Abstract
The rapid expansion in reproductive medicine services in recent decades, with notable technological advances in oocyte cryopreservation [1] and artificial reproductive techniques using mature gametes [2], has aligned with developments in oncology to support a rapid expansion in the provision of fertility preservation for cancer patients. Ongoing developments in oncology continue to improve statistics for long-term survival and cure, particularly in the pediatric oncology environment where the great majority of children with cancer can now expect to become long-term survivors [3]. It is, however, unfortunately the case that many such long-term survivors will have one or more significant health issues related to their disease and its treatment [4], and this includes the impairment to their fertility [5]. At the time of diagnosis, loss of fertility is regarded by patients as a keen concern [6].
The rapid expansion in reproductive medicine services in recent decades, with not able technological advances in oocyte cryopreservation [1] and artificial reproductive techniques using mature gametes [2], has aligned with developments in oncology to support a rapid expansion in the provision of fertility preservation for cancer patients. Ongoing developments in oncology continue to improve statistics for long-term survival and cure, particularly in the pediatric oncology environment where the great majority of children with cancer can now expect to become long-term survivors [3]. It is, however, unfortunately the case that many such long-term survivors will have one or more significant health issues related to their disease and its treatment [4], and this includes the impairment to their fertility [5]. At the time of diagnosis, loss of fertility is regarded by patients as a keen concern [6]. Thus, the provision of fertility preservation services is part of a broader survivorship agenda for patients, seeking to minimize the long-term adverse effects of cancer treatment and promote health in its broadest context.
While concerns regarding fertility mostly relate to biological effects on the reproductive system, it is also important to consider the wide range of issues that such patients may face [4]. For females, some of these will be specific medical issues relating to potential future pregnancy, such as radiation damage to the uterus or anthracycline-induced damage to cardiac function. Recommendations for surveillance in young cancer survivors have been published recently [5]. Breast cancer is the commonest indication for a patient seeking fertility preservation but carries its own specific issues, particularly the increasingly long-term endocrine treatment recommended for women with hormone-sensitive cancers [7], and concern that pregnancy may increase the risk of recurrence, although current evidence suggests that this is not, in fact, the case [8]. Other relevant concerns that apply to cancer survivors of either gender include the need to establish healthy long-term relationships, which is significantly reduced in survivors of brain and central nervous system (CNS) cancer [9, 10] and the psychological impact of a cancer diagnosis which may influence an individual’s desire to start or complete their family [11, 12]. While these are all important considerations, this chapter focuses on the cancer treatments that specifically affect fertility and the evidence that such treatments have effects on reproductive function, including data on fertility and reproductive lifespan although most of the literature is on surrogate markers of these key clinical outcomes.
The effects of cancer treatment on ovarian function have been known for several decades, with early studies showing the acute loss of growing follicles in children treated for leukemia, and the longer term consequences of ovarian damage with early menopause in women treated for Hodgkin lymphoma [13, 14]. This was particularly linked to treatment with the alkylating agent cyclophosphamide, and highlights the important endocrine function of the ovary in addition to its role in fertility. This is of particular importance given increasing longevity in women, with thus the likelihood of many decades of life in a state of estrogen deficiency. Growing ovarian follicles, with their rapidly proliferating granulosa cells, are particularly sensitive to cytotoxic chemotherapy and many therapies will therefore result in significant depletion of the growing pool, or often in induction of amenorrhea during treatment [15]. Of greater importance for long-term ovarian function is, of course, whether the primordial follicle pool is affected. While it appears that some therapies, particularly radiotherapy as well as some chemotherapy agents, will indeed directly affect the primordial follicle pool, it is also believed that a reduction in the inhibitory effect from the depleted pool of growing follicles will cause an increase in primordial follicle growth initiation, thus accelerating depletion of the primordial pool [15, 16]. While most research activity has focused on the ovarian follicles as the key target of cancer therapy, the ovarian vasculature and stroma are also potential targets and have been clearly demonstrated to be affected by chemotherapy in both women and animal models [17–19]. The consequences of this are poorly understood as it is difficult to measure and investigate in vivo but may be at least part of the basis for the impact of age on the extent of ovarian toxicity, in addition to the age-related depletion of the ovarian reserve. This has been recently demonstrated in a study of recovery from a low toxicity chemotherapy regimen for Hodgkin lymphoma, which showed that young women showed a full recovery, but in women over the age of 35, recovery was limited and this was not related to the pretreatment ovarian reserve [20].
For males, effects of chemotherapy and radiotherapy on gonadal function and sperm production are also well recognized [21]. As opposed to the situation in females, gametes are not produced in males until puberty. Furthermore, the continuous production of sperm in adulthood is dependent on the survival and maintenance of the spermatogonial stem cell (SSC) population. While low doses of chemotherapy or radiotherapy may deplete the pool of differentiating spermatogonia, studies in primates have shown that SSCs may survive, and existing spermatocytes and spermatids can continue their maturation into sperm [22]. Long-term recovery of sperm production following cancer treatment depends on the ability of mitotically quiescent SSCs to survive and resume self-renewal or differentiation, and if the treatment results in all SSCs committing to apoptosis then the patient becomes permanently infertile [21]. While it has previously been considered that prepubertal SSC are protected from damage relative to adults, it is clear that loss of spermatogonia may occur following exposure to alkylators in childhood [23, 24], and that the cumulative dose of exposure to cyclophosphamide during childhood is associated with an increased prevalence of azoospermia in adulthood [25].
Chemotherapeutic agents may be conveniently classified according to their risk to fertility (Table 4.1), which highlights the particular toxicity associated with the alkylating agents [26, 27]. However, while this is a useful general guideline, a number of caveats need to be remembered when interpreting these and similar data. For females, these include that much of the data derives from surrogate effects on fertility, particularly the prevalence of amenorrhea after treatment, and may not include the very important compounding effect of the patient’s age. For men the majority of the data involves measurement of sperm count, the timing of which is important to account for temporary azoospermia that may occur soon after treatment, or alternatively, the resumption of spermatogenesis that in some cases can occur up to 15 years after treatment [28]. Additionally, agents are generally given in combination and thus it must be recognized that detailed analyses on reproductive function of these combinations and the relative contributions of the specific components is challenging and often based on a limited amount of data. This box also does not include the newer biological agents, on which there are very limited data indeed. These include monoclonal antibodies and kinase inhibitors, and some, such as bevacizumab have activity against biological processes such as angiogenesis which are very likely to have adverse effects on gonadal function [29]. Radiotherapy is also well recognized to have adverse effects on gonadal function, and is dealt with specifically in the preceding chapter. These effects will be both dose-dependent and age-dependent [30], although estimating the dose to the gonad may be difficult. In females, pelvic radiotherapy will also significantly affect uterine function. Radiotherapy to the uterus carries clear adverse risk to a subsequent pregnancy, including increased risk of early and second trimester miscarriage, premature delivery and a range of other effects include growth retardation, stillbirth and post-partum hemorrhage [31–33] as well as other maternal complications [34]. In children and adolescents radiotherapy can prevent subsequent growth of the uterus, with a clear relationship between adult uterine volume and age at irradiation being demonstrated [35]. Normative values allow estimation of uterine volume compared to the normal population [36], but this highlights the wide range of ovarian volume in normal women, and identification of those at risk, other than in the more extreme cases, is therefore difficult. It may be the case that further analyses such as uterine vascular flow may be of value, but this has yet to be clarified. Radiotherapy can also affect hypothalamic and pituitary function in males and females. This may be subtle as demonstrated in a study of women who had been treated with craniospinal radiotherapy for childhood leukemia [37]. These women showed an increased risk of short luteal phases, which were linked to deficient mid-cycle LH surge. It also appeared that this risk increased with increasing time since therapy, that is, ongoing damage to hypothalamic function. Similarly, for males, evolution of hypothalamo-pituitary damage following irradiation may take up to several years with the gonadotrophin deficiency manifesting in some patients up to 8 years after treatment [38]. These data highlight the need for long-term surveillance of cancer survivors who have received these treatments [39].
| High or medium risk | Low risk | |
|---|---|---|
|
|
|
Much of the data on pregnancy and birth after cancer come from the large US and UK childhood cancer survivor studies. The USCCS has for many years provided excellent data on ovarian function, pregnancy and pregnancy outcome in childhood cancer survivors using siblings as the control group. These studies demonstrated the adverse effects of abdomino pelvic radiotherapy and alkylating agent chemotherapy [40, 41], and more recent studies have provided detailed information on specific chemotherapy agents used in the absence of radiotherapy on the likelihood of pregnancy and birth [42]. In this analysis the hazard ratio (HR) for live births in female childhood cancer survivors treated with chemotherapy only was 0.82 (confidence interval [CI], 0.76–0.89), with alkylating agents only demonstrated to have had adverse effects at the highest dose, and busulfan and lomustine were also demonstrated to have significant adverse effects. This population has also demonstrated the risk of subfertility in the absence of ovarian failure after childhood cancer [43], an effect that was also highlighted by a questionnaire-based analysis of adult cancer survivors [44]. This large study of 620 women who had received chemotherapy alone reported acute ovarian failure risks for the more common cancers affecting women of reproductive age and its relationship with age. However, importantly, this study also highlighted that the prevalence of infertility was approximately 40% at age 35 years and the probability of early menopause was at least 25% at age 30 years [44]. For male cancer survivors, there was a decreased likelihood of siring a pregnancy or having a live birth, with a greater impact than in female survivors (HR for live birth 0.63, CIs 0.58–0.69) [42]. Reduced fertility in males was significantly associated with upper tertile doses of alkylators and cyclophosphamide.
While all these studies are largely questionnaire based and thus open to potential bias they raise important questions that need to be addressed in specific studies with relevant design but thus far have largely not been undertaken in adult women. A population-based analysis of parenthood in female survivors of childhood Hodgkin lymphoma (aged less than 18 at diagnosis) showed that the probability of parenthood was the same as the national German population at age of diagnosis up to 40 years [45]. As with the USCCS study mentioned earlier, this study also demonstrated effects of the alkylating agents procarbazine and cyclophosphamide at the highest doses but did highlight the very marked effect of radiotherapy to the pelvis. Although it did not show clear effects of treatment protocol or age at treatment, a large body of work from the adult German Hodgkin’s group has shown clear effects of both factors in adult women [46], although the outcome measure is related to amenorrhea/resumption of menses rather than fertility. A comparable UK survey of the different treatment regimens for Hodgkin lymphoma showed very clear differences in the HR for early menopause with different therapies [47], with no effect detected for the non-alkylating-based therapy ABVD, whereas alkylating agent regimens and pelvic radiotherapy, and particularly the combination had substantially increased HRs of as high as 20–35. This again was questionnaire-based data, and information was obtained only from 50% of potential respondents. To try to provide a more comprehensive and unbiased analysis of the chance of pregnancy after cancer, we have recently undertaken an analysis of the latest data from the Scottish Cancer Database, together with national hospital discharge records. These provide accurate information on diagnosis and pregnancy and its outcomes, although do not include detailed information on specific treatments. This analysis of 23,201 cancer survivors diagnosed between 1981 and 2012, age 0 to 40 years at diagnosis, shows an overall reduction in the likelihood of pregnancy after diagnosis of 38% compared to women in the general population [20]. Strikingly, this deficit was noted across all diagnostic groups, including diagnoses such as skin cancer where there may not be marked effects of treatment on reproductive function (Table 4.2). This may highlight issues regarding changes in patients’ priorities for pregnancy after a significant diagnosis. A further analysis investigated women who had not been pregnant before cancer diagnosis, and for these women the rate ratio of achieving a subsequent pregnancy was 0.53 compared to controls age-matched for the time of diagnosis [20]. Marked changes in the impact of some specific diagnoses over time were also demonstrated, with an improvement in the adjusted HR for a first pregnancy after diagnoses including breast, cervical cancer, and Hodgkin lymphoma, but no significant changes over time for leukemia or brain/CNS cancers (Figure 4.1). Both those two diagnostic groups had a very marked impact on the chance of pregnancy after cancer, which has not changed over the last three to four decades. While this study does not provide information relating to specific treatments, these findings do highlight areas for further research and the focusing of clinical activity where there remains the greatest potential risk to loss of fertility.
| Diagnosis | SIR | 95% CI |
|---|---|---|
| Cervix uteri | 0.34 | 0.31–0.37 |
| Breast | 0.39 | 0.36–0.42 |
| Brain, CNS | 0.42 | 0.36–0.48 |
| Leukemia | 0.48 | 0.42–0.54 |
| Ovary | 0.63 | 0.57–0.69 |
| Hodgkin lymphoma | 0.67 | 0.62–0.73 |
| Non–Hodgkin lymphoma | 0.67 | 0.58–0.77 |
| Thyroid | 0.79 | 0.72–0.86 |
| Skin | 0.87 | 0.84–0.90 |
Data are standardized incidence ratios (SIR) and 95% CIs, from all women with cancer diagnosed at ages 0–40 in Scotland between 1981 and 2012 (total of 23,201 cancer survivors) compared to matched controls in the general population. Data from (Anderson et al., 2018), published with permission.
In addition to these important studies where fertility or other direct clinical outcomes were studied, much work has involved the use of surrogate markers of gonadal function. For women this particularly includes recovery of regular menses, or biomarkers of ovarian function, notably measurement of anti-Müllerian hormone (AMH). The importance of age comes through in all studies where this is investigated with the prevalence of ongoing menses after chemotherapy for, for example, breast cancer declining very dramatically with age from approximately 90% in women under the age of 35 years to less than 30% in those aged over 40 years [48]. Conversely, the risk of persisting amenorrhea varies by age and diagnosis. AMH was first demonstrated to be a potentially useful biomarker of ovarian toxicity in a study of childhood cancer survivors who, despite having regular menses, showed lower AMH concentrations without other reproductive endocrine abnormalities [49]. Subsequently, a large body of work has confirmed and extended these findings in both adult and childhood cancer survivors (reviewed in [50]). Prospective analyses of ovarian function during and following chemotherapy have shown how AMH falls rapidly during treatment, and to a much greater extent than other ovarian hormones, including inhibin B [51]. In women treated for breast cancer, who are generally in their 40s, very little recovery is apparent whereas in younger women treated for Hodgkin lymphoma, recovery is clearly dependent on the treatment regimen [20, 52, 53] (Figure 4.2). This is in close agreement with studies described earlier with either resumption of menstruation or risk of early menopause as outcomes although, perhaps surprisingly, there remains a lack of large studies with fertility as an outcome considering the importance of this diagnosis in young women. As mentioned earlier, however, it is clear that ovarian recovery after even low toxicity regimens such as AVBD is limited by age independently of the reduction in ovarian reserve associated with increased age [20]. In the context of fertility preservation, therefore, while younger women can be reassured, for women in their mid and later 30s fertility preservation becomes a more important consideration. The patient’s BMI has also been suggested to be a determinant of recovery of ovarian function after chemotherapy for early breast cancer [54]. These studies have also demonstrated that pretreatment AMH is an important predictor of posttreatment ovarian function [54, 55], and this information will be useful in counselling patients. While this has been most clearly demonstrated for breast cancer and for young women with Hodgkin lymphoma treated with ABVD, further prospective analyses are required for other diagnoses and treatments. It is, of course, important to remember in this context that AMH is not a good predictor of fertility where there remains even a low ovarian reserve, with the chances of pregnancy being similar in women with a high, medium, or low AMH having been clearly demonstrated to be equal in studies in both women in their 20s [56] and a more recent study in women largely in their 30s [57], although this was not following cancer treatment.
