Decoding the Follicular Fluid Microenvironment: Integrative Biomarkers and Multiomics Signatures to Predict IVF Outcomes.

The authors have nothing to report.

Extensive research on biomarkers of FF over the last decade has highlighted several consistent themes (Figure  2 ). First, FF is a rich source of potential biomarkers that reflect oocyte quality, as evidenced by numerous proteins, metabolites, and RNAs that show significant associations with fertilization, embryo development, and pregnancy outcomes. Many of these biomarkers also make biological sense. For example, higher levels of antioxidants or beneficial fatty acids in FF create a nurturing environment for the oocyte, whereas markers of cellular stress, such as cfDNA, homocysteine, and proinflammatory signals, indicate poor outcomes. Second, no single FF biomarker is perfectly predictive on its own; however, a combination may be required for robust prediction. This finding aligns with the multifactorial nature of pregnancy success. Conceptual overview of follicular fluid biomarker discovery and multiomics integration to predict ART outcomes. Multiomics approaches that integrate data from metabolites, proteins, and miRNAs represent a promising next step. With improved analytic techniques such as liquid chromatography‐MS and NGS, researchers can quantify a broad “FF fingerprint” for each follicle. Early predictive models have been explored; for example, machine learning of metabolomic profiles has identified panels of lipids that differentiate pregnant from nonpregnant IVF cycles [ 21 ]. Such models could eventually guide embryo selection; if a noninvasively collected FF sample from an aspirated follicle indicates a high‐quality oocyte or embryo, that embryo might be prioritized for transfer. To use FF biomarkers as criteria for embryo selection in this way, it is essential to collect FF from each follicle, analyze its components, and match the oocytes with their corresponding FF for evaluation. Even within the same individual, developing follicles exhibit distinct characteristics. Therefore, by analyzing each follicle individually and verifying the results against the corresponding oocyte, this approach demonstrates its true value as a biomarker at the follicular level. While the way is needed to collect oocytes in bulk rather than a single sample in order to reduce the number of punctures during oocyte retrieval, these biomarkers function effectively only when each follicle is isolated and collected. We should analyze the follicles and validate the heterogeneity of each follicle so that it can be used as a biomarker. However, translation into clinical practice is challenging. One major issue is heterogeneity. Many FF biomarkers show inconsistent results across studies owing to differences in patient populations, stimulation protocols, or assay methods. Findings observed in normal ovulatory young women may differ from those in patients with PCOS [ 103 , 104 ]. Consistently, as noted, infertility diagnoses impact FF composition; for example, PCOS follicles have altered metabolomic and cytokine profiles, and endometriosis can introduce inflammatory changes in FF [ 34 , 105 ]. Future research should stratify biomarkers according to these conditions or develop reference ranges specific to each subgroup. Standardizing FF collection and analysis is also necessary. Factors such as blood contamination of the FF, timing of the LH surge, and specific follicle size can affect the measured levels. International efforts will help ensure the comparability of results. To summarize these points, the current challenges in FF analysis include variability in sample handling and oocyte stimulation protocols, small sample sizes in omics studies, limited external validation, and dependence on different outcomes, such as embryo quality, pregnancy rate, and live birth rate. While it is difficult to standardize methods across different studies, it is essential to proceed with the evaluation of these biomarkers with a clear understanding of these limitations. Another consideration is clinical implementation. How and when should FF tests be used? One potential use case is during IVF, as the FF from each follicle is typically discarded after oocyte retrieval. Alternatively, it could be analyzed on‐site, perhaps with rapid bedside assays in the future, to generate a score for each follicle oocyte. This could aid in the selection of embryos that should be prioritized if multiple embryos are available. It might also help determine whether to freeze all embryos in suboptimal scenarios if FF biomarkers suggest low implantation potential, thereby avoiding fresh transfer. Therapeutic interventions may arise from this knowledge. For example, if low CoQ10 in FF is linked to poor outcomes, could supplementing CoQ10 before IVF improve oocyte quality? Similarly, ensuring sufficient vitamin D levels in patients might improve their FF vitamin milieu and potentially enhance pregnancy outcomes. Furthermore, if the development of nucleic acid therapeutics and related treatments proceeds smoothly, packaging miRNAs within artificial EVs and delivering them to developing follicles could potentially increase pregnancy rates. Technological innovation is already underway, and these hypotheses warrant controlled trials. It is also worth noting the potential of FF supplementation. One speculative approach is to add beneficial FF components to the culture media after oocyte retrieval. For instance, the addition of antioxidants or growth factors present in FF can mimic natural follicular conditions and enhance embryo development in vitro. Although largely experimental, this concept merges diagnostics with therapy. First, we identified the characteristics of a healthy follicle and then attempted to provide those factors to embryos from poor‐quality follicles. However, there are challenges related to establishing and standardizing workflows when introducing this technology into clinical practice. Furthermore, since ART facilities range from general hospital to clinics, further validation is required for practical implementation in clinical settings.

This work was financially supported by a Grant‐in‐Aid for Scientific Research from the Japan Society for the Promotion of Science (JSPS KAKENHI; grant numbers 24K19721), the Nitto Foundation, the Sumitomo Foundation, and the Senri Life Science Foundation.

FF steroid levels reflect intrafollicular endocrine activity and have been examined for associations with IVF success. The levels of serum estradiol (E2) are routine parameters used to monitor follicular development and oocyte maturation [ 26 ]. E2 is secreted by granulosa cells, and its concentration is high in mature follicles. Progesterone (P4) is another critical hormone secreted by the corpus luteum after ovulation [ 27 ]. Notably, P4 levels in FF were observed at higher concentrations in follicles that yielded normal fertilization, and E2 levels were higher in pregnant patients than in nonpregnant patients [ 28 ]. Several nonsteroidal hormones have been studied in FF. The anti‐Müllerian hormone (AMH), produced by granulosa cells, has drawn significant interest. AMH is a well‐known serum marker of ovarian reserve, but its intrafollicular level may indicate granulosa cell viability and oocyte quality [ 29 , 30 ]. A recent meta‐analysis revealed significantly higher mean AMH levels in the FF of pregnant women than in those of nonpregnant women after assisted reproductive technology (ART) [ 31 ]. This finding suggests that FF AMH could serve as a positive prognostic marker for oocyte developmental potential; however, the analyses showed substantial heterogeneity. One study found that FF AMH could predict blastocyst development [ 32 ], whereas another study found no significant correlation with IVF outcomes [ 33 , 34 ]; therefore, FF AMH is not ready for clinical use until further standardized research is conducted. We summarized the biomarkers in FF for IVF associated with hormonal factors that have been reported (Table  1 ). Hormonal predicting biomarkers in FF for IVF outcomes. In addition to hormones, various proteins and growth factors in FF have been evaluated for their predictive value. The follicle is a site of intense protein secretion and enzymatic activity during oocyte maturation. Therefore, protein markers can reflect processes such as inflammation, tissue remodeling, and nutrient transport. Matrix‐remodeling enzymes, such as matrix metalloproteinases (MMPs), have been studied. MMPs are zinc‐ and calcium‐dependent proteolytic enzymes that play roles in tissue remodeling and organization of the extracellular matrix during folliculogenesis [ 35 ]. MMP‐2 and MMP‐9 are members of the gelatinase group that degrade gelatin and collagen [ 36 ]. Gelatinase activity has been found in theca cells of growing follicles and in the stroma of the ovary [ 37 ]. These FF levels showed inconsistent correlations: the study by Atabakhsh et al. [ 38 ] found no correlation between MMP‐2 activity and the number of mature oocytes or fertilization rate; however, a positive correlation was observed between MMP‐2 activity and oocyte and embryo quality. Another group found that higher MMP‐9 levels in FF predicted pregnancy, a finding that was also confirmed in serum as well as in FF [ 39 ]. FF contains numerous growth factors critical for follicle and oocyte development. Bone morphogenetic proteins (BMPs) and growth differentiation factors (GDFs) are oocyte‐secreted factors that are critical for follicular development and female reproductive capacity [ 40 ]. BMP‐4 and BMP‐7 are produced by ovarian stromal and theca cells and are thought to enhance the transition from primordial to primary follicles [ 41 ]. BMP‐4 expression correlates with increases in the size of primary and secondary follicles, as well as with an increase in the number of primary follicles [ 42 ]. Takmaz et al. [ 43 ] reported that higher BMP‐4 levels in FF enabled a more accurate prediction of pregnancy (area under the curve [AUC] = 0.61). BMP‐15 is another BMP sub‐family member secreted by oocytes and is a key regulator of oocyte maturation [ 44 ]. BMP‐15 and GDF‐9 in FF play crucial roles in oocyte maturation; however, reports indicate that neither BMP‐15 nor GDF‐9 is a useful biomarker for predicting pregnancy [ 34 ]. Stem cell factor (SCF) is a paracrine and autocrine growth factor produced by granulosa cells during follicular development [ 45 ]. One study showed that SCF levels in FF predicted oocyte maturity [ 46 ], and another demonstrated that SCF levels in FF could predict both oocyte maturity (AUC = 0.69) and pregnancy (AUC = 0.7) [ 47 ]. Overall, individual proteins and growth factors showed modest predictive power. High‐throughput proteomics and peptidomics have expanded FF biomarker discovery beyond single‐analyte studies. Several mass spectrometry‐based studies comparing FF from cycles with favorable versus unfavorable outcomes have repeatedly implicated pathways related to complement/coagulation, extracellular matrix remodeling, and cell–cell/vascular adhesion, supporting the concept that coordinated inflammatory resolution and tissue remodeling underpin oocyte competence [ 48 , 49 , 50 , 51 ]. More recent proteomic datasets emphasize extracellular matrix regulation as a recurrent signature associated with oocyte quality and embryonic developmental potential [ 52 , 53 ]. Importantly, proteomic discovery has also moved toward clinically feasible formats, including targeted peptidomics and peptide “barcodes,” which may reduce sample input and analytical burden [ 54 , 55 ]. These findings collectively favor multivariate panels; however, the proteome is highly sensitive to stimulation protocols, blood contamination, and inter‐follicular variability, making rigorous preanalytical standardization and external validation prerequisites for clinical adoption. We summarized the biomarkers in FF for IVF associated with proteins and growth factors that have been reported (Table  2 ). Predicting biomarkers in FF for IVF outcomes by proteins and growth factors. Ovarian follicles exhibit immune and inflammatory activities that are critical for ovulation and are subject to oxidative stress. These factors can influence oocyte quality, and researchers have measured various cytokines and redox markers in FF. One notable finding was the presence of interleukin‐6 (IL‐6) among these cytokines. IL‐6 is a pro‐inflammatory cytokine produced by follicular and immune cells that plays a role in the ovulatory cascade [ 56 ]. Yang et al. [ 57 ] found significantly higher IL‐6 levels in FF from women who achieved pregnancy than in those who did not ( p  = 0.016). Moreover, higher IL‐6 levels were associated with better embryo development, as they correlated with lower embryo fragmentation, implying higher embryo quality. This result is notable, as excessive inflammation is typically expected to be detrimental. However, IL‐6 may reflect a normal inflammatory milieu necessary for follicle rupture and subsequent luteinization that supports early embryo development. Another study demonstrated that higher IL‐6 levels in FF predicted pregnancy [ 56 ]. IL‐6 levels in FF are significantly higher than those in serum [ 58 ], indicating that IL‐6 levels are a good prognostic marker for folliculogenesis. Conversely, monocyte chemotactic protein‐1 (MCP‐1), another proinflammatory cytokine, was not correlated with pregnancy outcomes in FF [ 59 ]. In humans, physiological levels of MCP‐1 are transiently elevated in periovulatory FF and ovarian stromal cells during the ovulatory process [ 60 ]. MCP‐1 plays an important role in ovulation; however, it does not appear to be a marker for predicting pregnancy. Earlier studies on cytokines such as IL‐8 and IL‐12 have shown controversial results. IL‐12 in FF is associated with lower pregnancy chances, whereas IL‐8 generally shows no clear association or exhibits a slightly positive trend in some reports [ 61 , 62 ]. Overall, IL‐6 appeared to be a consistently positive marker among inflammatory cytokines during normal ovulation, whereas other cytokines have not yielded reliable predictive value. In the oxidative stress domain, multiple studies have evaluated global redox capacity and oxidative damage markers in FF. Overall, evidence suggests that a more favorable antioxidant milieu is associated with improved oocyte and embryo competence, although thresholds and effects may vary by patient population and assay platform [ 63 , 64 , 65 , 66 , 67 , 68 ]. Methodologically, this heterogeneity underscores the value of pairing redox markers with mitochondrial function surrogates and/or omics features rather than relying on single oxidative readouts in isolation. The follicular environment requires a balance between reactive oxygen species (ROS) and antioxidants for proper oocyte maturation. Excess oxidative stress can damage cells and deoxyribonucleic acid (DNA), whereas ROS play a role in ovulation signaling. Coenzyme Q10 (CoQ10) is a mitochondrial electron transport chain component and antioxidant that prevents free radical formation [ 69 ]. This has emerged as a meaningful marker in FF. One study showed that higher FF CoQ10 levels were found in embryos that developed into high‐quality blastocysts compared with poor‐quality embryos, and that FF CoQ10 levels were significantly higher in women who became pregnant than in those who did not [ 70 ]. This suggests that higher antioxidant levels in follicles may reflect better mitochondrial function in the oocyte and support embryonic development. Conversely, homocysteine is a prooxidant amino acid that induces oxidative stress [ 71 ]. In FF, high homocysteine levels were associated with negative outcomes, and a subgroup analysis found that FF homocysteine levels were significantly higher in women who did not become pregnant than in those who did [ 72 ]. Elevated homocysteine levels can impair follicular vascularization and oocyte development by promoting free radical damage. Clinically, this suggests that antioxidant support may improve follicular conditions. We summarized the inflammatory and oxidative stress biomarkers in FF for IVF that have been reported (Table  3 ). Predicting biomarkers in FF for IVF outcomes by inflammatory and oxidative stress markers. Metabolomics has allowed the simultaneous assessment of many small molecules in FF, revealing metabolic signatures linked to pregnancy success. Several studies have profiled amino acids, vitamins, and lipids in FF, often identifying differences between the profiles of successful and unsuccessful IVF cycles [ 73 , 74 ]. Amino acids in the FF can serve as nutrients and signaling molecules. Notably, higher levels of certain amino acids in FF, including phenylalanine, leucine, and tryptophan, were significantly correlated with lower miscarriage rates [ 75 ]. Conversely, women whose follicles had higher concentrations of these amino acids were more likely to achieve successful pregnancies, suggesting that these nutrients may support embryo viability or reflect healthier follicular metabolism. Vitamins in FF originate from the diet and follicular transport, and some have been linked to IVF outcomes. Vitamin D has recently attracted considerable attention. An analysis by Ekapatria et al. [ 76 ] showed that higher FF 25‐OH vitamin D levels were associated with better oocyte maturity and quality, as well as a trend toward higher fertilization rates in vitamin D‐replete follicles. Another study showed that higher levels of 24,25(OH) 2 D 3 and a low 1,25(OH) 2 D 3 /24,25(OH) 2 D 3 ratio were associated with increased live birth rates [ 77 ]. These data suggest that sufficient vitamin D in FF enhances oocyte competence via genomic effects on granulosa cells or through calcium regulation in the oocytes. However, vitamin D levels in the FF did not predict pregnancy in another study [ 78 ]. This suggests that while vitamin levels might influence oocyte/embryo development, they may not individually determine whether pregnancy occurs. Lipid metabolites in FF are a major focus of metabolomics because lipids serve as energy sources, membrane components, and signaling molecules. FF lipid composition differs from that of serum and can indicate follicular metabolic status [ 74 ]. Phospholipids are the most abundant lipids in the cell membrane. Lysophosphatidylcholines (LysoPCs) are membrane lipid derivatives that have emerged as intriguing markers [ 79 ]. One study reported that excessively high LysoPC levels in FF were associated with higher miscarriage rates [ 75 ]. This suggests that excessive LysoPC levels indicate pathological stress or inflammation within the follicle, which may subsequently impair pregnancy. The study also revealed that higher FF levels of docosahexaenoic acid (omega‐3), linoleate (omega‐6), and oleate (omega‐9) were significantly associated with lower miscarriage rates [ 75 ]. These unsaturated fatty acids are critical components of cell membranes and may improve oocyte and embryonic physiology. This finding aligns with the general health literature, which shows that omega‐3 and omega‐6 fatty acids are important for reproductive success. Recent metabolomic studies have suggested that FF metabolic fingerprints encode both diagnosis‐specific biology and cycle‐specific readiness, including alterations in amino acids, lipids, and energy‐related metabolites. A systematic review highlighted that despite promising signals, study‐level heterogeneity in sampling, analytical platforms, and outcome definitions remains a major barrier to generalizable biomarkers [ 80 ], and subsequent critical appraisal has echoed the need for prospective external validation and clinically meaningful endpoints such as live birth [ 81 ]. Within this context, obesity and ovarian reserve phenotypes have been linked to distinct FF metabolomic profiles [ 82 ], and emerging lipidomics studies further indicate that specific lipid classes may stratify polycystic ovary syndrome (PCOS) phenotypes and correlate with embryo development [ 83 , 84 , 85 , 86 ]. We summarized the metabolic biomarkers in FF for IVF that have been reported (Table  4 ). Predicting biomarkers in FF for IVF outcomes by metabolomics. In addition to proteins and metabolites, FF contains genetic material that can serve as biomarkers. Two main sources have been investigated: cell‐free DNA (cfDNA) released from follicular cells and ribonucleic acids (RNAs), particularly micro RNAs (miRNAs), which are often carried in extracellular vesicles (EVs). These biomarkers can reflect cell death or regulatory processes in the follicles. Traver et al. [ 87 ] highlighted cfDNA in FF as a prognostic biomarker of IVF outcomes. They quantified DNA fragments in FF from 117 IVF patients undergoing IVF. Notably, they found that FF cfDNA levels were an independent and significant predictive factor for pregnancy outcomes (adjusted odds ratio = 0.69). The FF cfDNA cutoff value yielded an AUC of 0.73 for predicting clinical pregnancy. These data suggest that elevated levels of DNA fragments, likely from apoptotic granulosa and cumulus cells, indicate a less favorable follicular environment. Excessive cell death within a follicle can compromise oocyte quality or subsequent luteal support. FF cfDNA reflects follicular cell apoptosis and stress, and a low cfDNA level is a hallmark of healthy follicles with high pregnancy potential. This concept has been echoed in subsequent studies and reviews [ 88 ]. Although not yet used clinically, measuring FF cfDNA during oocyte retrieval could help identify embryos with a better prognosis without directly manipulating the embryo itself. Future studies should explore whether these interventions modulate follicular cfDNA levels and improve patient outcomes. Beyond concentration‐based cfDNA quantification, qualitative features such as cfDNA integrity and mitochondrial DNA abundance have been proposed as additional indicators of follicular health and oxidative injury, potentially capturing the apoptosis/necrosis balance within the follicle [ 89 , 90 ]. Similarly, broader miRNA profiling studies in FF support the feasibility of noncoding RNA panels as biomarkers, and recent syntheses have highlighted extracellular miRNAs as a particularly attractive class because they are stable and amenable to multiplex assays [ 91 , 92 ]. miRNAs are small noncoding RNAs that regulate gene expression, and they have emerged as potent biomarkers in reproductive biology [ 93 ]. In FF, miRNAs can be freely packaged into EVs. Cumulus and granulosa cells, as well as oocytes, secrete EVs that accumulate in FF, carrying miRNAs that reflect the physiological state of the follicle [ 94 ]. EVs are nanosized, membrane‐bound vesicles secreted by almost all cell types that carry several RNA species, including miRNAs [ 95 ]. Small noncoding RNAs in EVs have recently been identified as a mechanism of intercellular communication and as a source of biomarkers. Recently, we applied small RNA sequencing to FF‐derived EVs and identified a miRNA “triplet” that robustly predicts pregnancy [ 96 ]. We analyzed EVs from individual follicles and discovered 14 dysregulated small RNAs that differentiated pregnant from nonpregnant outcomes, then homed in on a combination of three miRNAs (miR‐16‐2‐3p, miR‐378a‐3p, and miR‐483‐5p) with a remarkable AUC of 0.96. Notably, miR‐378a‐3p and miR‐483‐5p have been implicated in embryo development in other spent media and in granulosa cells [ 97 , 98 ]. Moreover, we performed functional analyses to reveal that the three types of miRNAs may be associated with both embryo quality and follicular development. Our study concluded that these miRNAs in FF EVs are promising noninvasive biomarkers of pregnancy potential, awaiting validation in larger cohorts. The use of EVs is advantageous because they protect miRNAs from degradation, can be reliably isolated, and originate from key follicular cells, carrying signals of cell function. Parallel evidence indicates that EVs in FF vary with age and infertility, further supporting the utility of our biomarker. In addition to miRNAs, researchers have noted that p‐element‐induced wimpy testis (PIWI)‐interacting RNAs (piRNAs) are major types of small noncoding RNAs that are 21–35 nucleotides long and function together with PIWI proteins [ 99 , 100 ]. piRNAs are involved in transposon regulation in the germline, although their relevance to pregnancy remains unclear. Messenger RNA from cumulus cells can also be detected in FF but is not a routine biomarker, and one could envision a transcriptomic signature of cumulus cell health in FF [ 101 ]. For instance, a transcriptomic study of granulosa cells from FF found that distinct gene expression patterns correlated with embryo quality [ 102 ]. However, because our focus was strictly on FF, miRNAs and cfDNA were the primary nucleic acid biomarkers studied. We summarized the biomarkers in FF for IVF associated with genomic and transcriptomic factors including our previous study (Table  5 ), and also other types of predicting biomarkers in Table  6 . Predicting biomarkers in FF for IVF outcomes by genomic and transcriptomic factors. Other types of predicting biomarkers in FF for IVF outcomes.

FF is a complex mixture of metabolites, proteins, nucleic acids, and cells that reflects both systemic physiology and local ovarian processes (Figure  1 ). It originates from blood plasma filtrate and cellular secretions within the follicle [ 18 , 23 ]. FF provides a microenvironment for the oocyte, facilitating bidirectional communication between the oocyte and surrounding cumulus and granulosa cells, which are essential for oocyte maturation. Molecules in the FF influence critical pathways for folliculogenesis and ovulation, thereby affecting the developmental competence of oocytes. Studies have shown that variations in the FF composition are associated with the developmental potential of oocytes. Therefore, FF has been increasingly recognized as an indicator of the metabolic and endocrine states of follicles. Key components of FF include steroid hormones, peptide hormones, growth factors, cytokines, chemokines, metabolites, and nucleic acids [ 24 ]. This molecular milieu not only supports the oocyte but also contains potential biomarkers of oocyte quality. Novel analytical techniques enable detailed profiling of FF constituents, paving the way for biomarker discovery [ 25 ]. In the following sections, we discuss the major categories of FF biomarkers and their associations with IVF outcomes, with a focus on human studies. Follicular fluid liquid biopsy.

In conclusion, the use of FF biomarkers in human reproductive medicine has expanded significantly over the last decade. We now have candidate biomarkers, ranging from basic science to clinical trials that illustrate an ideal follicular environment. FF, once considered a mere by‐product, is now recognized as a treasure trove of information reflecting oocyte competence and pregnancy potential. The challenge is to integrate this information into predictive tools and validate them in large prospective studies. As technology and understanding advance, it is plausible that the next generation of IVF will incorporate FF‐based assays to personalize and improve embryo selection, ultimately increasing success rates while minimizing the physical and emotional burdens on patients. One day, reproductive clinicians may use a simple FF biomarker panel encompassing hormones, metabolites, and miRNAs to reliably forecast which embryo is most likely to result in a healthy pregnancy, fulfilling a long‐sought goal in reproductive medicine.

Infertility is diagnosed when a couple fails to achieve pregnancy after 12 months or more of regular unprotected sexual intercourse. Infertility is a global health concern and affects approximately 15% of couples of child‐bearing age, as reported by the National Institute of Child Health and Human Development [ 1 ]. Assisted reproductive technologies (ART), such as in vitro fertilization (IVF) and embryo transfer (ET), are widely used to help infertile couples conceive and have dramatically increased the pregnancy rates to approximately 20%–35% [ 2 , 3 , 4 ]. Unfortunately, despite technical advances and well‐established medical and surgical techniques used in IVF, clinicians still lack reliable molecular biomarkers to predict ART success. One method of embryo selection for ET is based on the morphological grading of embryos or blastocysts. Generally, cleavage‐stage embryos are assessed using the Veeck classification on day 2 or 3, and blastocysts are using the Gardner classification on day 5 or 6, considering the number of cleavages or blastomeres and the rate of fragmentation [ 5 , 6 ]. However, this morphological scoring system relies on the experience of embryologists, making it a subjective method. Therefore, the morphological scoring system alone cannot indicate the success rate of IVF [ 7 ]. Even embryos that appear morphologically normal may fail to implant or result in miscarriage. Time‐lapse imaging is another technology for embryo selection that uses artificial intelligence (AI) to learn comprehensively from large amounts of image data [ 8 , 9 ]. However, regardless of the presence of embryo selection using AI software, time‐lapse technology has failed to demonstrate an increase in pregnancy and live birth rates [ 10 , 11 ]. Time‐lapse imaging is essentially limited to observing morphological changes and does not directly evaluate molecular or biological integrity within the embryo. In particular, preimplantation genetic testing for aneuploidy (PGT‐A) is increasingly applied to distinguish aneuploid embryos from euploid embryos to increase pregnancy rates [ 12 ]. However, a recent meta‐analysis revealed that PGT‐A performed on blastocysts improved clinical outcomes only in women aged 35 years or older, showing no effect in the general population [ 13 ]. Furthermore, PGT‐A requires invasive biopsy, which may affect embryo viability. It can also yield results requiring further consideration, such as the management of mosaic embryos [ 14 , 15 ]. Therefore, high demand exists for new, noninvasive, and easily implementable embryo selection methods. Follicular fluid (FF) is the fluid surrounding the oocyte within the ovarian follicle and provides a noninvasive means of observing the immediate environment of the oocyte [ 16 , 17 ]. FF provides a complex microenvironment for oocyte maturation and contains several molecules secreted from the oocyte, granulosa, cumulus, and theca cells [ 18 , 19 ]. The cells lining these follicles provide hormones, essential nutrients, and signaling molecules for oocyte maturation. Importantly, FF can be readily collected during oocyte retrieval in IVF with minimal risk to the oocyte, making it an attractive source of biomarkers [ 20 ]. Over the past decade, research has increasingly focused on the FF constituents as predictors of oocyte quality, embryo viability, and pregnancy outcomes. Earlier studies examined single hormones or proteins, but recent efforts employ high‐throughput “omics” techniques (e.g., proteomics, metabolomics, and transcriptomics) to capture a holistic signature of the follicular microenvironment [ 21 , 22 ]. Over the past decade, advances in highly sensitive analytical technologies such as mass spectrometry (MS) and next‐generation sequencing (NGS) have enabled the comprehensive analysis of hormones, proteins, metabolites, lipids, and nucleic acid components such as micro ribonucleic acid (miRNA) in FF. Furthermore, rapid advancements in the application of AI and machine learning have accelerated efforts to integrate vast and complex multiomics data and to construct clinical prognostic prediction models. In this review, we examine the current understanding of FF‐based biomarkers, highlighting key discoveries from the last 10 years across all major omics, and discuss how these markers may improve prognostication in reproductive medicine. This report comprehensively evaluates the utility of diverse biomarkers in FF based on the latest research findings from the past decade. It details how FF analysis, ranging from conventional markers such as hormones and proteins to integrated AI‐based analysis, has the potential to transform the future of reproductive medicine.

The authors declare no conflicts of interest.