11.1.3 Long-term endocrine sequelae of cancer therapy

Over the past 40 years cure rates for childhood malignancies have improved at a remarkable pace. Overall 5-year survival improved from less than 30% in 1960 to more than 70% in 1990. With increasing cure rates, came recognition of the long-term detrimental effects of radiotherapy and chemotherapy on multiple organ systems. Five-year survival has, however, altered little over the last decade. To improve upon recent successes will probably necessitate the use of more complex treatment regimens, resulting in a higher prevalence of adverse treatment-associated long-term effects in these individuals.

Over the next decade the long-term sequelae of childhood cancer therapy is likely to have a significant financial and workforce demand on health services. It is estimated that one in 640 adults aged 20–39 years in the USA is currently a survivor of childhood cancer, and in the UK by 2010 one in 715 young adults is estimated to be a survivor of childhood cancer. Epidemiological data from the American Childhood Cancer Survivors Study (CCSS) reported survivors of more than 5 years to have a 10.8 fold excess in overall mortality (1). The majority of deaths (67%) relate to recurrence of the original tumour. After exclusion of deaths relating to recurrence or progression of the original tumour, mortality rates remained significantly increased. Standardized mortality rates for second malignancies (SMR 19.4), cardiac disease (SMR 8.2), pulmonary disease (SMR 9.2), and other causes (SMR 3.3) were significantly elevated. Long-term endocrine sequelae are particularly prevalent in childhood cancer survivors with 43% of the CCSS cohort reporting one or more endocrinopathies (2). Endocrine late effects include disturbances of growth and puberty, hypothalamopituitary dysfunction, hypogonadism, subfertility, thyroid dysfunction, benign and malignant thyroid nodules, hyperparathyroidism, and reduced bone mass (Table 11.1.3.1–11.1.3.3).

The impact of childhood cancer and treatment thereof has long been recognized to impair height velocity and final height (Table 11.1.3.1). Growth velocity is frequently impaired at diagnosis and during treatment of childhood malignancies, reflecting the acute illness, poor nutritional status, and ongoing cancer therapy. In addition, perturbations of the endocrine system, including radiation-induced hypothyroidism, precocious puberty, and growth hormone deficiency (GHD) impact adversely on growth. Survivors of childhood malignancies who previously received cranial irradiation achieve final heights significantly below those predicted from parental heights, even with irradiation doses as low as 18 Gy. Although radiation GHD is an obvious cause for the abrogated growth it is not universally present in all children with impaired growth velocity, thereby implicating additional mechanisms. A subanalysis of the CCSS survivors with brain tumours revealed 40% of patients to have a final height below the 10th percentile (3).

Spinal irradiation has a negative impact on growth above that of cranial irradiation alone, and relates directly to a reduction in spinal growth (4). Leg length SDS in patients who receive cranial and craniospinal irradiation are equivalent, whereas spinal growth is impaired only in the latter patients (4). The greater impairment of spinal growth results in disproportion, reflected by an increase in the leg length to sitting height ratio. The impact of spinal irradiation on the skeleton correlates with age; the younger the individual is at the time of irradiation the greater is the impairment of spinal growth and the greater the degree of disproportion (4). This observation simply reflects the fact that the younger an insult to growth occurs, the greater the loss in growth potential. Disproportion may be further amplified by the use of growth hormone replacement therapy in patients found to be GHD as although growth hormone replacement impacts favourably on growth of the long bones, the spine remains relatively resistant to the growth promoting effects of growth hormone (5). In children who received spinal irradiation, growth should be monitored by leg length velocity.

Catch-up growth is frequently observed, without growth-promoting intervention, once active treatment has been completed and remission achieved. Although the effect of cytotoxic chemotherapy on growth remains contentious, there is a suggestion that subsequent growth may be attenuated. Additionally, chemotherapy may potentiate the growth impairment resulting from craniospinal irradiation, but requires further study. Although the pathophysiological mechanism by which chemotherapy influences growth is unclear a reduction in growth factors including insulin-like growth factor (IGF-1), increased sensitivity of bone to irradiation damage, and a direct action on the growth plate have been postulated. High-dose cranial irradiation leads to gonadotropin deficiency, however, at lower doses results in early onset of puberty. The age of onset of puberty in children correlates to age at cranial irradiation. An early age at onset of puberty leads to premature completion of puberty, thereby restricting the time for growth and growth-promoting therapy in these individuals.

Radiation-induced hypopituitarism of varying degrees is a well-recognized sequela of external beam irradiation when the hypothalamopituitary axis falls within the field of treatment (Table 11.1.3.1). Hypopituitarism has been reported in patients irradiated for pituitary and parasellar tumours, intracranial malignancies, soft tissue sarcomas of the facial bones, and nasopharyngeal carcinomas, as well as patients who received cranial irradiation as part of their regimen for treatment of haematological malignancies, or total body irradiation (TBI) as preconditioning for bone marrow transplantation (BMT).

Selective radiosensitivity of the neuroendocrine axes means that growth hormone secretion is almost exclusively the first of the anterior pituitary hormones to be affected (6–8). Prospective data following irradiation of pituitary tumours and nasopharyngeal tumours suggest that deficiency of the gonadotropins occurs next, followed by corticotropin, with thyrotropin being relatively resistant to irradiation damage (6). Transiently elevated prolactin levels are frequently observed after hypothalamopituitary irradiation in excess of 40 Gy (7, 8). The hyperprolactinaemia is usually clinically silent and tends to return to baseline values over the following few years. Posterior pituitary dysfunction following irradiation is not described.

The radiobiological impact of a radiation schedule on hypothalamopituitary function is dependent on the total dose, fractionation, and duration over which the radiation is administered (6). The proportion of patients 5 years postirradiation of the hypothalamopituitary axis when administered a fractionated dosage of approximately 40 Gy during childhood, would be expected to be in the region of 60–100%, 30–60%, 20–40%, and 5–25% for growth hormone, gonadotropin, corticotropin, and thyrotropin deficiency, respectively (Fig. 11.1.3.1). Few data are available for adults exposed to hypothalamopituitary irradiation for nonpituitary brain tumours. One study reported the prevalence of growth hormone, gonadotropin, corticotropin, and thyrotropin deficiency at a median of 38 months to be 32%, 27%, 21%, and 9% respectively. In addition to the incidence of anterior pituitary hormone deficits, radiation dose also determines the speed of onset and severity of hormone deficits.

Fractionation is an important factor to consider when assessing the radiobiological impact of a radiation dose; in general division into a greater number of fractions of smaller size administered over a longer duration is less likely to result in hypopituitarism. The progressive development of anterior pituitary hormone deficits necessitates prolonged follow-up with yearly assessment of pituitary function in patients who have received cranial radiation.

With radiation doses less than 50 Gy, hypothalamopituitary hormone deficits are attributable to the cumulative damage from the delayed neurotoxic effects of irradiation on the hypothalamus, and secondary pituitary atrophy (Fig. 11.1.3.1). Higher irradiation doses are thought to additionally cause damage directly at the level of the pituitary. Support for the hypothesis that the hypothalamus is more vulnerable than the pituitary to radiation is derived from several sources. Following radiation, normal growth hormone responses to growth hormone releasing hormone (GHRH) may be seen in the setting of impaired growth hormone responses to the ITT, and subnormal gonadal function may be observed in the presence of a normal gonadotropin response to gonadotropin releasing hormone (GnRH). Prolactin levels are frequently transiently elevated, suggestive of a reduction in hypothalamic dopaminergic tone. Insertion of yttrium-90 implants (500–1500 Gy) in to pituitary adenomas results in a lower prevalence of anterior pituitary hormone deficits compared with conventional external beam irradiation (37.5–42.5 Gy); the likely explanation for this observation is that the field for conventional irradiation includes the hypothalamus, which is relatively spared with yttrium-90. The pathophysiological mechanism responsible for radiation-induced hypothalamic damage is unclear and may reflect either vascular or direct neuronal damage.

Although an association has been suggested, evidence for a direct association of chemotherapy with the development of anterior pituitary dysfunction is lacking. A number of chemotherapeutic agents may, however, modulate antidiuretic hormone release from the posterior pituitary resulting in the syndrome of inappropriate antidiuretic hormone. Cisplatin, cyclophosphamide, melphalan, vinblastine, and vincristine have all been implicated, but by no means provide a comprehensive list.

Nearly all children who receive more than 30 Gy to the hypothalamopituitary axis will show blunted peak growth hormone responses to stimulation tests at 5 years (10). In contrast, both children and adults who receive irradiation doses of less than 30 Gy leave a significant proportion of individuals with ‘normal’ stimulated growth hormone responses (9). Studies of spontaneous growth hormone release in these latter children have shown failure of the partially damaged hypothalamopituitary axis to meet the expected increase in demand for growth hormone during growth and puberty. This phenomenon of impaired spontaneous growth hormone secretion and normal stimulated growth hormone levels is termed neurosecretory dysfunction, the clinical relevance of which is currently unknown, but has been associated with an attenuated pubertal growth spurt. In contrast, in adults, compensatory hyperstimulation of a partially damaged growth hormone axis has been shown to maintain normality of spontaneous growth hormone secretion, in the context of partially abrogated stimulated levels to insulin-induced hypoglycaemia.

In individuals who receive irradiation to the hypothalamopituitary axis, growth hormone stimulation tests are frequently discordant; the arginine and GHRH-arginine stimulation tests generally being less attenuated than the ITT (11). Twenty-four-hour growth hormone profiles in individuals who fail all the above stimulation tests show preservation of the normal growth hormone pulsatility and diurnal variation, but profound amplitude reduction. It is probable, therefore, that before development of GHD, defined by stimulation tests, radiation causes dual hypothalamus and pituitary damage, and that normality of spontaneous growth hormone secretion can be maintained by compensatory overdrive of the residual somatotrophs (12).

Diagnosing radiation-induced GHD thus remains complex. During puberty, a failed response to the ITT represents minimal remaining functional reserve of somatotrophs already at near maximal stimulation, and an inability to respond to increased demands. The ITT is thus a good indicator of the need for growth hormone replacement at this stage of development. In the adult, an isolated abrogated response to the ITT may not necessarily be representative of GHD as appropriate spontaneous growth hormone secretion is frequently maintained. A failed response to the GHRH-arginine test, however, is almost always associated with impaired spontaneous growth hormone secretion and therefore truly representative of GHD (6).

Growth hormone replacement is commonly used to optimize final height in children diagnosed with radiation-induced GHD. Most studies have shown improvements in height velocity, however, data have been conflicting. Early studies showed disappointingly small differences in the height loss prevented by the use of growth hormone replacement, and much less than observed in children treated for idiopathic GHD (5). A number of factors contributed to these suboptimal results including spinal irradiation, early puberty, a prolonged interval between irradiation and initiation of growth hormone therapy, and inadequate growth hormone schedules. The predominant factor probably relates to the interval between hypothalamopituitary irradiation and initiation of growth hormone therapy. Since the risk of recurrence of childhood brain tumours is relatively low more than 2 years out from treatment and there is no evidence that growth hormone increases the risk of recurrence of brain tumours (13, 14), it is reasonable to consider growth hormone replacement at this time. The approach of clinicians is variable with some offering growth hormone replacement only to children who demonstrate a reduced peak growth hormone response to stimulation in association with a reduced height velocity, whereas others offer growth hormone replacement on the basis of the biochemistry alone on the premise of preventing height loss which once established may not be fully remediable. Where growth velocity and growth hormone stimulation tests are normal at 2 years post-treatment, growth should be monitored and the growth hormone stimulation tests repeated annually.

To date, there are no data specific to transitional use of growth hormone in adults with radiation-induced GHD, however, it is intuitive to surmise that the beneficial effects of growth hormone replacement in these individuals would be similar to GHD hypopituitary adults of other aetiologies. Before committing an individual who received childhood growth hormone replacement for radiation-induced GHD to transitional growth hormone replacement it is essential to reassess the growth hormone axis. This necessity is derived from the fact that all degrees of GHD are treated during childhood and only those with severe GHD qualify for treatment as an adult. Furthermore, the reproducibility of growth hormone stimulation tests is poor and recent data reassessing growth hormone status in childhood brain tumour survivors showed only 61% retested severely GHD at final height. Growth hormone doses used to treat GHD adults during transition should be aimed at normalizing the IGF-1 level in contrast to the weight-based regimens used during childhood (15).

GHD survivors of childhood cancer are phenotypically indistinguishable from GHD hypopituitary adults of other aetiologies (16). In the former group the relative contribution of GHD is difficult to disentangle from the direct effects of the primary pathology, irradiation, chemotherapy, high-dose glucocorticoids, insufficient exercise, and excess caloric intake. Physiological growth hormone replacement in adult GHD survivors of childhood cancer significantly improves quality of life, with the greatest benefit occurring in the domain of vitality (17). Only minimal beneficial effects on body composition, serum lipids, and bone density are observed (17). No beneficial effect of 18 months’ growth hormone replacement was observed on the spinal bone density of patients who previously received spinal irradiation. These data support a role of GHD in the aetiology of the impaired quality of life of GHD childhood cancer survivors. In keeping with NICE guidance (18) a trial of therapy is therefore appropriate in GHD adult survivors of cancer where quality of life is impaired.

Gonadotropin secretion is the second most frequently affected anterior pituitary hormone following cranial irradiation (7, 8); however, it is infrequent with radiation doses less than 40 Gy. A remarkable increase in incidence occurs with more intensive radiation schedules (8). As observed with other anterior pituitary hormones, the incidence of gonadotropin deficiency is both dose and time dependent. Gonadotropin deficiency following irradiation is present in a continuum from subtle abnormalities detectable only with GnRH testing to severe deficiency with clearly subnormal sex hormone levels.

In addition to gonadotropin deficiency, cranial irradiation doses of less than 50 Gy can result in precocious or early puberty in children (Table 11.1.3.1). Both genders are affected with irradiation doses employed in the treatment of brain tumours (25–50 Gy), whereas lower doses used for prophylaxis in treatment of acute lymphocytic leukaemia results in a predominance of girls developing precocious puberty. A linear relationship between age at irradiation and the age at onset of puberty is observed in patients who received cranial irradiation for brain tumours distant to the hypothalamopituitary axis. The onset of puberty occurs at a mean of 8.51 years in girls and 9.21 years in boys, plus 0.10 years for every year of age at irradiation. The mechanism responsible for early puberty is thought to result from disinhibition of cortical influences on the hypothalamus allowing GnRH pulse frequency and amplitude to increase prematurely. Is has been postulated that the cortical restraint on the onset of puberty is more easily disrupted in girls than boys by any insult, including irradiation. The impact of early puberty in a child with radiation-induced GHD is to foreshorten the time available for growth hormone therapy and thereby restrict the therapeutic efficacy of this intervention. It is for this reason that children with early puberty are treated with a combination of GnRH analogues and growth hormone replacement.

ACTH is more resilient to irradiation-induced damage than either the growth hormone and gonadotropin axes. As with other axes there is a clear dose dependency in the intensity of damage to this axis. There are only occasional reports of ACTH deficiency following TBI (9.0–15.0 Gy) used as preconditioning before BMT (9), or cranial irradiation during treatment of acute lymphocytic leukaemia (18–24 Gy) (19). Even with cranial radiation doses up to 50 Gy only around 3% of children develop ACTH deficiency, though the incidence increases dramatically with doses more than 50 Gy (7, 8). In survivors of childhood brain tumours, ACTH deficiency tends to occur late necessitating continued awareness and screening beyond 10 years after treatment of the primary disease (Fig. 11.1.3.1).

The thyroid axis is thought the least vulnerable of the anterior pituitary axes to radiation-induced damage (7, 8) (Fig. 11.1.3.1). Overt secondary hypothyroidism is uncommon with irradiation doses below 50 Gy (9). Diagnosis of central hypothyroidism is notoriously difficult as the thyroid-stimulating hormone (TSH) level can lie within, below, or slightly above the normal range, with free thyroxine levels in the lower reaches of the normative range or only slightly below. The slightly elevated TSH levels seen in central hypothyroidism are thought to be the consequence of an alteration in the predominant form of TSH secreted, resulting in an alteration in the ratio of bioactive/ immunoreactive TSH. At present, there is no convincing evidence to support the routine use of the TRH test or assessment of TSH surge to improve the diagnostic sensitivity and specificity of central hypothyroidism (21, 22).

As a result of the multimodality treatment regimens employed in the treatment of cancer, damage to the gonadal axis can occur directly at the level of the gonad and centrally at the hypothalamus and pituitary—as discussed above. Damage to the gonads and central structures are not mutually exclusive and it is not uncommon for an individual who has received multimodality cancer therapy to have involvement at both levels. Damage to the gonads can occur from irradiation exposure and cytotoxic chemotherapy (Table 11.1.3.2). Irradiation of the gonads occurs during treatment of gonadal tumours, testicular relapses of haematological malignancies, soft tissue sarcomas of the pelvis, TBI in preparation for BMT, and from scatter during spinal irradiation for certain brain tumours and relapsed haematological malignancies. Damage from cytotoxic chemotherapy is most frequently described following alkylating agents including cyclophosphamide, chloambucil, and mustine; however, nitrosoureas, procarbazine, vinblastine, cytosine arabinoside, and cisplatin have also been incriminated (23). In children, it has been suggested that the chances of maintaining or recovering gonadal function following multimodality cancer therapy are greater for girls than boys (24).

The testis is one of the most radiosensitive tissues in the body. A dichotomy between damage to the germinal epithelium and the Leydig cells is observed; very low doses of irradiation causing significant impairment of spermatogenesis whereas sex hormone production is impaired only with high radiation doses. As a consequence, puberty generally progresses normally in children and secondary sexual characteristics are maintained in the majority of adults who received irradiation to the testis. Testicular volumes are small reflecting damage to the germinal epithelium (9), and should not be relied upon for staging puberty. In contrast to most other tissues, dose fractionation increases gonadal toxicity.

The effect of single fraction low-dose radiotherapy on spermatogenesis is well documented. In general, the most immature cells, spermatogonia, are the most radiosensitive with doses as low as 0.1 Gy causing a significant reduction in sperm count and morphological changes in the spermatozoa (Fig. 11.1.3.2).

The majority of testicular radiation exposure occurs as a consequence of fractionated irradiation, which evidence suggests is more toxic to the germinal epithelium. Fractionated radiotherapy doses of less than 0.2 Gy have no significant effect on spermatogenesis, doses of 0.2–0.7 Gy cause a dose-dependent increase in follicle-stimulating hormone (FSH) and transient reduction in spermatogenesis which recovers within 12–24 months, and doses of 2.0–3.0 Gy frequently result in azoospermia with recovery of spermatogenesis often delayed for 10 years or more.

At the irradiation doses discussed, Leydig cell function is relatively spared, the vast majority of patients having normal testosterone levels albeit frequently at the cost of elevated luteinizing hormone levels. With time, the elevated luteinizing hormone level returns to normal. During adulthood, irradiation doses of 20–30 Gy used for carcinoma in situ in the contralateral testis following unilateral orchidectomy result in overt Leydig cell insufficiency, characterized by a fall in testosterone and a compensatory increase in luteinizing hormone levels. The fall in testosterone, however, is not so great as to require replacement therapy in the majority of adults. In contrast, there is a suggestion that individuals who have undergone a similar treatment regimen for testicular cancer during childhood may be more vulnerable to Leydig cell damage and frequently require testosterone replacement as an adult. It is noteworthy that an irradiation dose of 20–30 Gy will completely ablate the germinal epithelium.

Following TBI during childhood (9.0–15.5 Gy), FSH is elevated in the majority (68–90%) of pubertal and peripubertal boys and luteinizing hormone is elevated in 40–50%, whereas testosterone levels are infrequently low (0–16%) (9). There are no robust data documenting sperm counts, during adult life, in these individuals to determine the proportion that would be spontaneously fertile or fertile with assisted fertility techniques.

The ovaries are irradiated in the management of pelvic tumours, lymphoma, during the spinal component of craniospinal irradiation, and during TBI preconditioning prior to BMT. The effect of irradiation and chemotherapy on the ovary can best be explained by loss of oocytes from a fixed population, which once destroyed can not be replaced (Table 11.1.3.2). The natural history of the healthy ovary is for oocyte number to fall exponentially with ageing. Ovaries of older females are therefore much more sensitive to radiation-induced damage, and a dose of 6 Gy is liable to result in a permanent menopause in women aged 40 years or more. In contrast, in young women it is estimated that 20 Gy over a 6-week period will result in permanent sterility in around 50% (Fig. 11.1.3.3). Higher doses inevitably result in ovarian failure irrespective of age.

Pelvic irradiation during childhood that involves the uterus within the irradiation field leads to changes that result in failure to carry a child (26). In those patients who do conceive the risk of miscarriage and low birth weight infants is greatly increased.

The adverse impact of chemotherapeutic agents on the testis is directed primarily at the germinal epithelium. The extent of damage and potential for recovery of spermatogenesis is dependent on the chemotherapeutic agents used and the cumulative dosage (24). It has been suggested that the adult testis is more susceptible to damage than that of the prepubertal testis. During adult life, however, few studies suggest a relationship between age and risk of gonadal failure (24). In general combination chemotherapy is more toxic than use of single agents and the induced azoospermia is less likely to recover.

Although subnormal testosterone levels (<7 nmol/l) are infrequent there is irrefutable evidence for a more subtle impact of chemotherapy on Leydig cell function (28). The most frequent abnormalities of Leydig cell function are an elevated basal and GnRH stimulated luteinizing hormone level in the setting of a normal or low normal testosterone level. Physiologically, luteinizing hormone pulse amplitude is increased whilst pulse frequency remains unaltered. The compensatory increase in luteinizing hormone means testosterone replacement is rarely necessary. In 135 men treated with high-dose chemotherapy for Hodgkin’s disease, 31% were found to have an elevated luteinizing hormone in association with a testosterone level in the lower half of the normal range or frankly subnormal, and a further 7% showed an isolated raised luteinizing hormone level (28). These biochemical abnormalities support the hypothesis that a significant proportion of men treated with cytotoxic chemotherapy have mild testosterone deficiency. Studies of testosterone replacement in these individuals with elevated luteinizing hormone and testosterone levels within the lower reaches of the normative range have failed to showed significant benefits to date.

Ovarian damage presents clinically with amenorrhoea with or without symptoms of oestrogen deficiency, or failure to progress through puberty. Hormonally, the gonadotropins may be grossly elevated with an unrecordable oestradiol level, or show moderate elevation of the gonadotropins in association with a midfollicular oestradiol level. Similar to irradiation-induced ovarian damage, the susceptibility of the ovary to chemotherapeutic damage, speed of onset of amenorrhoea, and the potential for recovery is dependent on age and cumulative dosage (24). Smaller doses of chemotherapy are thus required with increasing age to induce ovarian failure.

In women with breast cancer treated with multiagent chemotherapy including cyclophosphamide, the average dose of cyclophosphamide to induce amenorrhoea in women in their twenties, thirties, and forties was 20.4, 9.3, and 5.2 g respectively. Intuitively, prepubertal and pubertal girls would be assumed to be at lower risk of ovarian damage; however, clinical and morphological studies reveal that, although infrequent, they are not totally resistant to cytotoxic ovarian damage. Following treatment of Hodgkin’s disease with the alkylating combination chemotherapy regimens MVPP, MOPP (mustine, vincristine, procarbizine, prednisolone), or ChlVPP (chrorombucil, vinblastine, procarbazine, prednisolone), 15–62% of survivors develop amenorrhoea. In those over 35 years, amenorrhoea is almost invariable. In many the onset is abrupt, whilst in others there is progression to oligomenorrhoea with later development of a premature menopause. In contrast, use of ABVD (adriamycin, bleomycin, vinblastine, decarbazine) is much less gonadotoxic. In treatment of acute leukaemias with standard regimens, persistent ovarian failure is reported in less than 20% of survivors.

Strategies aimed at prevention of gonadal damage have led to the use of chemotherapeutic regimens, such as ABVD for the treatment of Hodgkin’s disease, that have equivalent cure rates but significantly less impact on gonadal function. There remains some risk to gonadal function, however, with almost all cancer therapies and discussions as to strategies for preservation of fertility needs to be undertaken as early as possible prior to commencement of cancer therapy. At present, only two options for fertility preservation are widely accepted and available: sperm banking for men and embryo cryopreservation for women. All other techniques remain in the realms of research (Table 11.1.3.4).

Males at risk of azoospermia due to their impending treatment schedule can have sperm frozen for future use. This procedure is relatively simple and part of standard practice, but of no value in prepubertal males. Sperm storage is most effective where the sperm concentration, motility, and morphology are not affected by the primary disease process. A significant proportion of men with lymphoma, leukaemia, and testicular tumours, however, are oligospermic or have impaired semen quality at presentation. For prepubertal boys there has been interest in harvesting spermatogonal stem cells which can then be frozen and stored for future use. Reimplantion of this tissue in to the testis after attainment of remission from cancer and completion of puberty could result in restoration of spermatogenesis. In vitro maturation of spermatogonial stem cells is a further investigational methodology that has been examined in animals. Postchemotherapy an increase in genetic abnormalities are observed in the spermatozoa. Concerns over the potential transmissibility of genetic anomalies, however, have not materialized. Suppression of the gonadal axis with GnRH analogues prior to cancer therapy has shown gonadal protection in animal models; however, there is no convincing evidence to date for benefit in either sex in the human.

A large number of cytotoxic agents have been implicated as teratogenic to the fetus, and it is therefore important that during cancer therapy women use appropriate contraception until remission is achieved. In women who retain normal ovulatory cycles after having received cytotoxic chemotherapy and who spontaneously conceive, no evidence of an increase in birth defects has been detected. Recovery of ovarian function in amenorrhoeic women and the possibility of a premature menopause in women retaining a normal cycle is difficult to predict accurately following an insult to the gonads received during multimodality cancer therapy. The use of transvaginal ultrasound to accurately quantitate ovarian volume and antral follicle count, along with measurement of inhibin-B and anti-Mullerian hormone, have been proposed as guides of future reproductive potential following cancer therapy. Both inhibin-B and antimullerian hormone are secreted by granulosa cells and thus concentrations decline with depletion of follicles. Further work is required to optimize these predictive models.

Preservation of fertility in women who are to undergo intense treatment likely to result in infertility is a significant growth area. Treatment of Hodgkin’s disease frequently includes local irradiation of involved lymph nodes, including those along the iliac vessels. The ovaries lie adjacent to the iliac vessels and will receive a dose of approximately 35 Gy, inevitably resulting in premature ovarian failure. Oophoropexy to remove the ovaries from the irradiation field, combined with shielding, can reduced the dose of irradiation received by the ovaries to less than 6 Gy, thereby reducing the incidence of amenorrhoea by around 50% (29). The exact reduction in risk of amenorrhoea as a consequence of oophoropexy is controversial and needs to be assessed in the context of disease extent, patient age, and surgical expertise. Both oocytes and embryos can be frozen. Embryo storage requires the patient to be in a stable relationship and undergo controlled stimulation of the ovary for several weeks, along with regular ultrasonograph monitoring and aspiration of follicles. This techniques is time-consuming when there is a pressing need to start treatment, is invasive, does not permit natural conception, and is not applicable to prepubertal girls. Pregnancy rates approximate to 15–30% per cycle with thawed embryos. Oocyte cryopreservation can be considered for patients without a partner, requires stimulation of the ovaries for around 2 weeks before retrieval of the oocytes, but is associated with a success rate of less than 5% for thawed oocytes and thus must be regarded as experimental. Recently, interest has been directed towards cryopreservation of ovarian cortical strips rich in primordial follicles which are then later thawed and grafted back in to the patient at the original site (orthotopic) or elsewhere (heterotopic). This technique is available to both prepubertal and mature women. Several large centres are now storing ovarian strips; however, to date there have been only two live births following orthotopic regrafting in women treated for lymphoma. Concerns remain as to whether cancer cells may be transferred back to the recipient. Only time will tell if this technique will improve the fertility prospects of women who undergo multimodality cancer therapy.

Men with overt hypogonadism should have testosterone replacement instituted to improve body composition, prevent osteoporosis, and maintain sexual function and wellbeing. In pubertal boys, who fail to progress through puberty due to overt testosterone deficiency, testosterone replacement will need to be titrated to bring the individual through puberty and maintain body composition and wellbeing thereafter. In men with compensated hypogonadism, sexual function has been found to be reduced along with slight reduction in bone mass and subtle body composition changes. Testosterone replacement in these individuals has not resulted in a significant improvement in bone mass, body composition, serum lipids, or quality of life, with the exception of a reduction in physical fatigue and low-density lipoprotein cholesterol.

In women under the age of 50 years who have developed gonadal failure the impact is twofold, on fertility and sex steroid production. Sex steroid replacement is recommended to alleviate symptoms of hot flushes, mood changes, and vaginal dryness, as well as to prevent loss of bone mass. The impact of sex steroid replacement on cardiovascular events remains controversial in patients below the age of 50 years in light of recent data showing an increase in vascular events in postmenopausal women treated with hormone replacement therapy. Reassuringly, after stratification of the WHI study data by age, the relative risk of cardiovascular disease was not increased in those aged 50–55 years.

Thyroid pathology following multimodality cancer therapy may relate to abnormalities of thyrotropin secretion as discussed, or relate to a direct effect on the thyroid gland itself. Primary thyroid anomalies following cancer therapy are very common, and may present as autoimmune thyroid disease (hypothyroidism, Graves’ disease, Graves’ ophthalmopathy), thyroiditis, or nodules (both benign and malignant).

Large-scale epidemiological studies have unequivocally confirmed a causal relationship between external beam irradiation of the neck and the development of thyroid cancer (Table 11.1.3.3) (30). The thyroid is most commonly irradiated in treatment of lymphomas, head and neck tumours, total body irradiation, and during spinal irradiation in the treatment of some brain tumours and haematological malignancies.

The most robust data concerning the development of thyroid nodules is derived from individuals who received mantel irradiation in the treatment of Hodgkin’s disease. The increased incidence of thyroid carcinogenesis appears 5–10 years after irradiation and remains elevated for at least several decades. The actuarial risk of thyroid carcinoma in survivors of Hodgkin’s disease has been calculated as 1.7%, equivalent to a relative risk of 15.6 (30). Following stem cell transplantation the SIR for development of thyroid carcinoma has been estimated as 3.26. The risk of developing thyroid carcinoma following neck irradiation is greater in children compared with adults, and young children are more vulnerable than older children. The odds ratio for development of thyroid carcinoma following 10–20 Gy irradiation to the thyroid in children diagnosed before 10 years of age at their first cancer compared with those diagnosed after 10 years at their first cancer has been estimated to be 16.3 and 2.9, respectively. The risk of developing thyroid cancer has been assumed to be linearly associated with dose. More recent data confirms this to be true for radiation doses up to 20–29 Gy, however, at doses greater than 30 Gy a fall in the dose response is observed consistent with a cell-killing effect of radiation at high doses (Fig. 11.1.3.4). Women are at greater risk of developing thyroid cancer than men at all doses of radiation. No definite association of chemotherapy with an increased risk of thyroid cancer has been shown, and neither does chemotherapy modify the carcinogenic effect of radiotherapy.

Histologically, the majority of radiation-induced thyroid carcinomas are well-differentiated papillary carcinomas (∼80%), with follicular carcinomas accounting for almost all remaining neoplasms; similar to the distribution observed in nonirradiated populations. It has been suggested, however, that the prevalence of multicentric disease, local invasion, and distant metastasis is greater in the irradiated thyroid. Most thyroid cancers are in the main eminently curable, but in an individual previously treated for cancer, the diagnosis can be a substantial psychological and physical burden.

In addition to thyroid carcinomas, thyroid irradiation is also associated with an increased incidence of benign thyroid nodules including focal hyperplasia, adenomas, colloid nodules, lymphocytic thyroiditis, and fibrosis (Table 11.1.3.3). Studies suggest palpable thyroid abnormalities to be present in 20–30% of the irradiated population compared with 1–5% in the general population. Time since irradiation, female gender, and dose of irradiation to the thyroid are independently associated with a greater risk of developing benign thyroid nodules. Recurrence rates are high following surgical removal of radiation-induced benign thyroid nodules, though not dissimilar to the nonirradiated gland. In contrast to the nonirradiated gland, thyroxine therapy aimed at TSH suppression, following surgery, may reduce the rate of recurrence of benign radiation-induced nodules.

Long-term surveillance with yearly examination of the thyroid and neck in survivors of cancer who received irradiation to the neck is essential. Controversy remains as to the best method to accomplish this as ultrasound will detect neoplasms earlier than palpation, including many small benign lesions of no clinical significance. Given the increased risk of thyroid carcinoma in these individuals, there should be a low threshold for performing fine needle aspiration cytology and diagnostic lobectomies where there is any doubt as to the diagnosis of thyroid nodules.

In addition to thyroid nodules, patients who receive irradiation to the thyroid have a greatly increased risk of thyroid dysfunction (Table 11.1.3.3). Compensated or frank hypothyroidism is reported in 20–30% of patients following fractionated TBI (9.0–15.0 Gy) in preparation for BMT (9), with higher prevalences reported after single-fraction TBI. Patients treated for lymphoma or head and neck cancers can receive a dose of 30–50 Gy in multiple fractions to the thyroid over several weeks. The cumulative probability of developing hypothyroidism in these individuals is 30–50% (30). Around 50% of cases develop within the first 5 years with the incidence declining thereafter. Thyroid dysfunction is also frequent after exposure of the thyroid to ionizing radiation during craniospinal radiotherapy, and also in children treated for brain tumours where the thyroid does not lie directly within the radiation field (31).

Younger age at irradiation, female gender, previous neck surgery, time since irradiation, and radiation dose all increase the prevalence of hypothyroidism. In cases where the thyroid function tests are only mildly abnormal, recovery may be seen even after many years. Treatment of both compensated and frank hypothyroidism with thyroxine is, however, warranted as animal studies have shown TSH to promote tumourigenesis in the irradiated thyroid. Intuitively, when replacing thyroid hormones in these individuals the goal should be to place the TSH within the lower reaches of the normative range.

The prevalence of Graves’ disease and Graves’ ophthalmopathy is increased in patients who receive irradiation to the thyroid, particularly in those who receive a dose above 30 Gy. The relative risk of Graves’ disease is around eightfold higher than the normal population (30). A transient thyroiditis postradiation exposure is described. The high incidence of developing thyroid dysfunction after exposure to ionizing radiation warrants screening of this population on an annual basis, or earlier should the individual become symptomatic.

There is no conclusive evidence that any cytotoxic chemotherapy can alter thyroid function. The prevalence of thyroid dysfunction in children who have received both craniospinal irradiation and chemotherapy is, however, reportedly greater compared with those who received only craniospinal irradiation. Both interferon and interleukin-2 increase the risk of developing autoimmune hypothyroidism. This condition is usually transient though a small proportion of patients may remain hypothyroid in the long term. Hyperthyroidism in the form of Graves’ disease and a transient thyroiditis has also been described. A number of chemotherapeutic agents affect thyroid function by modulating thyroid hormone binding; 5-flurouracil increases total T₃ and T₄, but patients remain euthyroid with normal TSH and free T₄ levels; asparaginase decreases production of hepatic thyroid binding globulin and total T₄ levels through its widespread effects on protein and DNA synthesis.

Through retrospective studies showing a greater proportion of patients who developed hyperparathyroidism received neck irradiation compared with normocalcaemic controls, neck irradiation has been implicated in the aetiology of hyperparathyroidism (Table 11.1.3.3). The absolute increase in incidence is difficult to quantitate as a consequence of the long latency period between irradiation and the development of hyperparathyroidism (25–47 years). Clinically, patients warrant annual screening of calcium levels, which must be undertaken lifelong in view of the long latency period between the insult and gland dysfunction.

Endocrine late effects in survivors of cancer can lead to abnormalities of all endocrine axes. Both chemotherapy and radiotherapy play a role in the aetiology of the adverse sequelae observed. All individuals involved in the care of cancer survivors need to be vigilant to the development of late effects, which can occur even decades after completing cytotoxic therapy. Where patients are at significant risk of developing a treatment-associated endocrinopathy, patients should be referred to an endocrinologist and undergo regular screening, the frequency of which will be dictated by the therapies received.