Replacement of the intracytoplasmic sperm injection (ICSI) holding pipette with a microfabricated device (microICSI) reduces changes to oocyte shape during sham injection of human oocytes

Article information

Korean J Fertil Steril. 2026;.cerm.2025.08144

Publication date (electronic) : 2026 March 11

doi : https://doi.org/10.5653/cerm.2025.08144

Rebecca L. Kelley, Cody Thomas, David K. Gardner

Melbourne IVF, Collingwood, VIC, Australia

Corresponding author: Rebecca L. Kelley Melbourne IVF, 36 Wellington Street, Collingwood, VIC 3066, Australia Tel: +61-394734410 E-mail: rebecca.kelley@mivf.com.au

*The study was funded by Virtus Health and Melbourne IVF. microICSI devices and training were supplied by Fertilis.

*Parts of this manuscript were presented at the Scientists in Reproductive Technology Annual 2-Day Meeting in Melbourne, Australia on 17th May 2024, and at the Fertility Society of Australia and New Zealand Annual Conference in Perth, Australia on 15th September 2024.

Received 2025 April 28; Revised 2025 September 30; Accepted 2025 November 24.

Abstract

Introduction

Intracytoplasmic sperm injection (ICSI) was introduced into human in vitro fertilization (IVF) in 1992 and has been used for 30 years with relatively little advancement in technology [1,2]. To perform ICSI, a denuded oocyte is immobilized using a glass holding pipette with suction, while an immobilized sperm is injected using an injection needle. The injection needle applies considerable point pressure as it penetrates the zona pellucida, inducing conformational changes in oocyte shape [3-5]. Once through the zona, the needle must then penetrate the oolemma, again applying substantial pressure on the oocyte and its cytoskeleton. Because the oocyte is held in place by the holding pipette, it becomes distorted into a toroidal shape (appearing oblong in two-dimensional view (2D) as the cytoplasm is invaginated. This invagination, called the injection funnel, can extend to more than two-thirds of the oocyte diameter before the oolemma finally gives way and the needle enters. The physical trauma of ICSI is clearly observable under the microscope, and intracellular strain is greatest at the oocyte center [6]. Importantly, the degree of distortion has been correlated with altered embryo development and gene expression [5,7,8]. It is accepted that ICSI results in lysis of some of the injected oocytes; the Vienna Consensus on key performance indicators recommends up to a 10% degeneration rate in a competent IVF laboratory [9]. The lysis rate is influenced by oocyte fragility, oolemma rigidity, patient age, and the embryologist’s experience and technique [10-13].

Using 2-photon polymerization three-dimensional (3D) printing, it has become possible to fabricate sub-micron-resolution devices for IVF applications [14]. The polymers used for these devices are safe for gametes and embryos, as confirmed in mouse embryo culture using an earlier prototype that yielded fertilization and blastocyst rates equivalent to standard microdrop culture [15]. One such device, microICSI, was designed to reduce stress on the oocyte during ICSI [15,16]. Instead of a glass holding pipette and suction, a clear 3D-printed receptacle for the oocyte is integrated onto a polystyrene dish, which does not require suction to hold the oocyte. Using a cup-shaped receptacle rather than a pipette and eliminating the use of suction may therefore reduce trauma to the oocyte [15,17]. In porcine oocytes, microICSI produced higher blastocyst rates than conventional ICSI (cICSI) while maintaining equivalent fertilization rates and blastocyst cell numbers [16]. These animal studies indicate the device is safe and may support improved embryo development even without increased fertilization rates. Recently, McLennan et al. [16] performed repeated sham injections on 18 donated frozen human oocytes and observed only one case of lysis, suggesting that microICSI may be suitable for human IVF.

microICSI also reduces the complexity and difficulty of ICSI by removing the need for a second micromanipulator, thereby shortening the procedure and decreasing the number of hand movements [16]. This simplified approach could reduce ICSI duration, minimize variability in fertilization rates among embryologists, and shorten training time. Such improvements may help alleviate workload burden, as embryologists currently manage high caseloads and experience injuries and burnout [18-20].

The aim of this study was to perform sham ICSI without sperm on donated fresh human oocytes not suitable for clinical use and to measure procedure duration and any changes in oocyte shape during injection. 

Methods

4. Image analysis

ImageJ (National Institutes of Health) 1.54g was used to assess changes in oocyte shape from still images extracted from ICSI videos [21]. Oocyte area, perimeter, diameter, roundness (smoothness), and circularity were measured at three time points: pre-injection immediately before needle penetration, mid-injection immediately before the oolemma broke, and post-injection immediately after needle withdrawal (Figure 1). The x-axis refers to the diameter across the oocyte center parallel to the injection needle, and the y-axis refers to the diameter across the center perpendicular to the needle. Circularity was calculated as (4π×area/perimeter²), and roundness as (4×area/[π×major_axis²]). The area, width, depth, and angle of the injection funnel were measured at mid-injection and post-injection. Images were not captured in other planes, and no 3D measurements were performed.

Figure 1.

Measurements of oocyte shape and injection funnel during microICSI. Changes in oocyte shape were measured from still images taken at three timepoints from video recorded during microICSI. (A) Pre-injection, immediately before penetration of the needle. Example measurements of oocyte area, perimeter, roundness, and circularity are shown. (B) Pre-injection, immediately before penetration of the needle. Example measurements of the x-axis and y-axis diameters. (C) Mid-injection, immediately before the breaking of the oolemma membrane. Example measurement of the injection funnel area. (D) Post-injection, immediately after withdrawal of the needle. Example measurement of residual funnel width and depth. Images from conventional ICSI were measured using the same methods (not shown). ICSI, intracytoplasmic sperm injection.

Results

No differences were observed in patient demographics between the microICSI and cICSI groups (Supplementary Table 1) or in the maturity of oocytes at the time they were discarded from clinical use (Supplementary Table 2).

1. Oocyte shape

1) Pre-injection

The average area and perimeter of oocytes immediately before needle penetration were greater in the cICSI group than in the microICSI group (Table 1). The diameter along the x-axis was also larger in the cICSI group, whereas the y-axis diameter did not differ, resulting in greater circularity in the microICSI group. In addition, oocytes in the microICSI group were more round (i.e., smoother) prior to injection. Together, these findings indicate distortion of oocyte shape caused by suction from the holding pipette, which can elongate oocytes along the x-axis. Zona thickness did not differ between groups.

Table 1.

Pre-injection oocyte shape

2) Mid-injection

During mid-injection, oocytes in the cICSI group demonstrated a greater decrease in their x-axis diameter relative to their pre-injection shape compared with oocytes in the microICSI group (Table 2). cICSI oocytes also exhibited a greater increase in y-axis diameter from pre-injection measurements than microICSI oocytes. These results suggest that microICSI caused less compression along the x-axis and less elongation along the y-axis during injection.

Table 2.

Mid-injection oocyte shape

Both groups became less circular and less round during injection, although the decrease in roundness was slightly smaller in the microICSI group, and microICSI oocytes were more round during injection than cICSI oocytes. This may be caused by continued distortion produced by pipette suction in the cICSI group, or possibly due to support provided by the cup-shaped hemisphere in the microICSI device.

Both microICSI and cICSI oocytes showed decreased area and increased perimeter compared with pre-injection shape, with no differences between groups. The depth, width, and area of the injection funnel generated during injection were also not different between the two methods (Table 3).

Table 3.

Injection funnel mid-injection

3) Post-injection

Immediately after withdrawal of the needle, oocytes in both groups showed a shorter x-axis diameter than their pre-injection shape; however, the microICSI group exhibited a greater decrease in x-axis diameter than the cICSI group (Table 4). Both groups also displayed a reduction in area compared with pre-injection, with a larger decrease observed in the microICSI group. This may indicate that microICSI oocytes experienced a slower return to their pre-injection shape.

Table 4.

Post-injection oocyte shape

The perimeter of oocytes in both groups increased slightly from pre-injection, likely reflecting the presence of a residual injection funnel. There was no difference in the number of oocytes exhibiting a residual funnel immediately after needle withdrawal (Table 5), and the depth, width, and area of residual funnels did not differ between microICSI and cICSI.

Table 5.

Residual injection funnel post-injection

Oocytes in both groups were slightly less circular post-injection compared to their pre-injection shape. Interestingly, roundness increased slightly in cICSI oocytes but decreased slightly in microICSI oocytes. Because cICSI oocytes were less round initially, the post-injection roundness did not differ significantly between methods.

There was no significant difference in the rate of lysis or degeneration following injection. One oocyte lysed immediately after microICSI due to excessive pressure during removal from the hemisphere. Two additional oocytes degenerated during culture at 0.9 hour post-injection (hpi) and 4.5 hpi; one was due to incorrect focal plane selection during injection, and the other to loss of micromanipulator control. The micromanipulator incident was unrelated to microICSI and could also have occurred during cICSI. The other two cases occured when the operator was relatively inexperienced with the device, and such events would be unlikely after further training. In the cICSI group, one oocyte degenerated at 5.5 hpi for unknown reasons, although no resistance was noted during injection.

When oocyte shape was measured again from the first image captured by the EmbryoScope+ within 30 minutes post-injection, no differences were detected between groups in area (10,229±106 µm² vs. 10,115±87 µm²), perimeter (360±1.8 µm vs. 358±1.5 µm), diameter (117±0.7 µm vs. 116±0.5 µm), circularity (0.991±0.001 vs. 0.990±0.004), or roundness (0.954±0.006 vs. 0.960±0.004).

4) Effect of the hemisphere and holding pipette on oocyte shape

A separate cohort was used to compare the shape of the same oocyte in the microICSI well (representing its natural shape), in the microICSI hemisphere, and on the cICSI holding pipette. No significant differences were detected in oocyte area, perimeter, or zona thickness among the three positions (Supplementary Table 3). However, oocytes were significantly less round compared to their natural shape when positioned on the holding pipette or in the hemisphere, although circularity remained unchanged. This corresponded to an increase in the longest diameter (major axis) when oocytes were placed on the holding pipette and a trend toward a decrease in the shortest diameter (minor axis) when positioned in the microICSI hemisphere (p=0.0507). These findings indicate that some distortion occurred in both the pipette and hemisphere compared with the natural shape, although the type and degree of deformation differed.

2. Procedure duration

The time taken to prepare the dishes was a few seconds slower for microICSI (Table 6). In all but one replicate, one dish was created for each method. Microscope set-up time was twice as fast for microICSI because no holding pipette was required. In contrast, the time needed to load and retrieve oocytes was slower for microICSI because oocytes were added to and removed from microwells individually for traceability.

Table 6.

Procedure time

Two methods were used for the microICSI procedure: aligning all oocytes at the beginning of the procedure or aligning each oocyte immediately before injection. When oocytes were aligned first, the time per injection was shorter in the microICSI group, but overall procedure duration did not differ (Table 7). When oocytes were aligned at the time of injection, neither the time required for each injection nor the total time to complete the procedure differed between microICSI and cICSI.

Table 7.

Injection time

Discussion

This study found that the microICSI device was an effective replacement for the holding pipette for ICSI with human oocytes. The rate of degeneration following injection was low and comparable to cICSI. In addition, oocyte shape was better preserved before and during injection when the microICSI device was used instead of a traditional holding pipette.

Maintaining oocyte shape is considered beneficial because it indicates lower mechanical stress and preservation of cytoskeletal structure. Deformation of oocytes can damage the spindle in mouse oocytes, which may lead to errors in chromosome segregation and subsequent aneuploidies [22,23]. ICSI involves the application of physical stress during sperm injection, and oocyte shape is often visibly distorted. Performing a sham ICSI procedure on IVF-inseminated mouse oocytes resulted in slower blastocyst development than in oocytes not subjected to sham ICSI, suggesting that the mechanical act of injection alone can alter developmental kinetics and reduce viability [24]. Furthermore, clinical ICSI procedures that require greater manipulation—and therefore more extensive cytoskeletal disruption—such as aspirating large volumes of cytoplasm or performing multiple injections, have been associated with lower fertilization rates and poorer embryo development [13,25,26]. Several recent studies have shown that ICSI techniques that produce less mechanical distortion, such as Piezo ICSI and laser-assisted ICSI, result in lower degeneration rates and improved fertilization and embryo development [27-30]. Conversely, Iwayama and Yamashita [5] reported no correlation between the degree of oocyte elongation or contraction during ICSI and fertilization or blastocyst development, although their study involved Piezo ICSI, which naturally causes less distortion than cICSI. There is a lack of published data regarding the significance of the return to circular shape after ICSI. In the current study, oocytes in the microICSI group showed a greater immediate post-injection decrease in size and x-axis diameter relative to their pre-injection shape, suggesting they returned to baseline shape more slowly than cICSI oocytes. However, after 30 minutes, no differences in shape remained between the two groups. The significance of this observation is unknown, and it is unclear whether a faster return to baseline indicates a more viable oocyte or instead reflects a compromised cytoskeletal structure. Further work is needed to determine whether the reduced distortion afforded by microICSI improves blastocyst outcomes, as has been observed in animal studies [16].

An earlier version of the microICSI device was found to reduce injection funnel depth compared with cICSI in mouse and porcine oocytes [17,31]. In the current study, however, the size of the injection funnel during injection did not differ between microICSI and cICSI, which is consistent with previous work in porcine oocytes using a later version of the device [16]. Injection funnel size has been correlated with developmental outcomes: large funnels caused by high oolemma resistance have been associated with higher degeneration rates and poorer embryo development [5,32,33]; however, very small funnels caused by low oolemma resistance have also been linked to higher degeneration rates, reduced fertilization, and lower blastocyst development [10,12,13,34-36]. These adverse outcomes in oocytes with either very large or very small funnels due to abnormal oolemma resistance are likely attributable to underlying oocyte quality [37,38] rather than the ICSI technique itself. Therefore, the significance of injection funnel size when comparing different methodologies remains uncertain and warrants further investigation in future studies of ICSI techniques.

No difference was observed between cICSI and microICSI in either the presence or size of a residual injection funnel immediately after injection. As with injection funnel size, any variation in residual funnel characteristics may be attributable to oocyte quality, ICSI technique, or a combination of these factors. Modifications to ICSI technique, including the type of injection pipette used, can influence the formation of residual funnels after cICSI; for example, larger injection funnels have been associated with higher fertilization rates [39]. Persistence of a funnel at 2 minutes has also been linked to lower degeneration rates and higher embryo quality [40]. Conversely, blastocysts derived from oocytes without a persistent funnel were associated with higher pregnancy rates [40], and a recent study found no relationship between residual funnels and fertilization or embryo development outcomes [32]. As with funnel size, the connection between residual funnels and oocyte viability remains unresolved and warrants further evaluation.

The time required to perform ICSI with the microICSI device was largely consistent with the findings of McLennan et al. [16]. Because the holding pipette was replaced by the microICSI dish, microscope and needle set-up was faster than cICSI. However, loading and unloading oocytes into the microICSI dish took longer, as each oocyte needed to be placed into and removed from the microwells with care. As a result, device design has since been modified to widen the well and hemisphere openings to facilitate faster loading and removal, meaning this aspect may improve in future studies. In contrast to McLennan et al. [16], who reported faster injection times with microICSI, injection time in this study was comparable to cICSI. Preparation of the microICSI dish was also slightly slower than cICSI. Overall, the total procedure time for microICSI was similar to that of cICSI. Because this was a research study using discarded oocytes, future work should determine whether microICSI provides workflow advantages or time savings in routine clinical IVF.

A limitation of this study is the use of sham injections with late-maturing oocytes, as it is unclear how maturation in vitro without an in vitro maturation culture system affects the oolemma. Oolemma membrane tension changes as oocytes mature [41,42], and post-ovulatory aging of mouse oocytes has been shown to decrease membrane tension [43]. Oocyte distortion during injection may also be influenced by factors not controlled in this study, such as injection speed and force [3] and patient-related characteristics, including age [44,45]. In addition, measurements of oocyte shape were based on 2D images captured in a single plane, whereas oocytes are 3D spherical structures.

Future studies should investigate the use of the microICSI device in clinical human IVF to determine whether the reduction in oocyte distortion observed here also occurs with in vivo-matured oocytes. Clinical research should assess whether the improvements seen in this study translate into higher blastocyst rates, as reported in animal models [16], or perhaps reductions in aneuploidy due to decreased spindle disruption. The microICSI device could also be combined with other techniques that reduce oocyte stress, such as Piezo ICSI, and future studies should examine whether combined approaches provide additive benefits compared with each method alone.

Supplementary materials

Article information Continued