Regulation of follicular development, ovulation and luteal function is complex and multifactorial (61). Careful characterization of ovulation is done by frequently-timed transvaginal ultrasonography to measure follicular development and rupture and corpus luteum development, coupled with measuring blood levels of progesterone as well as estradiol, LH, and FSH. A six-point grading system was developed by Hoogland and Skouby in the early 1990s to characterize ovarian activity from “no activity” to “ovulation” based on follicular rupture, plus estradiol and progesterone, and has been widely used (24). This approach is sometimes modified by investigators based on research objectives. For example, research to quantify ovarian activity and ovulation among combined or progestin-only pill users have considered stratified luteal phase progesterone levels in the algorithm i.e., whether progesterone levels following rupture of a lead follicle were considered adequate to confer a theoretical risk of pregnancy (62–64). While evaluating multiple factors involved in the ovulatory process has been the gold standard for concluding an ovulation has occurred and for assessing whether it is functional, the demand on both participants and researchers is great. More commonly, presumptive ovulation has been assessed without using ultrasound, measuring only serum or urinary hormones i.e., progesterone (or progesterone plus estradiol) (65–68), urinary pregnanediol—which has been shown to reflect serum progesterone (69, 70), and sometimes including LH and FSH levels (64, 71, 72).

With regard to using progesterone as a biomarker of ovulation, detailed assessments of daily serum levels across the entire menstrual cycle have been published (69, 73) and the range of luteal phase progesterone values considered to be “normal” after ovulation (determined by observation of an LH peak) classified based on median, 5th and 95th percentile values. Many studies in the literature use more limited sampling, however, and the varied timing and frequency of sample collection as well as the differing cut-off levels of progesterone used to conclude an “ovulation” has occurred make precise comparisons difficult. Some definitions will lead to more liberal determinations when classifying a presumptive ovulation i.e., those that use a single low progesterone value. A recent study documented the variability in concluding a presumed ovulation when using several different progesterone level cut-offs (68). Finally, progesterone values on the low end of normal that have been considered “ovulations” in many studies e.g., near the 5th percentile, may not be functional ovulations indicative of “fertile cycles”. A recent review summarized the challenges with using progesterone alone when classifying normal or deficient luteal phases (74). However, research on pregnancy success following timed intercourse or intrauterine insemination reported that midluteal levels of progesterone were significantly lower in cycles where pregnancy was not achieved compared with cycles where pregnancy was achieved (75, 76).

With all hormonal methods, wide interindividual variability in steroid levels and responses are observed, and understanding is limited with regard to the impact of this variability on method-specific contraceptive effectiveness and side effects (65, 77–80). Pharmacogenomic differences, BMI differences, and ethnicity among other factors have been shown to influence pharmacokinetics and pharmacodynamics (81, 82). Evidence of the influence of stress on menstrual cycles and ovulation is also accumulating (83, 84). In addition, one study reported that the prevalence of luteal phase deficiency (low progesterone and shorten luteal phase) is more common in women with obesity during natural ovulatory cycles (85), adding further complexity to evaluating method-specific RTF based on ovulation (particularly progesterone alone) as a biomarker or surrogate marker.

Furthermore, fairly recent large population studies showed that about 1/3 of cycles with stated normal length (21–35 days) have subclinical ovulatory disturbances with low progesterone (68, 86). The authors speculate that anovulation periodically occurs in all or most women and note the need for more population-based studies to assess within-woman variability over time (68). With this as a backdrop, it may be challenging when using ovulation as a biomarker to determine precisely whether an observed delay in RTF following discontinuation of a method is truly method-related, or inherent in some variability of individual menstrual cycles.

Some highly effective hormonal methods, e.g., certain progestin-only pills (POPs), levonorgestrel implants, and the hormonal IUD, only partially suppress ovulation. Additional effects include altering the regulation of follicular development leading to anovulation, abnormal ovulation, inadequate luteal phase or a luteinized unruptured follicle (64, 66, 67, 71, 87, 88), reduced endometrial thickness (89), and progestin-induced thickening of cervical mucus causing it to become impenetrable to sperm (66, 88, 90–92). Cervical mucus quality is typically assessed using a modified Insler score as well as a sperm penetration assay (93). A review on the role of cervical mucus and contraception discusses the methodological limitations with these assays (91). However, a recent well-designed study of a levonorgestrel POP showed that cervical mucus quality did not favor fertility in users of this POP, even when a cycle was classified as ovulatory (88). This study is consistent with earlier studies on levonorgestrel implants (66) and helps to clarify the understanding of the important role of cervical mucus in the mechanism of progestin-only methods as well as its potential to influence RTF following use of certain progestin-only contraceptive approaches.

For methods where discontinuation of use leads to a rapid elimination of the circulating progestin (e.g., POPs, implants, hormonal IUD), the ability to conceive has been shown to resume without a lingering effect (61, 94). In contrast, injectable contraceptives that form drug depots e.g., medroxyprogesterone acetate or norethindrone enanthate, have a more variable and delayed return to ovulation and to fertility at the end of the period of high effectiveness (22, 65, 78) due to the continued impact of low, circulating levels of progestins and the mechanisms described above (66, 67, 71, 87, 88). This reinforces the concern that using only progesterone levels as a biomarker for putative ovulation may overestimate fertility potential if residual contraceptive progestin serum levels are detectable e.g., with injectable, non-removable approaches (22).

The contraception field can benefit by considering knowledge gained through decades of research related to infertility and perhaps leverage data on predictors of successful pregnancy following in vitro fertilization. A number of factors have been considered as biomarkers of ovarian reserves and as predictors of fertility or potential success in infertility treatment. These include antral follicle count (AFC), anti-Mullerian hormone (AMH), FSH, and InhibinB (95, 96). Whether any of these could potentially be used alone or combined with assessments of ovulation to more accurately predict RTF remains to be determined.

One study evaluated whether the use of long-term COC is associated with AFC suppression in patients seeking fertility preservation (FP) before cancer treatment (97). They reported that long-term use of COC is strongly associated with reversible AFC suppression, which returns to normal by 6–7 months after COC stoppage. This study agreed with other studies that found not only low AFC but also low AMH and ovarian volume during COC use (98, 99). However, greater precision in the timing of return to normal would be necessary to make these indicators useful. In addition, AFC and ovarian volume are determined through use of ultrasound that is more difficult than measuring a serum hormone such as AMH.

AMH is expressed by granulosa cells of growing follicles, is regulated by LH, FSH, estradiol and bone morphogenic proteins, and in turn regulates folliculogenesis through a variety of target genes (100). AMH is a marker of ovarian reserve that peaks around age 16, remains stable through the mid-20s and then declines with age (an inverse correlation with age). AMH levels have been shown to vary across the menstrual cycle; however, while statistically significant (with highest levels in the mid-follicular phase), changes across the cycle were small and not considered clinically important as related to timing of sampling for diagnostic purposes (100, 101). Investigators did note, however, that AMH levels were slightly higher in anovulatory cycles than in ovulatory cycles. Many studies have documented the inhibitory effect of estrogens on AMH during follicular growth, including one evaluation of the impact of continuous use of three types of estrogen-containing contraceptives (OCs, a vaginal ring and a patch) that noted significant declines (approximately 50%) in AMH after nine weeks of use in all groups (102). More recently, a retrospective cohort study assessed the impact of various contraceptives including COCs, a combined vaginal ring, POPs, the levonorgestrel IUD and non-hormonal methods on AMH and antral follicle count and reported decreases in AMH by COCs and POPs, but not the other methods (103). Another showed AMH levels to be markedly reduced among current users of COCs, a combination vaginal ring and DMPA, but not among current users of the levonorgestrel IUD (104). There was no significant association between AMH and time since last use for COC users, but AMH levels were lower among individuals who had used DMPA within the last year when compared to never users. Given DMPA's reduction in AMH in the study, the latter difference might be explained by the extended time for MPA levels to return to undetectable following a user's last injection (22, 78).

Several important factors need to be considered in assessing whether AMH might be used as an indicator of RTF, or as a component of other algorithms of RTF. These include the impact of various other influencers of AMH levels including age, race or ethnicity, BMI, and environmental or other factors that might confound result interpretation (100, 105). In addition, better information on the time course to return to a previous baseline, for example after COC discontinuation, would be necessary to assess operational questions such as whether an individual would need a baseline sample before initiating contraception in order to accurately assess recovery (RTF). Finally, in a prospective cohort study of 981 women without a history of infertility who were trying to conceive, the probability of conception was the same for women with low and normal serum AMH levels and the authors concluded that their findings do not support use of AMH to assess natural fertility or reproductive potential (106).

Another hormone that has been used to predict fertility potential in in vitro fertilization (IVF) programs is the pituitary hormone FSH. During a normal menstrual cycle, FSH levels rise during menses and stimulate granulosa cell production of estradiol and Inhibin B. Driven by the negative feedback of increasing estradiol and Inhibin B, FSH levels then decline by the mid-follicular phase (61). An FSH level on cycle day 3 has been considered a marker of ovarian reserves and has been a widely used test in the clinical evaluation of infertility. However, a review in 1999 concluded that “the evidence does not support the use of these tests for widespread screening of women for fertility potential” and “will result in an unacceptably large percentage of false-positive test results” if used in the general population (107). A 2006 systematic review similarly concluded that measurement of basal FSH in the early follicular phase has a limited capacity to predict IVF outcome, and therefore should not be routinely used for this purpose (95). More recently, in the prospective study of Steiner and colleagues noted above (106), women with high serum or urinary FSH (a potential indication of reduced ovarian reserves) did not have a significantly different probability of conceiving after 6 or 12 attempted cycles compared with women who had normal FSH values. They concluded that their data do not support the use of FSH to assess natural fertility. It follows, therefore, that an early follicular phase FSH level would not be a sensitive potential biomarker of RTF after discontinuation of contraceptives that inhibit FSH as part of their mechanism of action (contraceptives with an estrogen component).

Inhibin B belongs to the superfamily of transforming growth factor-B. It is secreted by granulosa cells and provides negative feedback at the level of the pituitary to inhibit FSH secretion during the normal menstrual cycle (108). Inhibin B is also considered one of the indicators of ovarian reserves that is used in the context of infertility assessment and treatment. As women age, Inhibin B levels gradually decline leading to progressive increases in basal FSH. While a number of studies have evaluated the use of basal Inhibin B, alone or in combination with other ovarian reserve indicators, as a predictor of success in infertility treatment, one systematic review and meta-analysis concluded that the predictive value is modest to poor (95). Although a recent study concluded that Inhibin B may have an advantage over FSH in predicting ovarian response to infertility treatments (109), another in a cohort of women without infertility noted above (106) concluded that Inhibin B levels were not associated with the probability of conceiving in a given cycle. Based on the composite of data, as with AMH and FSH, the use of Inhibin B as a potential biomarker of RTF following contraceptive use would not be recommended.

Contraceptive methods affecting the process of folliculogenesis, and RTF after discontinuation of methods thereof, may potentially be evaluated by markers associated with the development of the preovulatory follicle. There are a number of molecules involved in the maturation of the ovarian follicle that are readily obtainable and quantifiable, and thus may serve as biomarkers to be analyzed in contraceptive studies. Prime amongst them are relaxins and, in particular, insulin-like peptide 3 (INSL3). Relaxins are a family of peptide hormones known as neo-hormones, which are a class of paracrine and endocrine adaptations that relate to specific mammalian physiological functions, and are intimately involved with viviparity, placental development, and implantation, with their secretion typically occurring via the corpus luteum, ovarian follicles, and decidua (170). INSL3 has been shown to vary considerably across the menstrual cycle with episodic peaks in follicular and sometimes luteal phases, and is known to drop to undetectable levels in postmenopausal women (171). Relaxin-3 expression has been reported within oocytes in mid to late follicular stages of development, whereas receptor staining was localized to follicular cells (172). Further, it has been suggested that relaxin peptides may be involved in estradiol-dependent events in follicular development (172). Relaxin expression is also modified by medroxyprogesterone, RU-486, and glucocorticoids (173). H2-relaxin is another peptide hormone known specifically as a paracrine factor. Although H-2 relaxin has been identified primarily in the endometrium and corpus luteum, its gene is also constitutively expressed when granulosa cells have attained luteinized status under the effect of LH or hCG (174), making H2-relaxin a potential biomarker for the differentiation status of granulosa cells and thus follicular development. Additionally, oral contraceptives have been found to reduce H2-relaxin serum levels (175).

INSL3 specifically is an insulin-like peptide that is also part of the relaxin-like peptide family, and while it has been primarily considered a male hormone, it is also synthesized within the ovary. It is in part responsible for meiotic progression of oocytes in preovulatory follicles by increasing cAMP concentration through G protein-coupled receptor 8 (176). As a reproductive biomarker, INSL3 is relatively stable across the cycle of premenopausal women and very low or undetectable in postmenopausal women. Inverse correlation between INSL3 expression in theca cells of the antral follicles and atresia of granulosa cells has been observed (174). Of note sub-fertility and reduction in follicle numbers have been demonstrated in INSL3 knockout mice (177). INSL3 is quantifiable in serum and its levels have been associated with ovulatory and anovulatory cycles in adolescents (178).

Another potential biomarker of folliculogenesis is vascular endothelial growth factor (VEGF). Secretion of VEGF from granulosa and theca cells is correlated with increased angiogenesis of the ovarian follicle, increased vascularization coinciding with increased oocyte viability, and favorable pregnancy outcomes (179). Knockout of VEGF in female mice displayed not only decreased fertility and ovary size but a reduction in follicular development (180).

Another family of potential molecules of interest is the gap junction proteins called connexins. Cx43, for example, is associated with folliculogenesis and demonstrated to be elevated in growing granulosa cells but mostly absent in the corpus luteum and follicles undergoing atresia. A marked decrease in expression takes place following the LH surge in murine proestrus only in preovulatory follicles and associated cumulus cells, but not in more immature preantral follicles, establishing expression of Cx43 as a regulator and possible biomarker for follicular growth and development (181, 182).

Present in the hypothalamus and acting upon the hypothalamic-pituitary-ovarian axis, Kisspeptin (KISS1) is a neuropeptide lying upstream of gonadotropin-releasing hormone as an ovulatory trigger and has been experimentally identified to be associated with follicular development and oocyte maturation (183). Expression of KISS1 and its receptor have also been identified in human ovaries, including in theca cells, corpus luteum, cumulus cells, mural granulosa cells, and interstitial tissues and ovarian surface epithelium (184, 185). Conditional knockout of KISS1 and its receptor, GPR54, in mice leads to infertility and hypogonadotropic hypogonadism in addition to other maturation and reproductive system dysfunctions (186, 187).

Two members of the transforming growth factor-beta (TGFβ) superfamily highly expressed in oocytes and thought to affect oocyte function and development, GDF-9 and BMP-15, are promising candidates as well. Determination of GDF-9 levels in follicular fluid by Western blot analysis correlated higher levels with the nuclear maturation of oocytes via identification of the first polar body (188). GDF-9 mRNA expression has also been positively associated with oocyte maturation as well as fertilization and cleavage rates (189). BMP-15 is known to induce mitosis and proliferation in granulosa cells (188). Evaluation of BMP-15 levels in follicular fluid correlated higher concentrations with higher fertilization rate, cleavage, and embryo quality as defined by favorable morphology (189, 190). In knockout mice of both of these TGFβ members, subfertility and decreased ovulation have been demonstrated, but in GDF-9 knockouts specifically, cessation of folliculogenesis was also observed (191).

Endocannabinoids are lipid-signalling neurotransmitters with a broad dispersion and activity throughout the body. One specific endocannabinoid, anandamide, has been quantified in follicular fluid following oocyte aspiration during IVF treatment and was found to both be correlated to ovarian follicle size and maturity and to be synthesized by granulosa cells in developing follicles but not by oocytes, suggesting involvement in the antral phase of folliculogenesis (192, 193).

The mid-cycle LH surge in mammals sets in motion interconnected networks of signaling cascades to bring about rupture of the follicle and release of the oocyte during ovulation (194). Many mediators of these LH-induced signaling cascades are associated with inflammation, angiogenesis, steroidogenesis, and proteolysis, which act synergistically leading to a well-orchestrated process culminating in final oocyte maturation, follicle rupture and oocyte release. Novel contraceptive methods may alter molecular mechanisms mediating these processes and key molecules involved in them may be evaluated for prediction of RTF. A number of strong candidates for potential biomarkers can be found in this category.

Amongst the types of prostaglandins, PGE2 has been shown to be the most commonly involved in fertility and ovulation, with conditional knockouts suggesting its importance in the female reproductive system by demonstration of abnormal cumulus expansion, unruptured follicles, and subfertility when PGE2 or its receptor were targeted (195–199). Within the LH-induced signaling cascade, PGE2 has been shown to possess a specific role in cumulus expansion, inducing the upregulation of additional ovulatory genes mediated by cAMP and PKB-MAPK3/1 pathways (200). Additionally, antiestrogens and mifepristone have been shown to decrease uterine concentrations of PGE2 (201), while levonorgestrel decreased multiple prostaglandins, including PGE2 in a dose dependent manner (202).

Matrix metalloproteinases (MMPs) are a family of proteases primarily capable of degrading extracellular matrix (ECM) proteins. Members of this family are posited to regulate the complex system of endometrial connective tissue turnover and remodeling throughout the menstrual cycle, exhibiting a large degree of control over ovarian physiology (203). Expression of MMP has been shown to be induced by an LH or hCG surge, implicating it in the process of ovulation and follicular rupture (204). More evidence pointing to the role of MMPs in this process is the inhibition of ovulation by the metalloproteinase inhibitor GM6001 of rhesus macaque follicles inducing abnormality or absence of the normal stigmata indicative of follicle rupture (205). MMP10 specifically is a stromelysin theorized to potentially be involved in ECM modulation during follicular remodeling and regulating vascularity in the theca and granulosa cells. The regulation of MMP-10 has been studied in humans and rats and determined to be mediated by progesterone and EGF receptor signaling pathways, playing an important role in ovulation and luteinisation (206).

ADAMTS-1 is a disintegrin and metalloproteinase (ADAM) with thrombospondin type 1 motifs expressed in a wide variety of accessible bodily fluid sources. Examination of cumulus granulosa cells in women undergoing IVF revealed correlation between ADAMTS-1 expression and successful fertilization of denuded oocytes, with a three-fold higher expression rate of ADAMTS-1 in successfully vs. unsuccessfully fertilized oocytes (207). ADAMTS-1 expression has been shown to peak at a period 12 h after a simulated LH surge, when follicular rupture begins, and to be linked to progesterone receptor, the deficiency of which causes both under expression of ADAMTS and failure of follicle wall rupture (208). It has been demonstrated to be related to histological regulation of the uterus and ovaries and fertilization, with ADAMTS-1 null female mice ovulating only one-ninth of the total number of oocytes (209), as well as displaying impaired follicular development, implantation, and intrauterine development more broadly (210).

Luteinizing hormone receptor (LHR) mediates the binding action of LH, a key regulator of oocyte maturation and ovulation. LHR mRNA expression has been investigated in luteinized granulosa cells obtained from aspirated preovulatory follicles in women undergoing IVF, correlating a higher expression rate in cumulus cells with decreased fertilization rate (211). LHR gene expression was also found, however, to be higher in larger and more mature oocytes, to rise during antral follicular growth, and to be high in mural granulosa cells, preceding ovulation (211). Levels of LHR could be a marker of oocyte maturation and readiness for cumulus expansion.

Pregnancy-associated plasma protein-A (PAPPA) is a protease with an IGFBP substrate associated with hormone modulation found in cumulus granulosa cells. mRNA levels of PAPPA have been shown to be elevated both in more mature oocytes and those that resulted in live birth vs. those that did not (212). Additionally, PAPPA knockout mice were shown to have reduced ovulated oocytes, likely related to associated alteration of IGFBP4 proteolytic activity (213).

Some weaker but still promising candidates include serine protease inhibitor E2 (SERPINE2), which has been correlated with oocyte maturity and successful fertilization (214), and Cystatin S, isoforms of which were found to be heavily isolated to ovulatory vs. pre- or post-ovulatory phases (215, 216).

Endometrium plays a dynamic role throughout the menstrual cycle to allow for adequate gamete transport, fertilization and successful embryo implantation. These changes are governed by progesterone and estrogen responsive molecular pathways and involve several processes leading to cell differentiation and tissue remodeling. Impairment or dysregulation of the processes and mechanisms leads to pathology and infertility (217). In healthy fertile women, an array of cytokines, growth factors, transcription factors, cell adhesion molecules, and prostaglandins have been described during the receptive phase (218). The molecular events associated with the window of embryo receptivity include the expression of decidualization markers like prolactin and Insulin-like growth factor binding proteins (IGFBPs) (219). Additionally, the decidual stromal cells exhibit upregulation of homeobox proteins (HOX)A10, HOXA11, and leukemia inhibitory factor (LIF), which are signature biomarkers for endometrial receptivity.

Showing promise in tracking endometrial transport and receptivity, VEGF has also been identified as a variably expressed factor in the endometrium across the menstrual cycle. A peak in the mid-luteal phase of endometrial VEGF expression, as well as impaired levels in infertile patients during the same phase, signify a role in embryo implantation (220, 221). A decrease in VEGF expression in endometrial stromal cells has been demonstrated after administration of mifepristone (222) as well as oral desogestrel/ethinyl estradiol and etonogestrel implant (223). However, a levonorgestrel-releasing intrauterine system has been implicated in elevated endometrial tissue levels of VEGF mRNA (224) and certain isoforms were also elevated by medroxyprogesterone acetate, norethindrone, and levonorgestrel in Ishikawa cell lines (225).

There are two additional strong candidates for endometrial transport and receptivity biomarkers. Activin A is a dimeric multifunctional growth factor of the transforming growth factor beta (TGF-ß) family that plays a key role in endometrial receptivity, trophoblast invasion, and embryo implantation (226). Additionally, Activin A has been linked to the production of ovarian steroid hormones, oocyte maturation, and follicular development (227) and mice deficient in activin receptors showcase reproductive deficiency and suppression of follicle-stimulating hormone (228). The role of Activin A as a possible biomarker of fertility and implantation success has been supported by elevated levels found in endometrial washing fluid of women who subsequently became pregnant after intrauterine insemination vs. those who did not, as measured by ELISA (229). A counterpart to activin, α-inhibin is also a member of the TGF-ß and has a negative regulatory effect upon actin. An increase of α-inhibin gene expression in the endometrium during the mid-secretory phase of the menstrual cycle was correlated with higher rates of implantation failure vs. confirmed clinical pregnancies in women undergoing IVF and embryo transfer (230). Alteration of expression by progestin-only contraceptives has been demonstrated for both Activin A and related α-inhibin subunits at both the protein and mRNA level, as well as during the menstrual cycle and early pregnancy (231).

Endometrial LIF, a cytokine of the interleukin 6 family, is another such promising marker. Differences in expression in the secretory phase between fertile and infertile women are demonstrably striking (232) with immunostaining of endometrial biopsy samples finding strong LIF expression to yield a 6.4-fold higher pregnancy rate than weaker intensities in women undergoing IVF (233). Embryo attachment failure in mice following treatment with levonorgestrel, at a dose established to be equivalent to use as emergency contraceptives in humans, has been observed alongside significantly reduced uterine LIF expression (234). However, it should be noted that evaluations of the effect of emergency contraceptive levonorgestrel in humans generally does not support a demonstrable effect on endometrial receptivity and some other associated markers (235).

Glycodelin is a glycoprotein of the lipocalin superfamily present in the uterus as a secretion of decidualized endometrial glands and stromal cells and is a proposed biomarker for endometrial function (236). Glycosylated isoforms have been shown to be differentially expressed across phases of the menstrual cycle, with an altered profile of expressed glycoforms in endometriosis during the use of intrauterine devices, an evaluation which supports the potential of GDA as a biomarker of endometrial receptivity (237). Other potential markers whose expression is altered during use and after removal of IUDs include LIF, IGFBP family of proteins, keratin 8, folate receptor, and MMP10 (238).

The mucus in the cervical canal controls the ascent of spermatozoa and is a target of contraceptive action by progestin-only pills and levonorgestrel (89–91, 239, 240). A variety of mucins are expressed on the epithelia of the upper genital tract and have been posited to act as inhibitors of implantation, although examinations of the exact role of each, and of the more frequently studied MUC1 specifically, show some contrasting results (241–243). The membrane-bound MUC16 measured in endometrial samples correlated lower expression with recurrent pregnancy loss and conception after IVF, and higher expression with elevated progesterone during ovarian stimulation (244). MUC16 has been proposed as a potential marker for fertility during normal menstrual cycling, acting as a regulator of the local osmotic modulus and mucus layer integrity in the reproductive tract (152). Amongst the gel-forming mucins identified in humans, MUC2, MUC5AC, MUC5B and MUC6, MUC2 related polymorphisms have been related to infertility and endometriosis (245). MUC2 has also shown significant differences in expression under the effects of oral contraceptives (246). Some additional candidates identified that proved weaker under our specific scoring criteria but remain somewhat promising for predicting endometrial receptivity include Interleukin-18, hDP, Integrins, BCL6, Aromatase P450, L-selectin ligand, and urocortin.

There exist several limitations constraining effective clinical implementation of prospective biomarkers in tandem with experimental validation. Regulatory requirements that seek to verify certain standards of clinical utility can require an extensive process of testing and justification that contain any number of stumbling blocks for potential marker candidates. Specificity, sensitivity, and accuracy at predicting interruption and RTF need to be established before considering candidate molecules as true biomarkers. Providing or developing an assay method which meets standard for clinical utilization with regards to quality control and assurance, speed, and sample storage or transportation from readily accessible sources is a challenge in and of itself beyond the hurdles for proving analytical performance and clinical relevance in decision making. The FDA requires a multi-step process by which these factors are analyzed through their Biomarker Qualification Program (247).