Can Ovarian Reserve Assessment Predict Fertility following Breast Cancer Treatment?

“We now have two accurate markers of the ovarian reserve; AMH and AFC. These very clearly predict the number of eggs that can be obtained following ovarian stimulation, and there are increasing data on their long-term use to predict ovarian function after chemotherapy”

Breast cancer is the most common malignancy affecting women, with some 230,000 women each year being diagnosed with this condition in the USA. Mortality has been declining over recent years, and a significant contribution to this is the more widespread use of chemotherapy [1]. This success story is therefore tempered by the adverse affects of chemotherapy, prominent among which are its effects on reproductive potential.

“The more important effects of chemotherapy relate to the loss of the nongrowing pool of primordial follicles, which is the true ovarian reserve. This longer term effect of chemotherapy has also long been recognized.”

The adverse effects of chemotherapy on ovarian function have been recognized for many years. Both the oocyte and granulosa cells may be targets for different chemotherapy agents, but in fact there is remarkably little literature investigating the ovarian toxicity of these agents at this level [2]. An early study demonstrated that the ovaries of girls who died during chemotherapy for leukemia had fewer growing follicles than did those who died of accidental causes [3]. However, once follicles have started to grow the great majority of them are destined to undergo atresia anyway, with only a tiny proportion surviving to ovulate. These acute affects of chemotherapy may, however, result in falls in estrogen production and early amenorrhea.

The more important effects of chemotherapy relate to the loss of the nongrowing pool of primordial follicles, which is the true ovarian reserve. This longer term effect of chemotherapy has also long been recognized, with early studies also demonstrating that alkylating agents such as cyclophosphamide had significant gonadotoxic effects [4]. Other agents have much less gonadotoxicity [5] but therapeutic regimes often involve combinations of drugs. Classical chemotherapy regimes for breast cancer involve cyclophosphamide in combination with methotrexate and 5-fluorouracil in the ‘CMF’ regime, with anthracyclines and taxanes also widely used.

The number of primordial follicles in the ovary declines with age, reaching a critical threshold below which regular cyclical ovarian activity is no longer possible and the menopause ensues [6]. Acceleration of this loss will result in an earlier menopause, although the term premature ovarian insufficiency is increasingly used [7]. The number of primordial follicles cannot, however, be assessed in vivo and thus we look to the use of surrogates. The recognition that early growing follicles produce anti-Müllerian hormone (AMH) has led to a dramatic increase in research on the value of this hormone [8]. AMH concentrations are not determined by the later stages of follicle growth, that is, the larger antral and preovulatory follicles, which predominantly secrete estradiol and the inhibins. Circulating concentrations are therefore relatively stable across the menstrual cycle and for practical purposes blood sampling can be on any day of the cycle. It was rapidly recognized that AMH concentrations showed a very good relationship with the number of follicles that could be stimulated to grow during ovulation induction prior to IVF and with the number of eggs that would be retrieved, and it is increasingly part of pretreatment assessment in assisted conception. A recent analysis has reported AMH concentrations across the full female lifespan [9]. It is therefore a marker of the functional ovarian reserve, with the term ovarian reserve in fact largely used in this context to indicate the number of small growing follicles. The ultrasound measurement of the ovarian reserve is the antral follicle count (AFC), which is the number of follicles seen in the two ovaries of up to 5 or 10 mm diameter, with studies using both definitions. AFC also shows an excellent correlation with oocyte yield after ovarian stimulation and indeed there is an extremely close relationship between AFC and AMH concentrations. At present the evidence suggests there is little to choose between the two in their accuracy in predicting egg yield at IVF [10]. These are thus the two methods currently available to assess the ovarian reserve and while neither directly assess the number of primordial follicles it is clear that there is a significant relationship between the number of primordial follicles and AMH concentrations [11].

“These findings need to be validated in further larger cohorts but certainly provide encouraging evidence that AMH can independently predict long-term ovarian activity after breast cancer chemotherapy.”

Both AMH and AFC have been used to investigate the effects of chemotherapy on gonadal function. Both fall during chemotherapy treatment, with the fall in AMH occurring very rapidly [12]. AMH concentrations have been demonstrated to be lower postchemotherapy in cancer survivors than in controls in a number of contexts, both in adult women with a range of diagnosis including breast cancer [13,14] and in women who have been treated for cancer during childhood [15,16]. The gonadotoxicity of different regimes has been demonstrated in women with lymphoma, with a clear post-treatment rise in AMH following low toxicity therapy, whereas there was little recovery following alkylating agent therapy [17]. One might imagine that women with a higher ovarian reserve pretreatment will be more likely to have ongoing ovarian activity following chemotherapy. For a long time, one of the best recognized predictors of whether a woman will be amenorrheic or not following chemotherapy is her age [18], which is of course a surrogate for her ovarian reserve. This has been very clearly demonstrated in prospective studies in women with breast cancer [19]. We have recently investigated the value of AMH as a more direct index of ovarian reserve in a prospective study of women recruited at the time of diagnosis of breast cancer who were then followed up for 5 years [20]. Pretreatment of AMH did indeed predict whether or not the women had chemotherapy-related amenorrhea at long-term follow-up, with amenorrhea being validated as an index of ovarian function. In a multivariate analysis examining variables including age, AMH and follicle-stimulating hormone (all variables that showed a relationship by univariate analysis), only AMH was significant. This means that age did not add to the value of AMH, which is what one would predict if AMH is a sufficiently accurate direct marker of the ovarian reserve, whereas age is a surrogate. This study is limited by the small number of patients, but it has recently been validated in a second, albeit also small, independent cohort of women also treated for breast cancer [21] and is supported by another prospective study that also showed higher pretreatment AMH in women who retained menstrual activity after treatment for breast cancer [22]. These findings need to be validated in further larger cohorts but certainly provide encouraging evidence that AMH can independently predict long-term ovarian activity after breast cancer chemotherapy.

The question remains as to whether AMH or indeed other markers of the ovarian reserve can predict fertility following treatment for breast cancer. Clearly women with ovarian activity should be much more likely to be able to fall pregnant than women who are menopausal as a result of their chemotherapy. However, the relationship with pretreatment AMH remains to be demonstrated, and clinically useful predictive values determined. There are very limited data relating AMH to spontaneous conception, although such a relationship has been proposed [23]. This paper showed a surprisingly small effect of age on conception and it is important that it is repeated in other cohorts of normal women. As there is no clear evidence for a link between AMH and egg quality, a relationship between ovarian reserve and spontaneous fertility when only a single oocyte is being released at a time would perhaps be surprising.

AMH has also recently been demonstrated to predict spontaneous menopause in a study involving approximately 11 years of follow-up in a Dutch cohort [24], supporting previous data [25]. Age remained an important variable in that analysis unlike in the chemotherapy studies, which may be related to the dramatically deleterious effect of chemotherapy on the ovarian reserve compared with the more subtle and gradual effects of increasing age. These data do on the whole, however, support a relationship between ovarian reserve measured by AMH and long-term fertility prospects. Studies need to be undertaken with fertility as the primary outcome rather than surrogate measures of ovarian activity. Such studies will be challenging both in terms of the number of patients required and the duration of follow-up.

In summary we now have two accurate markers of the ovarian reserve; AMH and AFC. These very clearly predict the number of eggs that can be obtained following ovarian stimulation, and there are increasing data on their long-term use to predict ovarian function after chemotherapy. This would lend encouragement to their use in predicting future fertility but it must be emphasized that there are currently no robust data demonstrating this link and their use in this context cannot therefore be promoted. Adequately powered prospective studies will be needed to address this important question, which will be very helpful in guiding women with newly diagnosed breast cancer as to their individualized future fertility prospects and therefore their need to consider fertility preservation options.