cycles was reported by Tanaka et al.39 in 2018 in the largest study on this technique, and the potential epigenetic risks in born offspring are other important factors to be considered.

FUTURE DIRECTIONS

Despite being experimental, ROSI offers a prospective assisted reproduction alternative for males with NOA and sperm-negative testicular sperm extraction. For several couples, the opportunity to create biological offspring from both parents is seen as valuable. A study reported that up to 95% of couples whose spermatozoa were not found expressed a desire to attempt ROSI as an alternative to a donor even after being informed of its low success rate.39 Furthermore, the identification of spermatids in ejaculates may also be used as a simple yet important additional tool to distinguish between OA and NOA in azoospermic men.117 Consequently, future efforts should address critical issues in the clinical application of intracytoplasmic spermatid injection technology, including the precise selection of spermatids, low developmental efficiency, and long-term safety.

Advances in computer and artificial intelligence (AI) may be able to identify better live spermatids. A preliminary study had reported on the feasibility of AI use for sperm identification from testicular tissue samples retrieved by micro-TESE.118 Strategies such as in vitro germ cell maturation can enhance the quality and effectiveness of round spermatids. As demonstrated in an animal study, the procedure involves the transplantation of Sertoli cells into the testes, leading to an increase in testicular cell numbers, sperm count, and motility in cases of azoospermia.119 Likewise, coculturing with Sertoli cells or administering Sertoli cell-conditioned medium may improve the development of spermatozoa in cases of azoospermia.35,120 However, the evidence is still lacking and contradictory.

To effectively preserve the original methylation pattern of imprinted genes in embryos created from ROSI, an alternative method of utilizing the histone deacetylase inhibitors (Scriptaid and Trichostatin A) has been considered, although their benefits have only been studied in animal embryos.121 Additionally, in an animal study, A366 treatment (selective euchromatic histone lysine methyltransferase 2 inhibitor) may effectively correct epigenetic abnormalities, enhancing normal development and delivery rates of mouse ROSI embryos.111 Future ROSI embryos may also benefit from preimplantation genetic testing and the utilization of freeze-all cycles to achieve uterine and embryonic synchronization (Figure 3).9

Figure 3.

Proposed methods for optimization of intracytoplasmic spermatid injection. Created in BioRender. Elsuity MA. (2025) https://BioRender.com/h39g767.

For pre-medical treatment, animal studies have highlighted differences in aromatase activity and mRNA expression between round spermatids and spermatozoa. Cheng et al.40 found that the use of anastrozole may affect the outcome of ROSI, as the implantation rate is higher in those taking anastrozole compared to those not taking it (mean±s.d.: 5.9% ± 14.1% vs 2.4% ± 7.1%). However, as this study was not primarily designed to evaluate the effect of anastrozole, future research is still needed to clarify the potential effect of anastrozole on improving ROSI outcomes.

CONCLUSION

The clinical utility of intracytoplasmic spermatid injection in assisted conception is limited by the absence of rigorous, randomized controlled clinical trials, outdated and/or low-quality underpinning clinical studies, difficulties in implementation, low success rates, and epigenetic risks in born offspring. Accurate identification of spermatids, optimization of AOA, and in vitro germ cell maturation are expected to enhance the clinical success rates of intracytoplasmic spermatid injection. Large, well-designed clinical trials are required to increase the evidence related to spermatid use for men with NOA.

AUTHOR CONTRIBUTIONS

DP, MS, and RS conceptualized the review idea. DP and MS are the project administrators. DP, MS, and MAE contributed to the review methodology. DP and MS wrote the original draft and tables, and MAE, HS, and RS revised them. MAE and MS designed the illustrative figures. All authors read and approved the final manuscript.

COMPETING INTERESTS

All authors declare no competing interest.