Late-effects among survivors of leukaemia and lymphoma during childhood and adolescence

Authors


Leslie L. Robison, PhD, Division of Pediatric Epidemiology and Clinical Research, University of Minnesota Cancer Center, Mayo Mail Code 422, Minneapolis, MN 55455, USA. E-mail: robison@epi.umn.edu

With improvements in survival rates for children and adolescents diagnosed with leukaemia or lymphoma, issues relating to late treatment-related sequelae have increased in importance. Reviewed are selected late adverse outcomes, including second malignancies, cardiotoxicity, endocrinological effects, impact on neurocognitive/neuropsychological status, late mortality and implications of stem cell transplantation. While there has been considerable research on the late outcomes among survivors of acute lymphoblastic leukaemia and Hodgkin's disease, future research must be directed towards identification of risks associated with more recent treatment regimens, as well as the very late-occurring outcomes resulting from treatment protocols utilized three or more decades ago. As treatment- and patient-related factors impact the subsequent risk of late-occurring adverse outcomes, clear delineation of those leukaemia and lymphoma survivors who are at high risk of specific adverse outcomes is essential for the rational design of follow-up guidelines, prevention and intervention strategies.

Introduction

Over the last three decades, survival rates for many of the childhood cancers have increased at a remarkable pace. Improvements in therapy have increased the probability of 5-year survival from less than 30% in 1960 to over 70% in 1990. Population-based data from the United States Surveillance, Epidemiology and End Results (SEER) program demonstrates that much of the increase in survival was achieved in the 1970s and early 1980s, with the subsequent rate of increase in the overall proportion achieving 5-year survival being negligible for children diagnosed since 1985 (Ries et al, 1999). Five-year survival is currently achievable in the majority of children and adolescents diagnosed with leukaemia (80% in acute lymphoid but only 40% in acute myeloid) or lymphoma [91% in Hodgkin's disease and 72% in non-Hodgkin's lymphoma (NHL)]. With this success comes the need to consider the long-term morbidity and mortality associated with the treatments responsible for that increased survival. Because of the young age of these cancer survivors and, thus, the potential longevity, the delayed consequences of therapy may have a greater impact on their lives and on society at large than the acute complications of the cytotoxic therapies they have already experienced.

The subject of late effects among children treated for cancer has been the topic of numerous reviews (Boulad et al, 1998; Marina, 1997; Hudson, 1999; Bhatia et al, 2002a; Dreyer et al, 2002). To varying degrees, it has been shown that disease- or treatment-specific subgroups of long-term survivors are at risk of developing adverse outcomes including early death, second neoplasms, organ dysfunction (e.g. cardiac, pulmonary, gonadal), reduced growth and development, decreased fertility, impaired intellectual function, difficulties obtaining employment and insurance, and overall reduced quality of life. This review summarizes selected aspects of the spectrum of outcomes relating to the late-effects of therapy among children and adolescents treated for leukaemia and lymphoma.

Second malignancies

A recent publication from the Childhood Cancer Survivor Study (Neglia et al, 2001) reported a cohort of more than 14 000 5-year survivors diagnosed and treated for their original cancer during 1970–1986, for whom the relative risk (RR) of developing a second malignancy was significantly increased for survivors of NHL (RR = 3·2), leukaemia (RR = 5·7), and Hodgkin's disease (RR = 9·7). While the observed relative risks are impressively elevated, the corresponding cumulative incidence rates and absolute excess risks provide additional measures that reflect the relatively low proportion of survivors effected (Table I).

Table I.  Occurrence of second malignancies among 5-year survivors of childhood leukaemia and lymphoma (adapted from Neglia et al, 2001).
DiagnosisObserved/expected ratio (95% confidence interval)Cumulative incidence at 20 yearsAbsolute excess risk*
  • *

    Risk is expressed at number of malignancies per 1000 person-years of follow-up.

Leukaemia5·66 (4·37–7·22)2·1%1·20
Hodgkin's disease9·70 (8·05–11·59)7·6%5·13
NHL3·21 (1·76–5·39)1·9%0·89

Evaluation of very large cohorts of patients diagnosed with childhood acute lymphoblastic leukaemia (ALL) and entered on therapeutic trials of the Children's Oncology Group have shown that the cumulative incidence of second and subsequent malignancies is only approximately 2% at 15 years from diagnosis and treatment of ALL (Neglia et al, 1991; Bhatia et al, 2002b). This low cumulative incidence has also been seen in smaller cohorts of ALL patients.

Central nervous system (CNS) tumours, the most common second malignancy observed among survivors of childhood ALL, are predominantly associated with exposure to cranial radiation (Neglia et al, 2001). Histologically, the most frequently observed forms of secondary CNS tumours are high-grade gliomas, including glioblastomas and malignant astrocytomas; although less frequently observed, peripheral neuroectodermal tumours, epdendymonomas and meningiomas do occur (Neglia et al, 1991; Bhatia et al, 2002b). Other potential risk factors for secondary CNS tumours include younger age at radiation, inherited genetic predisposition to cancer and genetic profile with regard to genetic polymorphisms involved in metabolizing enzymes. An example of a polymorphism that has been found to be predictive of the risk of secondary CNS tumours in childhood ALL is thiopurine S-methyl-transferase (TPMT) (Relling et al, 1999).

Other commonly reported second cancers within the population of ALL survivors include thyroid cancer, lymphoma and acute myeloid leukaemia (AML). Secondary thyroid malignancies, typically papillary carcinoma, are generally associated with radiation exposure to the thyroid gland as part of CNS radiation, either prophylactically or for treatment of CNS leukaemia. Thyroid malignancy has been reported to represent between 6% and 17% of secondary cancers among large cohorts of ALL survivors (Neglia et al, 1991, 2001; Bhatia et al, 2002b), and typically develop 10 or more years from treatment. As is true of de novo thyroid malignancy, the long-term outcome for survivors diagnosed with a secondary thyroid malignancy is excellent. The risk of secondary AML following therapy for ALL is generally low, except among those patients treated with epipodophyllotoxin therapy, where a cumulative risk of 3·8% at 6 years has been reported (Pui et al, 1991). Secondary AML associated with Topoisomerase-II inhibitors is characterized typically by a shorter latency period (3–5 years from therapeutic exposure) than that seen for secondary AML following alkylating agents, by the presence of 11q23 rearrangements with mutations in the myeloid/lymphoid or mixed-lineage leukaemia gene (MLL), and is more dependent on the schedule of drug administration than total cumulative dose (Pui et al, 1995; Smith et al, 1999). Epipodophyllotoxin-associated secondary AML has also been documented among paediatric NHL patients (Hawkins et al, 1992).

Because the survival rate for paediatric Hodgkin's patients has been high for more than three decades, there exists a considerable amount of outcome-based research focusing on the occurrence of second malignancies. Survivors of paediatric and adolescent Hodgkin's disease clearly represent one of the subgroups of cancer survivors who are at very high risk of secondary cancer. This is particularly true for patients who received earlier regimens with predominantly radiation-based therapies where an approximate 10-fold increased risk has been reported (Neglia et al, 2001). A number of studies, including cohorts ranging from 499 to 5925 Hodgkin's patients, have reported cumulative risks of second malignancies ranging from 7·6% at 20 years to 18·0% at 30 years (Tucker, 1993; Beatty et al, 1995; Bhatia et al, 1996; Jenkin et al, 1996; Sankila et al, 1996; Wolden et al, 1998; Green et al, 2000; Metayer et al, 2000; Neglia et al, 2001). Early studies of Hodgkin's survivors identified the increased risk of secondary leukaemias among patients treated with MOPP (mechlorethamine, Oncovin, procarbazine, prednisone)-based therapy, which included cyclophosphamide (Tucker et al, 1988; Kaldor et al, 1990). These alkylating-agent-associated secondary leukaemias are characterized as having a relatively short latency period (mean of approximately 7 years), the presence of monosomy 5 or monosomy 7, and are often preceded by a phase of myleodysplasia (Felix, 1998). The risk of therapy-related leukaemia usually does not extend beyond the first 10–15 years after therapeutic exposure.

There is increasing interest in the investigation of associations between therapy-related myelodysplasia or acute leukaemia and polymorphisms in specific drug-metabolizing enzymes capable of metabolic activation or detoxification of anticancer drugs, such as NAD(P)H:quinone oxidoreductase (NQO1), glutathione S-transferase (GST)-M1 and -T1, and CYP3A4 (Wrighton & Stevens, 1992; Hayes & Pulford, 1995; Raunio et al, 1995; Smith et al, 1995; Felix et al, 1998; Naoe et al, 2000; Blanco et al, 2002). There is evidence to indicate that the NQO1 polymorphism is significantly associated with the genetic risk of therapy-related acute leukaemia and myelodysplasia. In addition, individuals with the CYP3A4-W genotype may be at increased risk of treatment-related leukaemia, by increasing the production of reactive intermediates that might damage DNA.

Breast cancer is the most commonly reported second malignancy among paediatric Hodgkin's survivors treated with mantle field radiation, and the risk remains markedly elevated for many decades following exposure (Bhatia et al, 1996; Sankila et al, 1996; Wolden et al, 1998; Metayer et al, 2000; Neglia et al, 2001). A recent update of the Late Effects Study Group cohort found female survivors to have a 55-fold increased risk of breast cancer compared with the general population, and the cumulative incidence of developing a secondary breast cancer approached 20% at 45 years of age (Bhatia et al, 2001a). Moreover, 40% of identified patients were found to have developed contralateral disease. Secondary thyroid cancer, the second most common solid tumour reported among survivors of Hodgkin's disease, is strongly associated with radiation therapy, occurs more frequently in women and is associated with an approximate 36-fold increased risk over the general population. As the cohort of paediatric Hodgkin's survivors continues to age, it is likely that an increasing number of other forms of malignancy, associated with an excess risk, will emerge. Extended follow-up of early cohorts have already reported excess risks for lung, gastric and colorectal cancers (van Leeuwen et al, 1995; Kreiker & Katten, 1996; Bhatia et al, 2001a; Deutsch et al, 2002).

Cardiotoxicity

Chronic cardiotoxicity usually manifests itself as cardiomyopathy, pericarditis and congestive heart failure. The anthracyclines, doxorubicin and daunorubicin, are well known causes of cardiomyopathy, which can occur many years after completion of therapy. The incidence of anthracycline-induced cardiomyopathy, which is dose dependent, may exceed 30% among patients receiving cumulative doses in excess of 600 mg/m2. A cumulative dose of anthracyclines greater than 300 mg/m2 has been associated with an 11-fold increased risk of clinical heart failure, compared with a cumulative dose of less than 300 mg/m2, with the estimated risk of clinical heart failure increasing with time from exposure and approaching 5% after 15 years (Kremer et al, 2001). Sorensen et al (1997) studied the prevalence of late cardiotoxicity in 120 long-term survivors of ALL treated with lower anthracycline doses, and reported a reduced incidence and severity of cardiac abnormalities with the lower anthracycline dose protocols (90–270 mg/m2) studied, compared with previous reports in which subjects had received moderate anthracycline doses (300–550 mg/m2). Twenty-three per cent of the patients were found to have cardiac abnormalities. Twenty-one per cent had increased end-systolic stress, while only 2% had reduced contractility. The cumulative anthracycline dose within the 90–270 mg/m2 range did not relate to cardiac abnormalities. The authors concluded that there may be no safe anthracycline dose to completely avoid late cardiotoxicity. A recent review of 30 published studies found that the frequency of clinically detected anthracycline cardiac heart failure ranged from 0% to 16% (Kremer et al, 2002a). In an analysis of reported studies, the type of anthracycline (i.e. doxorubicin) and the maximum dose given in a 1-week period (i.e. > 45 mg/m2) was found to explain a large portion of the variation in the reported frequency of anthracycline-induced cardiac heart failure.

The incidence of subclinical anthracycline myocardial damage has been the subject of considerable interest. Steinherz et al (1991) found 23% of 201 patients, who had received a median cumulative dose of doxorubicin of 450 mg/m2, had echocardiographic abnormalities a median of 7 years after therapy. In a group of survivors of childhood cancer who received a median doxorubicin dose of 334 mg/m2, it was found that progressive elevation of afterload or depression of left ventricular contractility was present in approximately 75% of patients (Lipshultz et al, 1991). A recent review of the literature on subclinical cardiotoxicity among children treated with an anthracycline found that the reported frequency of subclinical cardiotoxicity varied considerably across the 25 studies reviewed (frequency ranging from 0% to 57%) (Kremer et al, 2002b). Because of marked differences in the definition of outcomes for subclinical cardiotoxicity and the heterogeneity of the patient populations investigated, it is difficult to accurately evaluate the potential long-term outcomes within anthracycline-exposed patient populations or the potential impact of the subclinical findings.

Among anthracycline-exposed patients, the risk of cardiotoxicity can be increased by mediastinal radiation (Fajardo et al, 1968), uncontrolled hypertension (Minow et al, 1975; Prout et al, 1977), underlying cardiac abnormalities (Von Hoff et al, 1979), exposure to non-anthracycline chemotherapeutic agents (especially cyclophosphamide, dactinomycin, mitomycin C, dacarbazine, vincristine, bleomycin and methotrexate) (Kushner et al, 1975; Minow et al, 1975; Smith et al, 1977; Von Hoff et al, 1982), female sex (Lipshultz et al, 1995), younger age (Pratt et al, 1978), and electrolyte imbalances such as hypokalaemia and hypomagnesaemia (Pai & Nahata, 2000). Previous reports have suggested that doxorubicin-induced cardiotoxicity can be prevented by continuous infusion of the drug (Legha et al, 1982). However, Lipshultz et al (2002) compared cardiac outcomes in children receiving either bolus or continuous infusion of doxorubicin, and reported that continuous doxorubicin infusion over 48 h for childhood leukaemia did not offer a cardioprotective advantage over bolus infusion. Both regimens were associated with progressive subclinical cardiotoxicity, thus suggesting that there is no benefit from continuous infusion of anthracyclines.

Chronic cardiac toxicity associated with radiation alone most commonly involves pericardial effusions or constrictive pericarditis, sometimes in association with pancarditis. Although a dose of 40 Gy of total heart irradiation appears to be the usual threshold, pericarditis has been reported after as little as 15 Gy, even in the absence of radiomimetic chemotherapy (Marks et al, 1973; Martin et al, 1975). Symptomatic pericarditis, which usually develops 10–30 years after irradiation, is found in 2–10% of patients (Ruckdeschel et al, 1975). Subclinical pericardial and myocardial damage, as well as valvular thickening, may be common in this population (Perrault et al, 1985; Kadota et al, 1988). Coronary artery disease has been reported after radiation to the mediastinum, although mortality rates have not been significantly higher in patients who receive mediastinal radiation than in the general population (Hancock et al, 1993).

Given the known acute and long-term cardiac complications of therapy, prevention of cardiotoxicity is a focus of active investigation. Several attempts have been made to minimize the cardiotoxicity of anthracyclines, such as the use of liposomal-formulated anthracyclines, less cardiotoxic analogues and the additional administration of cardioprotective agents. The advantages of these approaches are still controversial, but there are ongoing clinical trials to evaluate the long-term effects.

Thus, certain analogues of doxorubicin and daunorubicin, with decreased cardiotoxicity but equivalent antitumour activity, are being explored. Agents such as dexrazoxane, which are able to remove iron from anthracyclines, have been investigated as cardioprotectants. Clinical trials of dexrazoxane have been conducted in children, with encouraging evidence of short-term cardioprotection (Wexler, 1998), however, the long-term avoidance of cardiotoxicity with the use of this agent has yet to be sufficiently determined.

Endocrinological effects

Radiation exposure to the head and neck is a known risk factor for subsequent abnormalities of the thyroid. Among survivors of Hodgkin's disease and, to a lesser extent, leukaemia survivors, abnormalities of the thyroid gland, including hypothyroidism, hyperthyroidism and thyroid neoplasms, have been reported to occur at rates significantly higher than found in the general population (Shalet et al, 1977; Robison et al, 1985; Hancock et al, 1991; Sklar et al, 2000a). Hypothyroidism is the most common non-malignant late effect involving the thyroid gland. Following radiation doses above 15 Gy, laboratory evidence of primary hypothyroidism is evident in 40–90% of patients with Hodgkin's disease, NHL or head/neck malignancies (Glatstein et al, 1971; Rosenthal & Goldfine, 1976; Hancock et al, 1991). In a recent analysis of 1791 5-year survivors of paediatric Hodgkin's disease (median age at follow-up of 30 years), Sklar et al (2000a) reported the occurrence of at least one thyroid abnormality in 34% of subjects. The risk of hypothyroidism was increased 17-fold compared with sibling control subjects, with increasing dose of radiation, older age at diagnosis of Hodgkin's disease and female sex as significant independent predictors of an increased risk. The actuarial risk of hypothyroidism for subjects treated with 45 Gy or more was 50% at 20 years following diagnosis of their Hodgkin's disease. Hyperthyroidism was reported to occur in only 5%.

Poor linear growth and short adult stature are common complications after successful treatment of childhood cancers (Sklar, 1997). The adverse effect of CNS radiation on adult final height among childhood leukaemia patients has been well documented, with final heights below the fifth percentile occurring in 10–15% of survivors (Berry et al, 1983; Robison et al, 1985; Sklar et al, 1993; Papadakis et al, 1996). The effects of cranial radiation appear to be related to age and sex, with children younger than 5 years at the time of therapy and female patients being more susceptible. The precise mechanisms by which cranial radiation induces short stature are not clear. Disturbances in growth hormone production have not been found to correlate well with observed growth patterns in these patients (Shalet et al, 1979; Blatt et al, 1984). The phenomenon of early onset of puberty in girls receiving cranial radiation may also play some role in the reduction of final height (Leiper et al, 1988; Didcock et al, 1995). In childhood leukaemia survivors not treated with cranial radiation, there are conflicting results regarding the impact of chemotherapy on final height (Katz et al, 1993; Sklar et al, 1993).

An increased prevalence of obesity has been reported among survivors of childhood ALL (Odame et al, 1994; Van Dongen-Melman et al, 1995; Sklar et al, 2000b). Craig et al (1999) investigated the relationship between cranial irradiation received during treatment for childhood leukaemia and obesity. Two hundred and thirteen (86 boys and 127 girls) irradiated patients and 85 (37 boys and 48 girls) non-irradiated patients were enrolled. For cranially irradiated patients, an increase in the body mass index (BMI) z score at the final height was associated with female sex and lower radiation dose, but not with age at diagnosis. Severe obesity, defined as a BMI z score > 3 at final height, was only present in girls who received 18–20 Gy irradiation at a prevalence of 8%. Both male and female non-irradiated patients had raised BMI z scores at latest follow-up, and there was no association with age at diagnosis. The authors concluded that these data demonstrated a sexually dimorphic and dose-dependent effect of cranial irradiation on BMI. In a recent analysis from the Childhood Cancer Survivor Study, Oeffinger et al (2003) compared the distribution of BMI of 1765 adult survivors of childhood ALL with that of 2565 adult siblings of childhood cancer survivors. Survivors were significantly more likely to be overweight (BMI 25–30) or obese (BMI ≥ 30). Risk factors for obesity were cranial radiation, female sex and age 0–4 years at diagnosis of leukaemia. Girls diagnosed under the age of 4 years who received a cranial radiation dose > 20 Gy were found to have a 3·8-fold increased risk of obesity.

Treatment-related gonadal dysfunction has been well documented to occur in both men and women following childhood malignancies, including leukaemia and lymphoma (Dreyer et al, 2002; Thomson et al, 2002). However, survivors of leukaemia and T-cell non-Hodgkin lymphoma treated with modern conventional therapy are at a relatively low risk of infertility and delayed or impaired puberty. There exists a reasonable body of research on topics relating to the long-term gonadal effects of radiation and chemotherapy, which provide a basis for counselling patients and parents of the anticipated outcomes on pubertal development and fertility.

Radiation effects on the ovary are both age and dose dependent. Amenorrhoea develops in approximately 68% of prepubescent girls treated for Hodgkin's disease with ovarian doses of 12–15 Gy, while 100% of women over 40 years of age will sustain irreversible ovarian failure following doses of 4–7 Gy (Lushbaugh & Casarett, 1976; Stillman et al, 1981). Spinal radiation for the treatment of childhood leukaemia appears to result in clinically significant ovarian damage in some survivors (Hamre et al, 1987). It has been estimated that there is a 50% depletion in oocytes following exposure of the ovaries to 2 Gy (Wallace et al, 2003).

The effects of radiation on testicular function, including germ cell number and Leydig cell function, have been investigated. Reduced sperm production has been observed following testicular doses of 1–6 Gy and follows a dose-dependent pattern (Rowley et al, 1974). Azoospermia has been reported among Hodgkin's patients with calculated testicular radiation exposures ranging from 1 to 3 Gy (Speiser et al, 1973). Testicular doses between 4 and 6 Gy have been associated with prolonged azoospermia and decreased testicular volume (Shamberger et al, 1981). The limited data on the long-term outcome of very young males treated for Hodgkin's disease suggest germ cell effects similar to those seen in older Hodgkin's patients (Green et al, 1981). Leydig cells, while also impacted by radiation in a dose-dependent fashion, require higher exposure levels to sustain damage than seen for the germ cells (Sklar, 1999). A testicular dose of 24 Gy among prepubertal boys has been reported to be associated with delayed pubertal development, and abnormal testosterone and gonadotropin levels (Shalet et al, 1985; Leiper et al, 1986; Sklar et al, 1990).

Ovarian and testicular damage can result from chemotherapeutic agents, with alkylating agents showing the strongest association. Effects of chemotherapy on gonadal function are typically sex, age and dose dependent. The ovaries tend to be less sensitive to the effects of alkylating agent exposure, compared with the testes. Ovarian dysfunction has been well documented in Hodgkin's disease patients treated with alkylating agents, either singly or in combination (i.e. MOPP regimen consisting of meclorethamine, vincristine, procarbazine and prednisolone or COPP regimen consisting cyclophosphamide, vincristine, procarbazine and prednisolone) (Chapman et al, 1979; Whitehead et al, 1982; Ortin et al, 1990). Data from patients treated with high-dose alkylating agents suggest that these subjects may be at an increased risk of an early menopause as they reach the third decade of life (Byrne et al, 1992).

Mackie et al (1996) assessed gonadal function in 101 postpubertal subjects after chemotherapy for childhood Hodgkin's disease. All had received ChlVPP (chlorambucil, vinblastine, procarbazine and prednisolone) chemotherapy alone, with no radiotherapy below the diaphragm. Gonadotropin levels were available in 46 (79·3%) male and 32 (74·4%) female subjects. The mean age at diagnosis in the male cohort was 12·2 years (range 8·2–15·3 years) and in the female cohort was 13·0 years (range 9·0–15·2 years). The males and the females were studied at a median of 6 years (range 2·5–11·1 years) and 4·3 years (range 1·9–11·5 years) from diagnosis respectively. Forty-one (89·1%) male subjects had elevated follicle-stimulating hormone (FSH) levels, confirming severe germinal epithelial damage. Germinal epithelial damage was seen in subjects up to 10 years out of therapy. Subtle Leydig cell dysfunction was identified in 24·4% with raised luteinizing hormone (LH) levels. All subjects, however, progressed spontaneously through puberty. Seventeen (53%) women had raised gonadotropin levels, with variable estradiol levels. Of these, 10 subjects presented with symptomatic ovarian failure and six received hormone replacement therapy (HRT). Nine women had 11 successful pregnancies; two of these patients had previously had symptoms of ovarian failure, with one requiring HRT. This study demonstrated that a much higher prevalence of ovarian failure was observed than had previously been considered in the prepubertal and pubertal female patients following combination chemotherapy. In another report, Viviani et al (1985) compared gonadal toxicity following two equally effective and non-cross-resistant regimens [MOPP and ABVD (adriamycin, bleomycin, vincristine, dacarbazine)] in 53 men with Hodgkin's disease. The median age was 29 years (range 16–45 years). MOPP produced azoospermia in 28/29 patients (97%) while ABVD induced oligoazoospermia in 13/24 patients (54%). Follicle-stimulating hormone levels were consistently and significantly increased after MOPP, while their median value remained within the normal range after ABVD. A sperm count was repeated in 34 patients. Recovery of spermatogenesis occurred in 3/21 patients treated with MOPP and in all 13 patients given ABVD. These findings confirm that the two alkylating agents, mechlorethamine and procarbazine, included in the MOPP regimen cause sterility in most patients, while the drugs included in ABVD are not associated with permanent gonadal dysfunction.

Excellent survival after Hodgkin's disease has allowed the focus to shift from more effective therapy to prevention of long-term treatment-related sequelae. Treatment with radiation therapy alone is recommended only for older patients with localized disease who have achieved skeletal maturity, but requires surgical staging and places greater volumes of normal tissues at risk for late carcinogenesis. Treatment with chemotherapy alone avoids the long-term growth, organ dysfunction and solid tumour induction associated with high-dose, extended-field radiation. However, these protocols prescribe higher cumulative doses of alkylating agent chemotherapy, which may increase the risk of treatment complications from myelosuppression, gonadal injury and secondary leukaemia. Combined modality therapy regimens have resulted in excellent treatment outcomes and reduced the incidence of treatment sequelae by utilizing lower doses and smaller volumes of radiation therapy and fewer cycles of less toxic chemotherapy in clinically staged children. Risk-adapted therapies, using two to four cycles of multiagent chemotherapy, and lower radiation doses and volumes, have maintained excellent disease-free survival rates in clinically staged patients with localized favourable disease presentations. Novel approaches, including compacted dose-intensive multiagent chemotherapy, are currently under investigation with the objectives of improving outcome and reducing treatment sequelae in patients with advanced and unfavourable disease. (Hudson & Donaldson, 1999)

Neurocognitive and neuropsychological effects

Long-term survivors of leukaemia and lymphoma, treated during childhood and adolescence, may be at risk of neurocognitive and neuropsychological sequelae. Among survivors of childhood leukaemia, neurocognitive late effects represent one of the more intensively studied topics. Adverse outcomes are generally associated with whole brain radiation and/or therapy with high-dose systemic or intrathecal methotrexate or cytarabine (Meadows et al, 1981; Ochs et al, 1991;Jankovic et al, 1994; Christie et al, 1995; Waber et al, 1995; Hertzberg et al, 1997). High-risk characteristics, including higher dose of CNS radiation, younger age at treatment and female sex, have been well documented. Results from studies of neurocognitive outcomes are directly responsible for the marked reduction (particularly in younger children) in the use of cranial radiation, which is currently reserved for treatment of very high-risk subgroups or patients with CNS involvement.

A spectrum of clinical syndromes may occur, including radionecrosis, necrotizing leucoencephalopathy, mineralizing microangiopathy and dystropic calcification, cerebellar sclerosis, and spinal cord dysfunction (Price, 1983). Leucoencephalopathy has been primarily associated with methotrexate-induced injury of white matter. However, cranial radiation may play an additive role through the disruption of the blood–brain barrier, thus allowing greater exposure of the brain to systemic therapy.

While abnormalities have been detected by diagnostic imaging studies, the abnormalities observed have not been well demonstrated to correlate with clinical findings and neurocognitive status (Peylan-Ramu et al, 1978; Riccardi et al, 1985; Wilson et al, 1991; Kingma et al, 1993; Hasle et al, 1995). Chemotherapy- or radiation-induced destruction in normal white matter partially explains intellectual and academic achievement deficits (Mulhern et al, 1999). Evidence suggests that direct effects of chemotherapy and radiation on intracranial endothelial cells and brain white matter as well as immunological mechanisms could be involved in the pathogenesis of central nervous damage.

Neurocognitive deficits, as a general rule, usually become evident within several years following CNS radiation and tend to be progressive in nature. Leukaemia survivors treated at a younger age (e.g. less than 6 years of age) may experience significant declines in intelligence quotient (IQ) scores (Kramer & Moore, 1989; Packer et al, 1989). However, reductions in IQ scores are typically not global, but rather reflect specific areas of impairment, such as attention and other non-verbal cognitive processing skills (Goff et al, 1980; Peckham et al, 1988). Affected children may experience information-processing deficits, resulting in academic difficulties. These children are particularly prone to problems with receptive and expressive language, attention span, and visual and perceptual motor skills, most often manifested in academic difficulties in the areas of reading, language and mathematics. Accordingly, children treated with CNS radiation or systemic or intrathecal therapy with the potential to cause neurocognitive deficits should receive close monitoring of academic performance. Referral for neuropsychological evaluation with appropriate intervention strategies, such as modifications in curriculum, speech and language therapy or social skills training, implemented in a programme tailored for the individual needs and deficits of the survivor should be taken into consideration (Moore et al, 1994). Assessment of educational needs and subsequent educational attainment have found that survivors of childhood leukaemia are significantly more likely to require special educational assistance, but have a high likelihood of successfully completing high school (Haupt et al, 1994; Mitby et al, 2003). However, when compared with siblings, survivors of leukaemia and NHL are at greater risk of not completing high school. As would be anticipated from the results of neurocognitive studies, it has been shown that survivors, particularly those under 6 years of age at treatment, who received cranial radiation and/or intrathecal chemotherapy, were significantly more likely to require special education services and least likely to complete a formal education.

Beyond the neurocognitive deficits, survivors of leukaemia and lymphoma are at risk of adverse neuropsychological outcomes that may impact overall quality of life. Results from a recent analysis of 5736 long-term survivors of leukaemia and lymphoma demonstrated that while a relatively low proportion reported symptoms indicative of depression (4·6%) and somatic distress (10·8%), they were significantly more likely to report depression or somatic distress when compared with sibling control subjects (Zebrack et al, 2002). There is growing interest in the reported occurrence of fatigue and sleep disturbances among cancer survivors, particularly those with Hodgkin's disease (Loge et al, 1999; Knobel et al, 2001; Green et al, 2002), which may contribute to depression.

Haematopoietic cell transplantation

The use of blood and marrow transplantation is considered a standard therapy for a select subgroup of high-risk patients with leukaemia and lymphoma. Transplantation may be the treatment of choice for patients with disease characteristics known to have a very poor prognosis with conventional chemotherapy/radiation regimens, or those patients who experience recurrent disease following treatment with conventional regimens. Therefore, when evaluating late effects of therapy in a transplant population, consideration needs to be given not only to the transplant-related therapy but also the previous therapy the patient may have received. Moreover, complications observed following blood and marrow transplant often have a multifactorial aetiology, encompassing factors related not only to prior cancer therapy, but also factors such as intensity of the preparative regimen for haematopoietic cell transplantation, graft-versus-host disease (GVHD) and other post-transplant complications (Sullivan et al, 1992; Thuret et al, 1995; Leahey et al, 1999; Brennan & Shalet, 2002; Leiper, 2002a, 2002b).

Previous studies have shown that marrow transplant recipients have a four to sevenfold increased risk of cancer, particularly Epstein–Barr-virus-related post-transplant lymphoproliferative disease, myelodysplastic syndromes and a variety of solid non-haematopoietic tumours (Witherspoon et al, 1989; Curtis et al, 1997; Socie et al, 2000; Bhatia et al, 2001b; Baker et al, 2003). Organ dysfunction has been reported among survivors of blood and marrow transplants. Significant late toxicity involving both the airway and lung parenchyma, which include restrictive lung disease and chronic obstructive lung disease, as well as bronchiolitis obliterans, is observed in 15% to 40% of patients after transplant. Restrictive lung disease is related to both pretransplant chemotherapy, as well as the therapeutic exposures during conditioning (Rovelli et al, 1995). Bronchiolitis obliterans has been shown to have a strong correlation with chronic GVHD, has been reported in up to 14% of allogeneic transplant recipients and carries a mortality rate of 50% (Holland et al, 1988). Liver disease due to chronic GVHD, hepatitis C virus infection, nodular regenerative hyperplasia and chronic fungal infection has been reported (Snover et al, 1989; Stechschulte et al, 1990; Strasser et al, 1999). Renal dysfunction may be the result of direct nephrotoxicity from preparative radiation and/or chemotherapy, tumour lysis, intravascular depletion and nephrotoxic medications; although generally seen as an acute toxicity, there are reports of renal dysfunction among long-term survivors (Tarbell et al, 1988; Miralbell et al, 1996).

As many preparative regimens are intensive, and include total body irradiation (TBI) and alkylating agents, patients successfully treated with a haematopoietic stem cell transplant are often at increased risk of endocrine dysfunction. Following administration of high-dose cyclophosphamide for aplastic anaemia among patients receiving a transplant, ovarian function has remained normal in females treated both prior to as well as after the onset of puberty (Sanders et al, 1988; Sanders, 1991). At least 50% of prepubertal girls undergoing blood and marrow transplantation retain adequate ovarian function to enter puberty and menstruate regularly (Sarafoglou et al, 1997). However, clinically recognized pregnancies among survivors of total body irradiation are at increased risk of spontaneous abortion (Sanders et al, 1996). Males treated with TBI-based preparative regimens, both single dose and fractionated, generally retain their ability to produce testosterone regardless of their age at treatment (Sanders et al, 1986; Ogilvy-Stuart et al, 1992). On the other hand, germ cell dysfunction is present in essentially all males treated with TBI (Sanders et al, 1996). Young boys and adolescent males exposed to cyclophosphamide alone at 200 mg/kg retain normal Leydig cell function, with normal concentrations of luteinizing hormone and testosterone, and normal onset and progression through puberty. Evidence of germ cell damage has been reported following this therapy and may be more common in those males treated during or after puberty compared with those treated prior to the onset of puberty (Sklar et al, 2001; Brennan & Shalet, 2002). Overall, most prepubertal boys undergoing chemotherapy and hyperfractionated TBI can expect to enter and progress normally through puberty (Socie et al, 2000). As seen in patients treated by conventional chemotherapy, radiation-containing preparative regimens are more likely to be associated with hypothyroidism (Sanders et al, 1991). Growth hormone deficiency has been reported in approximately 50–60% of patients receiving TBI. This risk is compounded among those who received pretransplant cranial radiation (Huma et al, 1995; Clement-De Boers et al, 1996).

Survivors of haematopoietic stem cell transplants are also at risk of neuropsychological and neurocognitive late effects. Prospective, longitudinal evaluations of intellectual and adaptive functioning of children receiving a transplant have revealed declines in intellectual function, particularly among those under 6 years of age at transplant (Kramer et al, 1997; Simms et al, 1998; Phipps et al, 2000). In general, studies have reported good levels of function and well-being in long-term survivors of transplantation, with the exception of patients receiving TBI at a very young age, especially among those with a history of prior craniospinal irradiation. In addition, there are increasing reports of fatigue, lack of energy and sleep problems (Sutherland et al, 1997; Heinonen et al, 2001). However, most studies addressing quality of life in transplant recipients focus on adult populations, with very little data available on survivors of transplantation performed in childhood.

Late mortality

Late recurring disease, as well as sequelae of the treatment for leukaemia and lymphoma during childhood, can have a direct or indirect impact on overall mortality. Several very large studies of late mortality among 5-year survivors of childhood cancer have been conducted (Mertens et al, 2001; Möller et al, 2001) with remarkable agreement in their findings (Simone, 2001). Among the 5-year survivors in the US study (Mertens et al, 2001), the proportion who were alive 25 years from the diagnosis of leukaemia, Hodgkin's disease and NHL was 87%, 81% and 90% respectively. The standardized mortality ratios (SMR) (i.e. observed deaths to expected deaths) for death due to subsequent cancer, cardiac complications or pulmonary conditions were statistically significantly elevated among leukaemia patients (SMR from 3·8 for cardiac to 17·4 for second cancer), Hodgkin's disease patients (SMR from 12·0 for pulmonary to 24·0 for second cancer) and NHL patients (SMR from 6·5 for cardiac to 15·6 for second cancer). Late mortality due to recurrent disease and complications of therapy has also been described among leukaemia patients who underwent a blood or marrow transplantation (Socie et al, 1999).

In conclusion, this review has detailed selected topics relating to the late-effects of therapy for childhood and adolescent patients with leukaemia and lymphoma (Table II–IV). Within the spectrum of potential late effects, for this ever-increasing population of cancer survivors, there are other important outcomes that clearly can have a significant impact on subsequent quality of life. Examples of other outcomes include muscoloskeletal effects (e.g. osteoporosis, avascular necrosis, scoliosis), dental complications (e.g. developmental defects of tooth enamel and roots), ocular function (e.g. Sjogren's syndrome, cataracts, glaucoma), pulmonary compromise (e.g. restrictive lung disease, fibrosis), reproduction (e.g. fertility, pregnancy outcomes, health of offspring, premature menopause), haematological and immunological function (e.g. bone marrow reserve, cell-mediated immunity), and psychosocial status (e.g. employment, marriage, education).

Table II.  Commonly occurring late effects after conventional therapy for Hodgkin's disease in childhood and adolescence.
Adverse outcomeTreatment associated
with an increased risk
Factors associated with highest risk
HypothyroidismRadiation to the thyroid gland
(neck, mantle, etc.)
Increasing dose, females, age at treatment
HypogonadismAlkylating agents, abdominopelvic
radiation, gonadal radiation
Males, treatment during peripubertal or postpubertal period
in girls, higher cumulative doses of alkylators
Breast cancerRadiation to the chestIncreasing doses, females
Therapy-related myelodysplasiaAlkylating agents, topoisomerase II
inhibitors
Increasing doses of alkylating agents, older age at therapeutic
exposure
Thyroid cancerRadiation to the thyroid gland
(neck, mantle, etc.)
Increasing doses, females, younger age at radiation
Skin cancer (basal cell,
squamous cell, melanoma)
Radiation therapyOrthovoltage radiation (prior to 1970): delivery of greater dose to skin, additional excessive exposure to sun, tanning booths
Cardiomyopathy/
congestive heart failure
AnthracyclinesHigh cumulative doses (> 550 mg/m2 in patients > 18 years of age;  > 300 mg/m2 for patients < 18 years of age), females, younger than 5 years at treatment, African–American race
Pulmonary fibrosis/
interstitial pneumonitis
Bleomycin, radiation to chest,
carmustine
Younger age at treatment, bleomycin dose > 400 U/m2
Myocardial infarctionRadiation to chestUnderlying risk factors for coronary artery disease (smoking,
hypertension, hyperlipidaemia, obesity, etc.)
Chronic hepatitis C and
HCV-related sequelae
(cirrhosis, hepatic failure,
hepatocellular carcinoma)
Transfusions before 1993Living in hyperendemic area, multiple transfusions
Psychosocial outcomes
(depression, anxiety,
post-traumatic stress)
 Females
Table III.  Commonly occurring late effects after conventional therapy for leukaemia in childhood and adolescence.
Adverse outcomeTreatment associated
with an increased risk
Factors associated with highest risk
Neurocognitive deficitsCranial radiation, intrathecal
methotrexate, cytarabine
Females, younger age at treatment, increasing dose
HypothyroidismRadiation to the thyroid gland
(craniospinal)
Increasing dose, females, age at treatment
Osteopenia/osteoporosisCorticosteroids, craniospinal radiation,
gonadal radiation, methotrexate
Associated hypothyroidism, hypogonadism, growth hormone
deficiency
HypogonadismAlkylating agents, craniospinal
radiation, abdominopelvic radiation,
gonadal radiation
Males, treatment during peripubertal or postpubertal period
in girls, higher cumulative doses of alkylators
Precocious pubertyCranial radiationFemales, younger age at treatment, radiation dose > 18 Gy
Avascular necrosisCorticosteroidsDexamethasone, adolescence
Growth hormone deficiencyCranial radiationYounger age at treatment, radiation dose > 18 Gy
ObesityCranial radiationYounger age at treatment (< 5 years), females, higher cranial
radiation dose
Secondary CNS tumourCranial radiationIncreasing dose, younger age at treatment
Therapy-related leukaemiaTopoisomerase II inhibitorsSchedule (weekly or twice weekly administration)
Skin cancer (basal cell,
squamous cell, melanoma)
Radiation therapyOrthovoltage radiation (prior to 1970): delivery of greater dose
to skin, additional excessive exposure to sun, tanning booths
Dental abnormalitiesCranial radiationYounger age at treatment
CataractsCranial radiation, steroidsHigher radiation dose, combination of steroids and radiation,
single daily fraction
Chronic hepatitis C and
HCV-related sequelae
(cirrhosis, hepatic failure,
hepatocellular carcinoma)
Transfusions before 1993Living in hyperendemic area, multiple transfusions
Cardiomyopathy/
congestive heart failure
AnthracyclinesHigh cumulative doses (> 550 mg/m2 in patients > 18 years of age; > 300 mg/m2 for patients < 18 years of age), females, younger than 5 years at treatment, African–American race
Table IV.  Commonly occurring late effects after conventional therapy for NHL in childhood and adolescence.
Adverse outcomeTreatment associated
with an increased risk
Factors associated with highest risk
Neurocognitive deficitsCranial radiationFemales, younger age at treatment, increasing dose
HypothyroidismRadiation to the thyroid gland
(neck, mantle, etc.)
Increasing dose, females, age at treatment
Osteopenia/osteoporosisSteroids, cranial radiation, abdominopelvic
radiation, gonadal radiation, methotrexate
Associated hypothyroidism, hypogonadism, growth
hormone deficiency
HypogonadismAlkylating agents, craniospinal radiation,
abdominopelvic radiation, gonadal
radiation
Males, treatment during peripubertal or postpubertal period in girls, higher cumulative doses of alkylators
Avascular necrosisCorticosteroids, high dose radiation to
any bone
Dexamethasone, adolescence, males
Secondary CNS tumourCranial radiationIncreasing dose, younger age at treatment
Therapy-related myelodysplasiaAlkylating agents, topoisomerase II
inhibitors
Increasing doses of alkylating agents, older age at
therapeutic exposure
Skin cancer (basal cell,
squamous cell, melanoma)
Radiation therapyOrthovoltage radiation (prior to 1970): delivery of greater dose to skin, additional excessive exposure to sun, tanning booths
Chronic hepatitis C and
HCV-related sequelae
(cirrhosis, hepatic failure,
hepatocellular carcinoma)
Transfusions before 1993Living in hyperendemic area
Cardiomyopathy/
congestive heart failure
AnthracyclinesHigh cumulative doses (> 550 mg/m2 in patients > 18 years of age; > 300 mg/m2 for patients < 18 years of age), females,
younger than 5 years at treatment, African–American race
Dental abnormalitiesCranial radiationYounger age at treatment

Through ongoing research efforts, knowledge of the late-effects associated with cancer in children and adolescents continues to increase. However, much of the available information relates to outcomes within the first decade following treatment, and only minimal data address the longer-term outcomes that may occur later in adulthood. It is critical that we also understand the potential long-term impact of cancer therapy if we are to effectively counsel survivors and offer effective intervention strategies to prevent or minimize the impact of adverse late effects. Research is needed to more clearly define those survivors at greatest risk of specific outcomes. The process of identifying high-risk populations is essential to the development and testing of rational and effective interventions. These interventions need to include scientifically valid evidence-based recommendations for the clinical follow-up of survivors, which will not only include screening for potential late effects but application of proven approaches for health promotion.

Whether focusing on children, adolescents, young adults, middle-age adults or the elderly, the landscape of cancer survivorship research is continually changing. As cancer treatment approaches change through the introduction of new: (1) therapeutic agents or combinations of agents, (2) radiation oncology techniques, (3) surgical procedures, or (4) supportive care, the potential for late effects of treatment also changes. This is particularly true in paediatric leukaemia and lymphoma where, within the context of excellent survival, the knowledge gained on the occurrence of long-term adverse outcomes is being applied to the design and testing of treatment strategies that will maintain excellent survival while avoiding treatment-related late effects.

There will be an ongoing need to systematically follow the populations of survivors produced by these new treatment strategies. These efforts should be carried out within the context of scientifically valid research, which will probably require large, multi-institutional initiatives through existing co-operative clinical trials groups or consortia established expressly for the study of long-term survivors. An example of the latter is the Childhood Cancer Survivor Study, which has proven to be a valuable research resource for investigations among adult survivors of childhood and adolescent cancers (Robison et al, 2002). Similar initiatives are underway in the UK and Canada. Through a multidisciplinary approach to the diagnosis, treatment and long-term follow-up of paediatric leukaemia and lymphoma patients, the goal of cure, while minimizing the occurrence of long-term adverse outcomes, is clearly achievable.

Ancillary