Fertility concerns and preservation in younger women with breast cancer

3.1. Treatment options for different stages of breast cancer 

As illustrated in Table 2, therapy for breast cancer patients is based on stage of disease and prognostic potential [47].

Treatment of breast cancer in premenopausal women with invasive breast cancer currently is most often multimodal, comprised of surgery, adjuvant radiotherapy and chemotherapy [29], [30], [31], [32], [33], [34], [48], [49], [50], [51], [52], [53], [54]. Patients for whom chemotherapy is planned with drugs such as alkylating agents are at risk of decreased ovarian reserve, and are therefore candidates in whom to consider fertility preservation options.

3.2. Mechanisms of action of chemotherapeutic agents 

Most chemotherapeutic regimens employed in breast cancer treatment interfere with some aspect of the cell cycle, thus their effects are randomly distributed to all proliferating cells in the body including healthy cells. Furthermore, drugs are chosen to attack neoplastic cells that share a significant degree of functional, histological and behavioral homology to germinal cells. Therefore many of these drugs also destroy germinal tissue such as steroid-producing cells of the ovary (granulosa and theca cells), as well as oocytes, with resultant POF and early menopause [16], [17], [55], [56]. Histologically streak ovaries with fibrotic changes from chemotherapy demonstrate marked follicular loss after treatment with alkylating agents such as cyclophosphamide [57]. In fact studies demonstrate a dose-dependent loss of oocytes and granulosa cells [28], [58], [59], while in mice primordial follicle destruction is noted at a minimal exposure to cyclophosphamide [28], [59]. The extent of ovarian injury is primarily dependent on the age of the patient (with younger patients demonstrating less damage than older patients), and the duration of exposure to chemotherapy [16], [17], [19], [22], [31], [36], as well as overall dosage of alkylating agent used which in treatment of breast cancer is usually a cyclophosphamide based therapy [60].

Chemotherapy-related amenorrhea has been reported in nearly 40% of women under age 40 who were treated for breast cancer with cyclophosphamide, methotrexate and 5-FU (CMF) for as little as 3 months. Cyclophosphamide, an alkylating agent, works on non-proliferative cells by interfering with cellular DNA function. Methotrexate mediates its antimetabolic effects by depleting cellular folate, and has not been found to cause germ cell damage. Although high-dose methotrexate employed in the treatment of sarcoma was found to cause post-treatment infertility, at the doses used in breast cancer therapy it is not believed to have any significant effect on causing POF [61], [62], [63]. 5-FU, an antimetabolite, interferes with DNA replication in proliferative cells. Abusief et al. studied the effects of paclitaxel (T) an antitubulin, in addition to adriamycin and cyclophosphamide (CA) chemotherapy, and dose density (q2 week treatment, dd), on follow-up menstrual status in breast cancer patients. They found that using T or dd did not appear to statistically increase the risk of amenorrhea on follow-up, however a larger cohort analysis is planned [64]. In the Hellenic Cooperative Oncology Group (HeCOG) breast cancer studies, evaluation of weekly [65] or q3 weekly administration of paclitaxel with 2 years of GnRH ovarian suppression in premenopausal patients [66] did not demonstrate any significant toxicity, although amenorrhea was not specifically assessed in these patients. Other clinical studies have reported similar low toxicity from regimens consisting of palcitaxel in the treatment of prostate cancer [67].

Gonadotoxicity of chemotherapeutic drugs is of significant concern to patients desiring future fertility. The primary medications employed in the treatment of breast cancer are listed in Table 3. There are many factors associated with the gonadotoxicity of these medications that include type of medication, dosing frequency, duration of treatment and especially the age of the patient. In addition, though not routinely assessed, the ovarian reserve of the patient at the time of treatment probably also contributes to the extent of ovarian damage. Since most regimens involve a combination of chemotherapeutic medications, it remains difficult to accurately ascertain the individual toxic effects of each drug. Animal studies suggest that alkylating agents are the most toxic and are most likely to result in permanent ovarian damage [43], [76]. Antimetabolites are mostly active in proliferating populations of cells and therefore demonstrate minimal to no impairment of fertility in patients [16], [61], [63]. The anthracyclines and vinca alkaloids appear to trigger apoptosis in target cells and it is unclear whether these drugs indeed cause ovarian injury. Hormone treatments appear to have the least impact on gonadotoxicity and may be part of a desirable adjuvant therapy if appropriate.

Table 3. Drug mechanisms of action.

	Mechanism of action	Drugs	Gonadal toxicity
	Anthracyclines-antibiotics	Doxorubicin	Questionable [1], [68], [69], [70]
		Epirubicin
		Mitoxantrone	Suspected
	Taxanes-antitubulins	Paclitaxel	Variable [29], [64]
		Docetaxel
	Vinca alkaloids	Vincristine, vinblastine	Demonstrated [16], [71]
	Alkylating agents	Cyclophosphamide	Demonstrated [16], [55], [63], [72], [73]
		Thiotepa	Demonstrated [48]
	Antimetabolites	Methotrexate	Does not appear to be [61]
		Fluorouracil	Does not appear to be [63]
	Hormones	Tamoxifen	No [31]
		Letrozole	No [24], [74], [75]
	Other	Cisplatin	Unknown
	HER2/Neu antibody	Herceptin	None known

Treatment regimens utilizing doxorubicin in younger breast cancer patients for less than 2 years does not appear to have any permanent effects on ovarian function [1], [68], [69], [70]. While even as short as a 3-month treatment course with CMF in women under 40 years of age can result in greater than 40% of patients demonstrating chemotherapy-related amenorrhea (CRA) [95% confidence interval CI, 36–44%] and greater than 75% CRA in women over 40 years [95% CI, 74–78%] with an average CRA rate of 68% [95% CI, 66–70%]. Much of this toxicity is believed to stem from the use of cyclophosphamide (Table 4). All alkylating agents are noted to be gonadotoxic resulting in increased risk of CRA and premature ovarian failure when used for greater than 3 months.

Table 4. Rates of chemotherapy-related amenorrhea.

	Adjuvant chemotherapy	Months treated	Rate of CRA
			<40 years	>40 years
	CMF [18], [19], [79]	12	52–61%	89–95%
	CEF [80], [81]	6	51%	ND
	l-PAM [48]	24	22%	76%
	AC [16]	3	34%	ND
	FAC [82]	24	9%	ND
	TAC	3	51%	ND

CRA=chemotherapy-related amenorrhea, cyclophosphamide (C), methotrexate (M), 5-flourouracil (5FU), epirubicin (E), doxorubicin (A), l-PAM (l-phenylalanine mustard), docetaxel (T).

Initial studies of breast cancer adjuvant chemotherapy looking at the use of the alkylating agent thiotepa indicated 40% POF in-patient's treated with this medication in contrast to 3% POF in controls [48]. Most reports on CRA and POF look at results after a 12-month chemotherapeutic course. The duration of treatment appears to increase the cumulative dose and reluctant ovarian insult. Reports by Padmanabhan et al. indicate that CMF treatment after 3, 6 and 12 months results in an increased rate of CRA of 50, 70 and 80% at the aforementioned time intervals relative to 3, 10 and 17% in controls clearly demonstrating the effect of increased dosing on diminishing ovarian function [77], [78]. Accordingly a single treatment of cyclophosphamide is associated with approximately 10% rate of amenorrhea, up to 61% of young patients’ demonstrated ovarian failure with 12 months of continued treatment [17]. The greater the dose of each chemotherapeutic treatment used greater the risk of ovarian dysfunction. Other agents such as taxol, vincristine, vinblastine, and doxorubicin have varying degrees of gonadotoxicity. Mouse studies demonstrate that a mechanism of action of doxorubicin may be due to stimulation of programmed cell death-apoptosis.

After 3, 6, 12 and 24 months of chemotherapy, rates of CRA with different drug regimen range from 9% after 24 months of FAC to 61% after 12 months of CMF in patients <40 years of age. By contrast older patients demonstrate significantly greater rates of amenorrhea related to treatment protocol and range as high as 95%.

There are at least three components that influence development of infertility with chemotherapy and these include: age of patient, duration of treatment as well as the specific type and dose of agent employed. The ovaries of younger patients appear to be more resistant to chemotherapy while older patients demonstrate longer and more often permanent sequelae. This may in part reflect ovarian reserve and the ability to sustain drug insult. The duration of treatment reflects at least two components of treatment; the (i) first being cumulative drug-dose and the other (ii) repetitive injury to the gonadal tissue with each treatment cycle. And as previously discussed specific mechanisms of action of the various agents employed impact cellular physiology in diverse fashions, some mediating significantly greater detrimental effects on target tissue while the effect of others may be relatively mild and transient. Additionally each medication is cleared at different rates by the body and therefore has a range of half-lives within the systemic distribution.

Bergh et al. report that adjuvant polychemotherapy at 10 years results in an absolute mortality reduction for patients younger than 50 years by 12% for node positive (34% relative mortality reduction corresponding to an estimated median survival prolongation of several years) and 6% for node negative patients [83]. Current treatment protocols for breast cancer include cyclophosphamide, anthracyclines, taxanes, 5-FU as well as hormone based therapy [29], [30], [31], [32], [33], [34]. Based on a systematic review of chemotherapy trials in breast cancer patients reviewing 233 randomized studies, 9 meta-analysis of randomized studies, a population-based cohort study and 18 overviews/retrospective analyses including a total of 155,243 patients Bergh et al. report an absolute survival benefit of 3% at 5 years with the use of anthracycline-containing combination chemotherapy. They also propose the addition of taxanes to the treatment regimen with an alkylating agent-cyclophosphamide. The addition of anthracyclines increases the response rate and statistically significantly improves the survival compared with non-anthracycline-containing chemotherapy, except for CMF combined with prednisone/prednisolone, which significantly improves survival relative to some anthracycline combinations. Thus most non-hormone based chemotherapy regimens will negatively impact ovarian function in reproductive age women with breast cancer.

3.3. Infertility with radiation 

Radiation therapy for breast neoplasms is localized and directed above the diaphragm thereby sparing the adnexa. Studies demonstrate that even with abdominal field radiation fertility is preserved with radiation doses <5Gy [84], [85], [86]. The cytotoxic effects of radiation therapy as well as chemotherapy are both dose and age dependent [85], [86]. The prepubertal ovary is relatively resilient to treatment side effects perhaps reflective of the significantly high ovarian reserve. The most sensitive appears to be the perimenopausal adnexa, which may in part reflect loss of follicles in the face of already, diminished ovarian reserve. Thus the number of primordial oocytes during the period of radiotherapy influences the age for onset of POF. Animal studies demonstrate at that exposures of the adnexa to <2Gy does not show any evidence of fetal malformations or anomalies [85]. The radio sensitivity of oocytes is <2Gy. Utilizing this data in a mathematical model that also considers the average age of menopause a differential equation known as the Faddy–Gosden model of ovarian primordial oocytes decline may be employed to predict ovarian failure for any given dose of radiotherapy. In the event of whole body radiation, employing the various radiologic imaging modalities such as computerized tomography and magnetic resonance imaging one might tailor specific dosing in the region of the adnexa. Since the primary sequelae stemming from radiotherapy is directly related to ovarian exposure supraabdominal or localized thoracic radiotherapy for breast neoplasm should not have a significant bearing on future fertility.

3.4. Surgical infertility 

Infertility arising from breast cancer surgery is highly unlikely due to the anatomic separation of the surgical site from the adnexa. The ovarian vessels arise from the posterior retroperitoneal plane form the abdominal aorta and course along the pelvic brim within the suspensory ligament. Thus supradiaphragmatic surgery has little or no impact on the adnexa. If in patients such as those with BRCA mutations, where prophylactic oophorectomy is performed obviously surgery would render the patient's infertile.

5.1. Ovarian suppression with gonadotropin releasing hormone 

Increased atresia of oocytes is noted in the post-chemotherapy patient as reflected by their diminished ovarian reserve [27], [28], [59], [95], [96]. The notion that hormonal suppression of ovarian function prior to chemotherapy minimizes cytotoxic injury has led to the use of gonadotropin agonists as adjuvant therapy [2], [28], [97], [98]. Early animal studies using mice suggested that by suppressing testicular function with gonadotropin releasing hormone agonists prior to exposing their adnexa to gonadotoxic agents, one might be able to decrease permanent damage [99]. Similar observations were noted in female rodents in which D-Trp6-lutenizing hormone-releasing microcapsules protected rat adnexa from cyclophosphamide toxicity [28], [76]. An extension of these observations to the female rhesus monkey demonstrated that adjuvant treatment with gonadotropin releasing hormone (GnRH) decreased follicular loss mediated by cyclophosphamide [100]. Furthermore the protective effect was observed to persist post-treatment as reflected by a decreased rate of follicular loss. At least by some retrospective reports such a protective effect of GnRH is seen with its pre and simultaneous administration at time of chemotherapy [97]. Menses and spontaneous ovulation was reported in 93.7% of Hodgkin and non-Hodgkin lymphoma patients within 3–8 months after treatment with concomitant GnRH agonist administration during chemotherapy with MOPP/ABVD (MOPP=mechlorethamine-alkylating agent, Oncovin or vincristine vinca alkyloids and interferes with cell-division, procarbine and prednisone, ABVD=adriamycin/doxorubicin, bleomycin, vinblastine and decarbazine) followed by mantle field radiation therapy. By contrast only 39% of the non-GnRH chemotherapy group resumed menses and ovulatory function in a comparable period [97]. However only – patients were in the treated and – in the untreated group.

Of note, at least one study has shown that GnRH suppression of gonadal function as measured by suppressed gonadotropin levels, has not minimized gonadotoxicity in the treatment of Hodgkin's disease [101]. In this study following chemotherapy profound oligospermia and amenorrhea was observed even in the setting of confirmed gonadotropin suppression pretreatment with a GnRH agonist.

The cytotoxic effects of cyclophosphamide and other alkylating agents manifests as primordial follicle atresia [43], which theoretically was ameliorated by suppression of the pituitary gonadotropic axis by some unknown mechanism [28], [98]. The mechanism of this presumed protective effect may not merely be related to the suppression of gonadotropin stimulation of receptors as demonstrated conflicting results from various GnRH antagonist studies. Recently it in one series of studies it has been demonstrated that cetrorelix, a GnRH antagonist can rescue ovarian primordial follicles in mice from cyclophosphamide gonadotoxicity in a dose-dependent fashion when started pretreatment [28]. Control groups of mice treated with a similar dose of cyclophosphamide demonstrated 39% more follicular atresia in the absence of cetrorelix pre- and co-treatment. The protective effect of cetrorelix appeared to diminish with increasing doses of the alkylating agent suggesting a cumulative dose and dose-dependent effect. By contrast Danforth et al. demonstrate a destructive effect of GnRH antagonists such as cetrotide on primordial follicles in mice [102]. Both these studies are intriguing and warrant further investigation. Pereyra Pacheco et al. reported that GnRH co-administration in adolescent cancer patients undergoing chemotherapy resulted in all patients resuming menstruation on completion of their chemotherapy and cessation of GnRH administration [103]. By contrast control patients who did not receive GnRH during chemotherapy displayed hypergonadotrophic hypoestrogenic amenorrhea. Collectively these studies suggest that it might prove beneficial for a young patient with newly diagnosed breast cancer to at least consider hormone mediated gonadal suppression prior to starting chemotherapy for preservation of their fertility. However, as none of these studies were randomized or placebo controlled, inherently different groups of patients may have opted for gonadal suppression. In the absence of placebo controlled randomized studies gonadal suppression must still be viewed as unproven, and not standard of care.

5.2. Embryo cryopreservation 

In 1972 preimplantation cryopreservation techniques of mammalian embryos were reported. The first successful frozen embryo transfer however, was not done until 1983 by Trounson and Mohr [104]. Today utilization of frozen embryos in assisted reproductive medicine is well established and has been employed for the past 20 years [105], [106], [107], [108], [109], [110], [111], [112], [113]. This technology has been successfully used in the treatment of cancer patient's chemotherapy [114]. In a case report from 1989 a 22-year-old woman with chronic myeloid leukemia was hormonally stimulated to generate multiple oocytes that were then fertilized in vitro, frozen and transferred back into this patient post-chemotherapy and radiation treatment 4 years later with a resultant successful pregnancy with delivery of a normal infant. In vitro fertilization in cancer patients for embryo “banking” for future use has been reported by several other groups [6], [9], [110]. The CDC report of 2003 showed a delivery rate on average of 29.4% per embryo thaw in patients <35 years age to 16.5% in patients >40 years age in US IVF programs. As patients undergoing IVF with embryo cryopreservation are generally not infertile, pregnancy rates might well be higher than those achieved by couples who potentially have poorer oocyte and sperm quality. While embryo cryopreservation is a proven option for breast cancer patients several concerning issues must be addressed.

5.3. In vitro fertilization 

In our experience, there is generally a 6–8-week period of time between the patient's definitive surgery, and the time when chemotherapy is planned. Ideally this is the time when an elective cycle of IVF can be undertaken in order to cryopreserve embryos. A typical IVF cycle consists of a series of tightly organized steps starting with a preparatory phase during which a patient's endogenous hypothalamic–pituitary reproductive axis is shutdown with continuous exogenous GnRH agonists or GnRH antagonists. With continuous exposure to a GnRH agonist the endogenous secretion of gonadotropins is shut down, this approach therefore is functionally the same as employing a GnRH antagonist. The end result is prevention of a premature luteinizing hormone (LH) surge which would result in triggering inappropriate ovulation during the following step of IVF, which is controlled ovarian hyperstimulation [40]. If premature ovulation were to occur then the oocytes could be immature with resultant cancellation of the cycle and lost treatment time for the patient. Thus suppression of the hypothalamic–pituitary reproductive axis is a crucial step to this process.

During the next phase ovarian follicular development and maturation is mediated by controlled ovarian hyperstimulation (COH) using exogenous gonadotropins, usually follicle-stimulating hormone (FSH). On average patients require 7–12 days of COH to obtain ovarian follicles that are near maturation and ready for retrieval. When the desired response is obtained the third phase consists of triggering final follicle maturation by administering human chorionic gonadotropin (hCG), 36h after which the final step of the procedure is completed which is the surgical transvaginal retrieval of oocytes. These oocytes are then fertilized in vitro to obtain embryos for intrauterine transfer, or cryopreservation as in the case of breast cancer patients who are about to initiate chemotherapy. This entire procedure can take approximately 3–4 weeks depending on the range of response to medication by individual patients, and where the patient is in her menstrual cycle when she presents for fertility evaluation and treatment. This time frame can however be extended if any complications from the treatment ensues therefore delaying initiation of chemotherapy. A primary complication seen in some patients who overly respond to COH is elevated estradiol concentrations and the development of large numbers of ovarian follicles and significant ovarian enlargement. This can lead to ovarian hyperstimulation syndrome (OHSS) which involves third spacing of exudative fluid into the abdominal cavity and in severe cases pleural space, with intravascular depletion, hemoconcentration and risk of deep venous thrombosis if untreated by intravenous fluid replacement. In some cases liver function tests become mildly elevated, potentially causing a problem with the ability to initiate chemotherapy. Symptoms of abdominal pain and severe bloating and distension typically start a week following hCG administration and persist for approximately 1 week in non-pregnant patients as would be the case for breast cancer patients. Recent reports by Humaidan et al. and Kolibianakis et al. in separate studies suggest that GnRH agonists may be used to trigger ovulation following ovarian stimulation [115], [116]. Although there appears to be a decreased ongoing pregnancy rate with this protocol, the estradiol levels were significantly lower in patients that received GnRH for ovulatory trigger. Collectively any complications from IVF could result in a delay in initiating chemotherapy. However one of the potentially most important concerns with ovarian hyperstimulation syndrome is increased estradiol levels that can be detrimental to a breast cancer patient with an estrogen receptor positive tumor. Such protocols should only be initiated in direct consultation with the oncologist. It would be prudent to consider alternate stimulation protocols such as using letrozole (discussed elsewhere) to relatively decrease estradiol levels.

After the patient is cleared for pregnancy, presumably having been unable to conceive on her own, or following the occurrence of POF, transfer of the embryos into her uterus can be performed. Aspects of preparing the uterus for the embryo transfer need to be taken into particular consideration for breast cancer patients. The endometrial lining of the uterus is prepared by hormonal stimulation employing fairly high-dose estrogen in advance of thawing and transferring the embryos. This would obviously be of concern in patients with ER positive tumors and decisions pertaining to the theoretical level of risk this would pose to the patient would need to be coordinated as a team effort consisting of the oncologist, reproductive endocrinologist, and the patient. Patients who are felt to be poor candidates for pregnancy may consider employing gestational carriers to carry a pregnancy for them, as will be discussed later in this chapter.

As previously indicated, the question of hormonal stimulation in breast cancer patients may be of significant concern since 30–50% of breast tumors in premenopausal women are estrogen receptor positive [ER+] [117], [118], [119], [120], [121], [122]. If histopathological classification of tumor subtype demonstrates estrogen receptors then standard ovarian stimulation may cause concern due to the high circulating estradiol levels leading to possible expansion of clonal populations of estrogen responsive tumor cells [119], [123], [124], [125]. While non-stimulated ovulatory cycles may be used to generate embryos for cryopreservation, the yield from such a cycle typically results in a single embryo [14], [24], [126], [127], [128], [129]. Although a viable option, this approach may not lead to a satisfactory “preservation” of future fertility, since the likelihood of survival of a single cryopreserved embryo in the best hands is at best 80–90%, and the implantation rate of a single thawed cryopreserved embryo is approximately only 10%. Thus to ensure a higher probability of success it is useful to generate and freeze larger number of embryos.

Standard ovulation induction protocols that have been developed for standard IVF treatment will optimize the number of embryos obtained for the patient. Concerns related to their use is that peak estradiol levels are in the 2000–3000pg/mL range, as compared to peak levels of 250–500pg/mL in the normally cycling premenopausal woman. It is unknown what impact if any of the 11 days or so elevation in estradiol levels will have on the breast cancer patient who is awaiting chemotherapy. In our experience oncologists feel that in light of the fact that these patients desperately want to attempt to maximize the likelihood that they will be able to have biological children in the future, the unknown risk may be worth it to them. In addition, one could argue that the brief duration of estrogen exposure during the IVF cycle pales in comparison to the long-term elevations of estradiol and progesterone levels that occur during pregnancy.

Alternative ovarian stimulation protocols that may reduce estradiol exposure, and then theoretically not stimulate residual cancer cell growth are under investigation [14], [24]. Selective estrogen receptor modulators such as tamoxifen have been employed with minimal side effects to facilitate controlled ovarian stimulation thereby promoting the development of a greater number of primordial follicles making it possible to harvest a larger number of eggs relative to an unstimulated cycle [6]. Since tamoxifen, an anti-estrogen that is a non-steroidal is widely in use for the treatment and prophylaxis of breast cancer [123], [124], [125], [126], [127], [128], [129], [130], [131], [132], the idea of employing this medication for ovarian stimulation in breast cancer patients is enticing. Oktay et al. [14], [24] examined the utility of tamoxifen 60mg per day for 6.9+0.6 days for ovarian stimulation in breast cancer patients (TAMIVF) relative to natural cycle in vitro fertilization (NCIVF). Their results suggest that tamoxifen stimulation does indeed generate more embryos; however notably, their study did not demonstrate any significant difference in pregnancy rates between the two groups (2/6 TAMIVF vs. 2/5 NCIVF). None of the breast cancer treated groups showed a recurrence of disease during the 15-month follow-up period, but clearly this is inadequate duration of follow-up on which to base any conclusions. While this implies a possibly safe option for ovarian stimulation and embryo cryopreservation in such patients, of interest the estradiol concentrations in the treated patients was significantly greater than that of women undergoing natural cycle IVF. The 60mg tamoxifen dose was chosen in order to theoretically counteract the higher estradiol levels that were anticipated, however it is unknown whether the high estradiol levels may worsen prognosis, or what amount of estradiol the breasts are actually being exposed to with the use of that dose of TAM. Clearly there are no data regarding the impact of ovulation induction on the long-term prognosis of these patients. Regardless the most important point with ovarian hyperstimulation syndrome is increased estradiol levels that can potentially be detrimental to a breast cancer patient.

Aromatase inhibitors have also been used to determine whether these might be used for ovarian stimulation in lieu of selective estrogen receptor modulators (SERM) circumventing the concern of elevated estradiols [24]. These drugs may in fact serve a dual purpose. Breast cancer research suggests that aromatase inhibitors may indeed be superior to SERM's such as tamoxifen in the long-term suppression of breast cancer recurrence in postmenopausal women [132], [133]. Furthermore the aromatase inhibitors do not result in significant elevations in estradiol levels when used for ovulation induction alone. Initial studies using letrazole [24] with low dose follicle-stimulating hormone for ovarian stimulation in breast cancer patients indicates that a comparable number of follicles develop in letrazole/FSH groups relative to tamoxifen/FSH groups (Table 5 from [14]). Moreover further follow-up of breast cancer patients in Oktay's study showed no increased rates of recurrence with letrozole use.

Table 5. Comparison of cycle characteristics and embryo yield among Tam-IVF (12 patients, 13 cycles) TamFSH-IVF (seven patients, nine cycles), and letrozole-IVF (11 patients, 11 cycles) patients.a.

Adapted with permission from Oktay et al. [14].

	Variable	Mean±standard deviation			P
		Tam-IVF (a)	TamFSH-IVF (b)	Letrozole-IVF (c)	a vs. b	a vs. c	b vs. c
	Age, years	36.6±1.6	38.3±1.9	38.5±1	NS	NS	NS
	Baseline FSH, mU/mL	9.4±1.5	9.4±1.5	6.2±1.1	NS	NS	NS
	Peak E2, pg/mL	419±39	1182±271	380±57	<.05	>.05	<.05
	Total follicles, No.	2±0.3	6±1	7.8±0.9	<.01	<.001	>.05
	Follicle >17mm, No.	1.2±0.1	2.6±0.4	3.2±0.4	<.05	<.001	>.05
	Total oocytes, No.	1.7±0.3	6.9±1.1	12.3±2.5	<.05	<.001	>.05
	Mature oocytes, No.	1.5±0.3	5.1±1.1	8.5±1.6	<.05	<.001	>.05
	Total embryos, No.	1.3±0.2	3.8±0.8	5.3±0.8	<.05	<.001	>.05

Abbreviations: Tam-IVF, tamoxifen alone followed by in vitro fertilization; TamFSH-IVF, tamoxifen combined with low-dose follicle-stimulating hormone followed by in vitro fertilization; letrozole-IVF, letrozole combined with follicle-stimulating hormone followed by in vitro fertilization; NS, not significant; E2, estradiol.

a One patient underwent both TamFSH-IVF and letrozole-IVF treatments. Peak E2 as measured on the day of human chorionic gonadotropin administration. TamFSH-IVF results in significantly higher peak E2 levels compared with Tam-IVF and letrozole-IVF.

This study also confirmed a lower estradiol concentration with letrozole treatment, though still higher than one would expect in a natural menstrual cycle. However, recently the manufacturer put out an alert suggesting that letrozole not be used in ovulation induction due to concerns about teratogenicity. A disadvantage of both the TAM and letrozole employing ovulation induction protocols is that we have little knowledge about the embryo quality resulting from these treatments. Data on patients undergoing embryo cryopreservation with standard IVF regimens also appear to result in significantly more embryos cryopreserved [9], [134]. In addition, since most patients undergoing IVF with embryo cryopreservation prior to chemotherapy must wait at least 2 years prior to conception. As some will spontaneously conceive, only those who have developed ovarian failure use their cryopreserved embryos. Therefore the pregnancy rates to be expected for these treatments are unknown. Obviously it will be vital to follow patients who have undergone IVF prior to chemotherapy both to ascertain which treatments are most effective from the standpoint of resulting in live births, but also to determine if there is a detrimental effect related to cancer recurrence.

5.4. Oocyte cryopreservation 

In addition to embryo cryopreservation, more recently oocyte cryopreservation is evolving as an alternative for fertility preservation [12], [111], [135], [136], [137], [138], [139], [140], [141], [142], [143], [144], [145], [146], [147], [148]. The first frozen oocyte related birth was reported by Chen [139]. Subsequently Porcu et al. reported the first live human birth from ICSI fertilization of cryopreserved oocytes [145]. Some of the concerns regarding oocyte cryopreservation include cryodamage of cytoarchitecture from the freezing process with resultant aneuploidy, highly variable low survival rates from a freeze–thaw cycle (30–95%), and variable fertility rates ranging from 14 to 70% with our current technology [59], [135], [141], [149], [150]. There have been several case reports and series of pregnancies resulting from oocyte cryopreservation [135], [136], [137], [148]. To date fewer than 300 pregnancies from cryopreserved oocytes have been attained. With new freezing techniques and media, the survival rate of freeze–thawed oocytes has relatively stabilized and improved to around 68% with a fertilization rate of 48% however the pregnancy rate remains very low at 1.7% [59].

Advantages of oocyte freezing include the ability to harvest oocytes from women while they are young and saving these for fertilization at a later age. The fact that the oocytes are retrieved from a younger patient would theoretically yield healthier eggs that may then be subsequently fertilized and transferred into the patient's uterus later in life when pregnancy is desired despite the chronologic aging of the patient. Another attractive aspect of this option is that if the patient is single at the time of oocytes harvest and freezing or if her partner is unable to produce sperm at the time of oocytes retrieval, one might choose to just freeze the gametes until a future suitable time presents for fertilization. Some of the cellular problems associated with cryodamage include mitotic spindle breakage, mitochondrial membrane damage [150], and adverse effects on the zona pellucida. When using thawed oocytes intracytoplasmic injection (ICSI) of sperm nuclei appears to facilitate fertilization [141]. To avoid transfer of abnormal embryos preimplantation genetic diagnosis has been successfully employed [137]. Ongoing research is trying to identify the most ideal cryoprotectant media components and protocols to minimize damage to oocytes and allow for successful thaw, fertilization and embryo transfer [141], [149]. While it has been proposed that immature oocytes may be more resistant to damage from freezing due to the absence of a metaphase spindle, problems with this approach include promoting in vitro maturation and development. In unstimulated follicle cycles Toth et al. demonstrate in vitro maturation rates as high as 58% from prophase I to metaphase II [142]. Survival rates ranged from 15% to greater than 43% depending on the mode of freezing employed. Freezing of mature oocytes in metaphase II are more likely to survive the freeze–thaw cycle, rather than prophase I oocytes [151]; however even with this approach the success rate of fertilization is relatively low at about 13% [94].

The widespread use of oocyte freezing remains controversial and is in the experimental stages in the United States. Oocyte freezing is in more widespread use in Europe. It remains an attractive option for women who do not have a partner and who do not want to use donor sperm in order to undergo in vitro fertilization. At this time however, it is clearly experimental, numbers of patients having undergone the treatment are small, and neither predictable oocyte thaw survival rates nor pregnancy rates are available.

5.5. Ovarian tissue freezing 

In March of 2004 Lee et al. reported the first primate birth from heterotopic transplantation of fresh ovarian tissue in the Macaca mulatta resulting in a live birth of a monkey [152]. On the heels of this report Donnez et al. reported a human live birth after orthotopic transplantation of cryopreserved ovarian tissue [25]. The Donnez report, however, was of a woman who still had an ovary present in situ, and had documented ovulation prior to replacement of the cryopreserved ovarian tissue, so that the pregnancy may in fact not have been from the cryopreserved tissue. Albiet to date no pregnancy may have occurred, one embryo was generated from a heterotopically located ovarian tissue. Nevertheless, both these reports represent advances in the field of reproductive medicine and have moved the field into a new and potentially exciting area of investigation that ultimately may provide patient's with novel fertility preservation options [7], [15], [22], [25], [98], [152], [153], [154], [155], [156], [157], [158].

While the aforementioned options provide limited alternatives to preserve fertility, the concept of ovarian tissue freezing pretreatment potentially lends a host of added advantages. Besides being able to preserve one's fertility, if indeed one could freeze ovarian tissue prior to initiating gonadotoxic therapies, autologous transplantation may allow restoration of hormonal function and the benefits associated with these hormones. In the case of ER+tumors one might opt to have transient replacement of tissue for the purposes of pregnancy following which the tissue may be removed. In patients with no contraindications to estrogen replacement the transplanted tissue may continue to function in a heterotopic or orthotopic location providing the necessary hormonal levels. Ovarian transplants in cynomolgus monkeys have been shown to restore menstrual cycles [159]. Furthermore Radford et al., Cajello et al. and Oktay et al. in separate studies demonstrated the feasibility of orthotopic reimplantation of cryopreserved ovarian cortical strips with some restoration of hormonal function in the latter study [156], [157], [158], [160], [161]. Indeed more recent studies in cancer patients with either breast, Hodgkin's, non-Hodgkin's leukemia or a CNS tumor suggest that autologous transplantation of cryopreserved ovarian tissue in either an orthotopic or heterotopic position restores ovarian hormonal function [26]. In all these patients ovarian tissue was cryopreserved prior to the initiation of treatment and replaced on completion of chemotherapy more than 18 months later. More recently Oktay et al. have reported functional heterotopic ovarian tissue transplants in a cervical cancer patient's forearm for up to 2 years [158]. Both menstruation and spontaneous ovulation were noted. In mouse studies orthotopic grafting of frozen ovaries into mice has not only resulted in normal ovarian function with delivery of normal size litters, but also restoration of a normal reproductive lifespan of the adnexal function [144], [162]. Such reproductive results are yet to be demonstrated in humans. Albeit still very much in its experimental stages ovarian tissue cryopreservation and orthotopic or heterotopic transplantation post-treatment may eventually become a routine alternative for breast cancer patients. Currently few centers perform this procedure internationally.

Cryopreservation of ovarian tissue has variable degrees of success with regards to primordial follicle survival with some reports indicting >70% survival of primordial follicles after freezing and thawing [153]. The primary issue remains restoration of ovarian functions subsequent to transplantation since it takes nearly 48h to establish a vascular supply to the grafted tissue [13]. During the interim there is significant hypoxic damage to the tissue as well as ischemic reperfusion injury subsequently. Resultantly primordial follicle survival following transplantation ranges from 5 to 50% [163], [164]. Despite these limitations there have been reports of hormonally functional ovarian tissue as well as maturation of ovarian follicles following autologous transplantation of either fresh or cryopreserved and thawed human ovarian tissues [156], [160]. The main concern of autotransplantation in cancer patients is the possibility of cancer cell transmission. While some studies suggest the apparent safety of ovarian transplantation in lymphoma patients [158], it has been argued that this approach may not be feasible in breast cancer patients secondary to concerns of disease recurrence [165] or occurrence of related malignancies such as from BRCA mutations. Obviously, autologous transplantation of ovarian tissue is not an option for patients with metastatic disease.

While several groups have attempted cryopreservation of ovarian tissue with auto grafting into patients, hitherto no clear pregnancy has been reported despite documentation of various degrees of resumed ovarian endocrine function as monitored by measuring estrogen, progesterone and gonadotropic hormone levels [166]. In the M. mulatta monkey study, Lee et al. demonstrate that abdominal transplants of ovarian tissue wedges into 54 different sites in 4 monkeys results in functional tissue with cyclic production of estrogen, progesterone and FSH as monitored by radioimmunoassay [152]. Furthermore nearly 50% of the abdominal grafts produced follicles, 16 of which were retrieved and fertilized by ICSI. Two morulae a 5 cell and 8 cell embryo were obtained and transferred to the oviduct of the monkey resulting in a single live birth. While these results are promising, it must be noted that the ovarian tissue was not cryopreserved prior to transplantation. Thus this tissue was not exposed to any of the ischemic events and cryodamage that ovarian tissue banking would entail. However frozen-thawed cortical ovarian strip grafts placed in a patients forearm or abdominal muscle have been shown to resume function as monitored by production of gonadotropins, albeit for only 3–4 months [161]. In other human studies Oktay et al. [157] demonstrate the feasibility of autologous transplants of ovarian tissue behind pelvic peritoneum in a hypothalamic amenorrheic patient. This patient showed cyclic estrogen and progesterone production from the graft tissue as well as development of a follicle when exogenous gonadotropins were administered. Survival of the graft in this study was monitored for 6 months with the tissue producing follicles in response to exogenous gonadotropins stimulation confirming viability of the same. Similar graft function was also documented in a cervical cancer patient with stage III b disease [7]. Six weeks after the transplantation, the patient's follicle-stimulating hormone (FSH) and luteinizing hormone (LH) were 47 and 35mIU/mL, respectively. Four months after transplantation, mean (SE) FSH was 13.6 (0.54) (range, 11–18.1mIU/mL), with a further decrease to 8.6 (0.4) (range, 6.2–12.4mIU/mL) (P<.001). Similarly, LH decreased from 18.3 (1.5) (range, 9–29mIU/mL) to 12.8 (0.8) (range, 6–26mIU/mL) (P=.002) by 7 months postoperatively. In order to cryopreserve embryos, the patient underwent controlled ovarian hyperstimulation with harvest of 3 oocytes. These were retrieved percutaneously. Two oocytes from 15.5-mm follicles were post-mature; there was no oocyte in the 11.5-mm follicle. The 14-mm follicle contained an oocyte in metaphase I. Following in vitro maturation of this oocyte, fertilization was attempted with ICSI but failed. Obviously these results collectively demonstrate that ovarian tissue transplantation and cryopreservation while a promising area of research remains experimental.

Numerous ethical and medical considerations need to be addressed before embarking on cryopreservation and transplantation of ovarian tissue in breast cancer patients [22]. As mentioned, although transplantation of cryopreserved ovarian tissue from a cancer patient post-treatment may preserve fertility, one must be cautious as to whether by transplanting un treated ovarian tissue are any malignant cells being reintroduced into the patient. While several laboratories are attempting to develop protocols to screen for this, with disseminated metastatic disease this is a matter of significant concern [13], [165]. In patients who have BRCA1/BRCA2 mutations, reintroducing ovarian tissue would also restore the potential risk of ovarian cancer. Along the same lines, in the event a patient has an estrogen responsive tumor, then what are the risks associated with resumption of ovarian function in such patients following ovarian grafting and more so what are the risks associated with the estrogen levels of the ensuing pregnancy? At least in one analysis in over 400 patients followed through a pregnancy, associated estrogen levels did not appear to alter the course of the disease progression [167], [168], [169]. Other frontiers of research to explore include heterologous transplantation of ovarian issue and the feasibility of such an approach.

Amongst several ethical considerations one must ask what the risk of disease recurrence is in such patients post-treatment and the risk of relapse resulting in the patient succumbing to the disease leaving an infant without a parent. By one account breast cancer survivors with node negative disease have been reported to have a 5-year survival rate of 85% and 51% for patients who have nodal involvement [22]. These issues are influenced by several individual patient characteristics but should be addressed in pretreatment counseling so that patients may make informed decisions regarding what venue of care to pursue.

Another cause for concern is that a number of facilities throughout the country are commercially advertising ovarian tissue cryopreservation not only for cancer patients but to women who are concerned about future loss of fertility. The vast majority of these facilities have done no preclinical or clinical research to determine whether in fact their procedures can result in viable ovarian tissue or pregnancy. The American Society for Reproductive Medicine has clearly stated that ovarian tissue and oocyte cryopreservation are investigational and should be undertaken under the auspices of an IRB only.

5.6. Donor embryos, gestational carriers and adoption 

Despite the various options available for fertility sparing in cancer patients, success is not a guarantee. Alternatives that must be included in a physician–patient discussion include the option of donor embryos and adoption. Often even with optimal treatment and ovarian stimulation conditions, the number and quality of eggs retrieved from cancer patients may be sub-optimal [9]. An alternative choice for such patients may be donor eggs or embryos. Accepting eggs for fertilization from a healthy young donor often will allow the individual to carry a pregnancy fertilized by their partner's sperm. If on the other hand the situation is compounded by male factor infertility in addition to the patient's cancer treatment, then one might opt to pursue donor embryos. Such a process may be followed by involving an anonymous donor or a mutually agreed upon known donor. The ethical and legal issues surrounding such a process are fairly complicated. Thus one must proceed with the appropriate counsel and guidance from their fertility specialists. Adoption remains a viable and preferred option for many couples, with its major limitation being cost, which averages over $30,000 per adoption in the US. Here again there may be individual differences in state law and one must contact an adoption agency to facilitate this process.

5.6.1. Gestational carriers (GC) 

Pregnancy leads to breast engorgement and difficult surveillance. Therefore for some breast cancer survivors, the prospect of pregnancy may cause too much anxiety for the patient and concerns for the oncologist. In these cases the use of a gestational carrier is possible. A gestational carrier is an individual who carries a pregnancy for a woman who for a variety reasons such as cancer, surgically absent uterus or morbidities that would preclude pregnancy is unable to do so herself [92], [170], [171]. A GC is typically genetically unrelated to the embryo, although this may not be true in cases where a sibling may opt to carry the pregnancy for her sister or brother. The legal and emotional issues that surround the use of a GC are highly complex and variable between individual patients. The cost of using a GC can range anywhere from $15,000 to well over $35,000 in addition to the cost of the care associated with the pregnancy itself [172]. Furthermore there are several medical issues such as compatibility and potentially transferable risks of pregnancy such as preeclampsia, gestational diabetes and peripartum cardiomyopathy [173], which are believed or proposed to have etiologic contributions from the developing trophoblastic cells of the placenta [174], [175]. Thus these factors and the associated risks would be carried by transfer of the embryo into a GC. The GC must also be prepared for embryo transfer by giving her hormonal stimulation with estradiol to prepare the endometrial uterine lining. There are of course obvious risks with the use of hormonal stimulation such as coagulopathy and therefore it is critical to evaluate the medical history of the GC so as to exclude patients with a history of any coagulopathy, uterine anomalies or even cardiac conditions such as hypertension or renal disease since these would further compound the obstetrical risks and complications for a GC [92]. The legal issues that may influence the intended parents vary from state to state. Thus this is not an arrangement to be lightly entered into. If the breast cancer patient is unable to bank her own embryos prior to treatment, she may want to use donated embryos in a GC to establish her family. Again the ethical, emotional and legal issues for such an arrangement need to be well explored and the patient appropriately counseled prior to entering into such an arrangement. GCs have been successfully employed by several cancer patients for a variety of reasons [170], [171].

Finally, other lines of research include attempts to generate gametes from embryonic stem cell lines generated from ones somatic cells by transferring the recipient's diploid nucleus to an enucleated donor oocyte [176], [177], [178], [179], [180]. While such studies clearly have significant ethical and scientific considerations, these may provide promising new alternatives for fertility preservation and restoration in future. Collectively, there are several options available to young breast cancer patients with continually emerging new medical technologies for fertility preservation and possibly even restoration (Table 6).

Table 6. Summary of fertility preservation options in young breast cancer patients.

	Treatment approach		Pretreatment options	Post-treatment options with gonadal failure
	Surgery+	Surgery+	GnRH gonadal suppression with agonists	IVF with donor oocytes
	Chemotherapy	Radiation ±		Gestational carriers
		Chemotherapy
	Chemotherapeutic optionsa		IVF with embryo cryopreservation	Donor embryos
				±Gestational carriers
	↑ Gonadotoxicity	Alkylating agents	Experimental Approaches	Adoption
		Vinca alkaloids	Oocyte cryopreservation
			Ovarian tissue freezing
		Antimetabolites
		Anthracyclines-antibiotics
		Taxanes-antitubulins
	↓ Gonadotoxicity	Hormones
		HER2/Neu antibody

This table summarizes the various alternatives for fertility preservation in newly diagnosed young breast cancer patients.

a Obviously the choice in chemotherapeutic agents will depend on the tumor stage, grade and histological type. When presented the opportunity to choose from equally efficacious treatment approaches, one should in consultation with their oncologist, consider agents shown to have less gonadotoxicity (Table 3). Also summarized are pre- and post-treatment fertility preservation options.