Chemotherapy

The relationship between chemotherapy and SMN is likely more complex than that with RT, because agents appear to have different oncogenic potential, and a number of different mechanisms are associated with SMN development. Table 62-5 summarizes some of the common chemotherapeutic agents and subsequent malignancies with which they are associated.

Chemotherapy-related SMNs have been studied for several decades. Acute myeloid leukemia, acute lymphoblastic leukemia, and myelodysplastic syndrome are largely associated with alkylating agents often, but not exclusively, after Hodgkin lymphoma. In fact, Hodgkin lymphoma serves as a model for the study of balancing efficacy with toxicity when considering primary treatment. Hodgkin lymphoma is a primary malignancy that is associated with a high cure rate, but the very therapy that resulted in that cure rate shares the etiologic limelight for risk of subsequent leukemia. This phenomenon reflects the use of chemotherapy agents that unfortunately also have leukemogenic potential, such as mechlorethamine (nitrogen mustard) and procarbazine. To better quantify the risk of leukemia associated with alkylating agents, Meadows and colleagues⁶⁵ developed an alkylating agent score to describe the number of different alkylating agents administered and the duration of therapy for any patient. With use of this model, the risk of subsequent leukemia was found to increase with alkylating agent score.⁶⁵ Replacement of mechlorethamine by cyclophosphamide in Hodgkin lymphoma protocols and/or use of doxorubicin, bleomycin, vinblastine, and dacarbazine instead of or to replace some of the alkylating agent have both led to reductions in subsequent leukemia risk.^66–69

Alkylating agents cause unstable positively charged ions that participate in covalent bonds with electron-rich sites on DNA bases. Resultant adduct formation can induce apoptosis or mutagenesis. Interstrand cross-links and breaks also occur. DNA adducts may lead to configurational changes in the base and subsequent mismatch. Bulky DNA adducts or strand breaks that are large or occur at critical areas can result in disruption of DNA replication. Point mutations or deletions may result in either activation of an oncogene or mutation of a tumor suppressor gene. Alkylating agent–associated leukemia or MDS is typically characterized by abnormalities of chromosomes 5q or 7q, where putative tumor suppressor genes are postulated. Accompanying mutations in RAS oncogenes are also present in a number of cases.70–74

Topoisomerase II inhibitors are another class of chemotherapy agents associated with subsequent leukemia. Topoisomerase II is a nuclear enzyme that catalyzes and regulates both DNA cleavage and ligation and is essential for DNA replication, transcription, and repair. Topoisomerase II inhibitors form a complex between the topoisomerase II and the DNA, and the ligase activity for DNA repair is inhibited; in addition, illegitimate recombination events occur. Lastly, active metabolites can lead to DNA adduct or free radical formation, leading to mutagenesis.75,76 Most of the translocations involve the MLL gene at 11q23 and fuse MLL with a wide variety of partners. However, subsequent leukemia with other chromosomal translocations such as 8;21, 3;21, inv16,8;16, 15;17, and 9;22 are reported.^77–81 Epipodophyllotoxins, which are strong topoisomerase II inhibitors, specifically etoposide, and in the past teniposide, have been associated with subsequent leukemia after treatment for leukemia and a number of solid tumors. Risk appears related to the schedule of administration, cumulative dose, and concomitant administration of either alkylating agents such as cyclophosphamide or other topoisomerase inhibitors such as doxorubicin.^48,82–89 In a case-control study of 61 patients with secondary leukemia who were matched with 196 control subjects, two factors were found to increase the risk of leukemia in multivariate analysis, namely, the type of the first tumor, with an excess risk in patients with Hodgkin lymphoma (RR = 6.4) or osteosarcoma (RR = 5), and exposure to epipodophyllotoxins and anthracyclines. The risk of leukemia increased regularly with the cumulative dose of etoposide. Patients who received between 1.2 and 6 g/m² of epipodophyllotoxins or more than 170 mg/m² of anthracyclines had a sevenfold increased risk of leukemia.⁹⁰ In the more recent era of chemotherapy for Hodgkin lymphoma, etoposide has been incorporated in a number of upfront regimens. In these regimens, doxorubicin was also used, and dexrazoxane was used for cardioprotection, resulting in the administration of three topoisomerase II inhibitors in a relatively short period. In an analysis of two such Hodgkin lymphoma studies, 239 patients were assigned randomly to receive dexrazoxane for cardioprotection and 239 did not receive this agent. SMNs developed in 10 patients. Six of eight patients in whom both leukemia and solid tumors (osteosarcoma and papillary thyroid carcinoma) developed were recipients of dexrazoxane. The SIR for all SMNs was 41.86 versus 10.08 with and without dexrazoxane, respectively (P = .02).⁹¹

In the most recent comprehensive analysis of SMNs from the CCSS, in a multivariable model, risk of any subsequent neoplasm was increased by female sex (RR = 1.5), older age at primary childhood cancer diagnosis (RR = 1.3), radiation therapy exposure (RR = 2.7), and primary diagnosis of Hodgkin lymphoma (RR = 1.5), but not by chemotherapy. However, secondary breast cancer was associated with increased anthracycline exposure, together with primary diagnosis of Hodgkin lymphoma and radiation therapy. Sarcomas, which represented 15% of the subsequent malignancies, were associated with alkylating agent and anthracycline exposure, as well as radiotherapy and primary sarcoma diagnosis.⁴⁶ These data highlight the complexity of risk factor analysis in the era of multimodality therapies.

Chemotherapy-associated leukemia is not limited to treatment of childhood cancer or Hodgkin lymphoma. In a case-control study of 96 women with secondary leukemia and 272 control subjects in a population-based cohort of 28,971 women in North America and Europe who were treated for ovarian cancer between 1980 and 1993, among the women who received platinum-based combination chemotherapy for ovarian cancer, the RR of leukemia was 4.0 (6.5 for carboplatin and 3.3 for cisplatin), with evidence of a dose-response relation, with RRs reaching 7.6 at doses of 1000 mg or more of platinum.⁹² Similar to data from Hodgkin lymphoma treatment, in a cohort study of 11,386 two-year survivors of non-Hodgkin lymphoma, 35 case patients with secondary leukemia were matched with 140 control subjects. Leukemia risk was elevated with exposure to prednimustine (RR = 13.4), regimens containing mechlorethamine and procarbazine (RR = 12.6), and chlorambucil with cumulative doses of ≥1300 mg (RR = 6.5). Cyclophosphamide regimens were associated with a small, nonsignificant increased risk.⁹³

Modifications of Treatment-Related Effects on SMN Risk

Although some chemotherapeutic agents appear to be more oncogenic than others, it is clear that whereas most patients with cancer are exposed to chemotherapy, an SMN develops in only a small proportion. Thus interindividual differences may be related to the interaction between the exposure to the chemotherapeutic agent with inherent host differences in drug disposition and pharmacodynamics and concurrent use of other medications competing for the same metabolic pathways. These pathways are all genetically determined, and thus relevant polymorphic gene variants may confer risk for SMN. These pathways include drug membrane transport, phase I and phase II metabolic pathways (described later), and DNA repair. Because there is considerable interindividual variability in the risk of SMN, despite common therapeutic exposures, it is possible that polymorphic variations in these pathways modify risk of SMN.

The ABCB1 gene product, P-glycoprotein, is a membrane efflux pump that extrudes chemotherapeutic agents by active transport, and the presence of P-glycoprotein in the plasma membrane confers a multidrug-resistant phenotype. Another membrane protein pump that is important for some chemotherapeutic agents is multidrug resistance–associated protein, and doxorubicin, which is associated with secondary leukemia, is one of its substrates. One may hypothesize that such a transport pump may act as a host factor, influencing SMN risk from doxorubicin by modulation of the intracellular accumulation.⁹⁴

Although there are known disorders of DNA repair that infer high individual attributable risks for cancer, such as Fanconi anemia,⁹⁵ they are uncommon in the general population. However, polymorphic variants of a number of DNA repair genes likely exist that may contribute to a gene-environment interaction, whereby both the polymorphic variant and the exposure (e.g., chemotherapy or radiotherapy) increase risk for SMN. This interaction might explain some of the interindividual variability in treatment-associated SMN. Mismatch repair abnormalities have been associated with treatment-associated myelodysplasia and leukemia, wherein patients are found to have microsatellite instability and methylation of MLH1.^98–98 In the homologous repair pathway, a polymorphism of RAD51 has been shown to be overrepresented in treatment-associated myelodysplasia and leukemia, with further increased risk conferred by the presence of a polymorphism in XRCC3, a gene in the base excision repair pathway.⁹⁹ However, even within these repair pathways, variation of effect by gene occurs, as shown by a protective role of a variant of XRCC1 for treatment-associated myelodysplasia and leukemia and BCC.¹⁰⁰ An identified variant in the nucleotide excision pathway, ERCC2, has also been associated with increased risk for treatment-associated myelodysplasia and leukemia.

Phase I metabolism is primarily performed by the cytochrome P450 (CYP) family of enzymes and involves activation of substrates into highly reactive electrophilic intermediates that can damage DNA. Phase II pathways, such as glutathione S-transferase or arylamine N-acetyltransferases and NAD (p) H : quinone oxidoreductase-1, participate in inactivation or detoxification of genotoxic substrates.103–103 However, as was seen with the DNA-repair enzymes and SMN risk, associations with variants in these metabolizing genes are modest, and although many studies have attempted to demonstrate associations between polymorphic variants in these DNA repair pathways and SMN, results are mixed or inconclusive. Thus research in this area is ongoing.^104–110

Environmental Exposures

Although it is clear that for many SMNs, particularly solid organ tumors, risk continues to rise with ongoing time since treatment, that period is also marked by lifestyle and environmental exposures such as tobacco, alcohol, ultraviolet light, obesity, a sedentary lifestyle, immune suppression, and hormonal changes, all of which have a role in primary cancers. The intersection of such factors in the etiology of SMN is difficult to determine, because one would need to simultaneously evaluate genetic risks, therapeutic exposures from the primary cancer, and many environmental exposures in the same cohort, which would be difficult to assemble and for which it would be difficult to provide adequate statistical power. However, tobacco and alcohol are both risk factors for multiple malignancies involving the lung, gastrointestinal system, biliary system, oral mucosa and cavity, pharynx, and upper aerodigestive track, where the effect is synergistic. In the most recent SEER monograph on SMN, it was estimated that in terms of absolute excess risk, tobacco- and alcohol-related cancer sites account for 35% of SMNs.1,111

The role of diet and nutrition in SMN risk is much less well defined. Associations between colorectal cancer, ovarian cancer, endometrial cancer, and breast cancer are well known,¹¹² as are data suggesting some role in decreased recurrence or mortality risk in breast cancer survivors who made dietary changes that included increased soy, decreased fat, and increased fruits and vegetables; however, results are not fully consistent.^115–115 No studies have been performed to investigate the role of dietary or exercise intervention specifically to reduce the risk of SMN.

Similarly, sun exposure increases the risk for primary cutaneous melanoma and BCC. Both of these cancers are now seen after primary cancer treatment, particularly after childhood cancer and hematopoietic cell transplantation.²⁷ However, the role of sun exposure remains relatively unknown because these malignancies in both the childhood cancer and stem cell populations are clearly related to the receipt of radiation therapy. Whether sun exposure creates an additive risk is unclear, but cancer survivors are advised to avoid excess sun exposure and use sunscreen. It is plausible to hypothesize that the radiation therapy from cancer therapy causes sensitization such that excess ultraviolet exposure from the sun has an additive effect.

Childhood Cancer Survivors

In the most recent analysis from the CCSS referenced earlier, the cumulative incidence at 30 years after the childhood cancer diagnosis was 20.5% for all subsequent neoplasms, 7.9% for second malignant neoplasms (excluding nonmelanoma skin cancer), 9.1% for nonmelanoma skin cancer, and 3.1% for meningioma. Excess risk was evident for all primary diagnoses, with the highest being for Hodgkin lymphoma and Ewing sarcoma. In the Poisson multivariable analysis, female sex, older age at diagnosis, earlier treatment era, diagnosis of Hodgkin lymphoma, and treatment with radiation therapy were associated with increased risk of a subsequent neoplasm.⁴⁶

In two analyses from the CCSS on subsequent thyroid cancer, more defined risks related to radiation dose and concomitant alkylating agent therapy were elicited. The 30-year cumulative incidence of thyroid cancer is 1.3% for women and 0.6% for men. Thyroid cancer risk increased linearly with radiation dose up to approximately 20 Gy, where the relative risk peaked at 14.6-fold, after which a downturn in the dose-response relationship was observed. A higher radiation risk was found with younger age, among females, and with longer time since exposure. Among patients with thyroid radiation doses of 20 Gy or less, treatment with alkylating agents was associated with a significant 2.4-fold increased risk of thyroid cancer. Chemotherapy-related risk was evident for thyroid radiation doses more than 20 Gy.116,117

Moreover, childhood cancer survivors are at risk for carcinomas that are uncommon in young persons, as well as multiple subsequent neoplasms. In another report from the CCSS, at a median age of 27 years and a median elapsed time of 15 years, 71 carcinomas were diagnosed in the genitourinary system (35%), head and neck area (32%), gastrointestinal tract (23%), and other sites (10%). Fifty-nine patients (83%) had received radiotherapy, and a second malignant neoplasm developed in a previous radiotherapy field in 42 patients (59%).¹¹⁸ In a recent CCSS analysis of 1382 survivors with one subsequent neoplasm, a second subsequent neoplasm developed in 27.9%, and of those, 39.6% experienced more than two subsequent neoplasms. Radiation increased risk for more than one subsequent neoplasm, and persons with nonmelanoma skin cancer had a 15-year cumulative incidence of SMN (excluding nonmelanoma skin cancer) of 20.3% compared with only 10.7% for persons exposed to radiation and whose first subsequent neoplasm was an invasive SMN.¹¹⁹

In another large cohort of childhood cancer survivors (a combined cohort of 8884 North American, 2893 British, and 1574 Nordic subjects with Wilms tumor diagnosed before 15 years of age during 1960-2004), 174 solid tumors (exclusive of BCCs) and 28 leukemias were ascertained in 195 subjects. The cumulative incidence of solid tumors at age 40 years was 6.7%. Leukemia risk was highest during the first 5 years after Wilms tumor diagnosis. Standardized incidence ratios for solid tumors and leukemias were 5.1 and 5.0, respectively.¹²⁰

Sarcomas

Sarcomas of the bone and soft tissues (osteogenic sarcoma, Ewing sarcoma, and soft tissue sarcoma) are associated with increased risk for SMN, and similarly, sarcomas are common second malignancies following childhood cancer. In a report from the CCSS, Henderson and colleagues¹²¹ reported that 108 of 14372 participants had secondary sarcomas, diagnosed a median of 11 years after the diagnosis of childhood cancer, with a risk ninefold that of the general population. Risk was elevated with radiation therapy (RR = 3.1), primary diagnosis of sarcoma (RR = 10.1), history of other secondary neoplasms (RR = 2.2,), and treatment with anthracyclines (RR = 2.3) or alkylating agents (RR = 2.2).¹²¹ In the CCSS, the SIRs for SMN after primary soft tissue, Ewing sarcoma, and osteosarcoma were 5.8, 8.5, and 4.2, respectively. Risks of secondary bone or soft tissue sarcoma were 19.0-fold and 8.1-fold that expected in the general population.⁴⁶ Of particular interest is the risk of SMN after diagnosis and treatment of Ewing sarcoma. In an analysis of 1166 patients with Ewing sarcoma in the SEER registry who were treated between 1973 and 2005, an SMN developed in 35 patients. The estimated risks of SMN at 5, 10, and 20 years after original diagnosis were 2.1%, 4.4%, and 8%, respectively. The risk of SMN was higher in patients diagnosed before the age of 20 years in the earlier years of the cohort (though 1985) and in those treated with radiation therapy.¹²² In a report from the CCSS, 36 SMNs were reported in 34 of 403 patients who had Ewing sarcoma, with a cumulative incidence of 9%. The SIR in all Ewing sarcoma survivors, those treated with radiation, and those not treated with radiation was 5.9, 6.6, and 3.3, respectively. Notably, 13 breast cancers developed in 11 women, 7 of which were in the radiation field. The SIR for subsequent breast cancer after whole-lung radiation was 36.0.¹²³ In a review from St. Jude Children’s Research Hospital, among 237 patients treated for Ewing sarcoma, the cumulative incidence of SMN was 3% and 4.7% at 5 and 10 years, respectively, translating to a SIR of 17.8.¹²⁴

Hodgkin Lymphoma Survivors

Several large cohort studies have established the high risk of SMN after diagnosis and treatment of Hodgkin lymphoma. Among 35,511 survivors from 14 population-based cancer registries in Nordic countries and North America during 1970-2001, 217 were diagnosed with leukemia, whereas it was expected that 10.8 would be diagnosed. Excess absolute risk was highest during the first 10 years after Hodgkin lymphoma diagnosis but remained elevated thereafter. Risk was higher for those whose Hodgkin lymphoma was diagnosed at ≥35 years of age than in those diagnosed before 35 years of age. The risk declined after 1984, which may be associated with modifications in chemotherapy.¹²⁵

In an international population-based cohort of 3817 female 1-year survivors diagnosed from 1965 to 1994, for survivors who were treated at ≤25 years of age with a chest radiation dose of at least 40 Gy without alkylating agents, the estimated cumulative absolute risks of breast cancer by ages 35, 45, and 55 years were 1.4%, 11.1%, and 29.0%, respectively. Cumulative absolute risks were lower in women who were treated with alkylating agents.¹²⁶ In a case-control study from this cohort, women who received a ≥5-Gy radiation dose to the breast had a 2.7-fold increased breast cancer risk, compared with those given less than a 5-Gy dose.¹²⁷

In another cohort of 18,862 5-year survivors from 13 population-based cancer registries in North America and Europe, 1490 SMNs were identified, with 850 estimated to be in excess. For most cancer sites, risk decreased with increasing age at Hodgkin lymphoma diagnosis and showed strong dependencies on attained age. Thirty-year cumulative risks of solid organ SMNs for men and women diagnosed at 30 years were 18% and 26%, respectively, compared with 7% and 9%, respectively, in the general population. For young patients, risks of breast and colorectal cancers were elevated 10 to 25 years before the age when routine screening would be recommended in the general population.¹²⁸

Within a population-based cohort of 19,046 patients diagnosed from 1965 through 1994, a case-control study of lung cancer was conducted, comparing 222 patients in whom lung cancer developed and 444 matched control subjects. Treatment with alkylating agents without radiotherapy was associated with increased lung cancer risk, as was radiation dose to the lung of 5 Gy or more without alkylating agents. Risk increase was associated with both an increasing number of cycles of alkylating agents and increasing radiation dose. Excess risk after treatment with radiotherapy began 5 years after treatment and persisted for more than 20 years. Tobacco use increased lung cancer risk more than twentyfold.¹²⁹ In the Late Effects Study Group cohort of 1380 children diagnosed with Hodgkin lymphoma from 1955 to 1986 in patients aged 16 years or younger, with 212 SMNs reported, the cohort was at an 18.5-fold increased risk of the development of SMNs compared with the general population. The cumulative incidence of any second malignancy was 10.6% at 20 years, increasing to 26.3% at 30 years, and the cumulative evidence of solid malignancies was 7.3% at 20 years, increasing to 23.5% at 30 years. Breast cancer was the most common solid malignancy, but other commonly occurring solid malignancies included thyroid cancer, bone tumors, and colorectal and gastric cancers. Third neoplasms developed in 32 patients of the 212 SMNs, with the cumulative incidence approaching 21% at 10 years from diagnosis of second malignancy.¹³⁰

In the CCSS, 277 invasive SMNs have occurred among 257 five-year survivors, with the median time to first SMN being 18.7 years from diagnosis (range, 5.0-33 years). The 30-year cumulative incidence was 10.9% among men and 26.1% among women, driven by the incidence of invasive breast cancer. The risk was highest for bone, thyroid, and breast cancer. The cumulative incidence of nonmelanoma skin cancer was also high among Hodgkin lymphoma survivors at 16.7% at 30 years after diagnosis.¹³¹

Hematopoietic Cell Transplant Survivors

In a multiinstitutional cohort of 28,874 allogeneic transplant recipients, 189 solid malignancies occurred, with an observed-to-expected ratio 2.1-fold and threefold among patients followed up for 15 years or more after transplantation.¹³² Although a substantial increase in SMN occurs with the use of chemotherapy alone, the addition of TBI clearly causes an increase in second tumor formation. In a cohort of 8079 patients who had undergone hematopoietic cell transplantation (HCT) at the Fred Hutchinson Cancer Research Center and 5806 patients who had survived beyond 100 days, 381 SMNs were reported. Risk factors included treatment with TBI and, furthermore, use of TBI in a single dose. In addition, as the dose of fractionated TBI increased, the hazard ratio increased. Patients with TBI exposure had a 25-year incidence of SMN of 21%, compared with 10% for patients not exposed to TBI. Risk continued to rise with elapsed time since transplantation, without an obvious plateau. Of particular interest is that for patients younger than 10 years at HCT, those with TBI exposure had a 25-year incidence of SMN of 19%, compared with 5% for patients not exposed to TBI. For patients older than 10 years at hematopoietic stem cell transplantation, those with TBI exposure had a 25-year incidence of SMN of 21%, compared with 11% for patients not exposed to TBI. The effect of TBI in the entire cohort appeared to be due to younger patients without TBI having a very low incidence of SMN and older patients without TBI having a somewhat higher cumulative incidence of SMN.⁴⁵

In another report from the Fred Hutchinson Cancer Research Center, among 4810 patients who received allogeneic HCT and survived for at least 100 days, 237 experienced at least one BCC or squamous cell carcinoma (SCC) of the skin or mucosal surfaces (BCC, n = 158; SCC, n = 95). Twenty-year cumulative incidences of BCC and SCC were 6.5% and 3.4%, respectively. For BCC, TBI was a significant risk factor, most strongly among patients younger than 18 years at the time of HCT. Acute graft-versus-host disease increased the risk of SCC (P = .02), whereas chronic graft-versus-host disease increased the risk of both BCC and SCC.²⁷ Among 3337 female 5-year survivors who underwent allogeneic HCT at the Fred Hutchinson Cancer Research Center or at one of 82 centers reporting to the European Bone Marrow Transplant Registry, breast cancer developed in 52 survivors with a standardized incidence ratio of 2.2 and 25-year cumulative incidence of 11%. In multivariable analysis, increased risk was associated with longer time since transplantation, use of TBI, and younger age at transplantation.¹³³

Prevention and Early Detection of Subsequent Malignancies

No clear evidence exists for strategies for primary prevention of SMN, although changes in therapy that limit exposures to alkylating agents, topoisomerase II inhibitors, and radiotherapy will serve to decrease risk. However, the significant advances in survival from pediatric and adult cancers during the past few decades should not be overshadowed by the risk of SMN. Although the balance of efficacy with short- and long-term toxicity should always be considered in the choice of therapy for a primary cancer, curative therapeutic options should not be withheld, even if they produce risk for SMNs. Survivors who are at increased risk for SMNs should certainly avoid additional known carcinogenic environmental exposures but should also be educated about their risk for SMNs and any specific screening that is recommended (Table 62-6). The Institute of Medicine has recognized the need for a plan of surveillance and monitoring for adverse long-term effects of cancer and its therapy among survivors of childhood and adult-onset cancers, including SMN. Recommendations also include enhanced communication between oncologists and primary care physicians and an interdisciplinary model of care, but no consensus has been reached about how such care should be delivered, in what setting, by what providers, and at what intervals.^134,135 Most childhood cancer institutions now have survivorship clinics to follow-up patients and advise or conduct specific screening for subsequent cancers. A number of cancer centers have similar programs for survivors of adult-onset cancer and/or adult survivors of childhood cancer. Guidelines and recommendations that can inform screening and follow-up are found in some of the disease-specific guidelines of the National Comprehensive Cancer Network and the American Cancer Society, and more detailed consensus-based screening guidelines based on exposure are available from the Children’s Oncology Group, and for those treated with stem cell transplant, from the National Marrow Donor Program. Table 62-7 provides a brief summary of common sites for subsequent cancers, risk factors, and screening or prevention, adapted from the Children’s Oncology Group. In a recent review on SMN, Wood and colleagues¹¹¹ have provided some suggestions for survivors of breast, colon, ovarian, endometrial, skin, lung and prostate cancer. Table 62-8 provides other sources for follow-up recommendations for cancer survivors in greater detail, where information is regularly updated.