National Institutes of Health Hematopoietic Cell Transplantation Late Effects Initiative: Consensus Recommendations for Subsequent Neoplasms

Abstract

Subsequent neoplasms (SN) following hematopoietic cell transplantation (HCT) cause significant patient morbidity and mortality. Risks for specific SN types vary substantially, with particularly elevated risks for post-transplant lymphoproliferative disorders, myelodysplastic syndrome/acute myeloid leukemia, and squamous cell malignancies. This consensus document provides an overview of the current state of knowledge regarding SN after HCT and recommends priorities and approaches to overcome challenges and gaps in understanding. Numerous factors have been suggested to impact risk, including patient-related (e.g., age), primary disease-related (e.g., disease type, pre-HCT therapies), and HCT-related characteristics (e.g., type and intensity of conditioning regimen, stem cell source, development of graft-versus-host disease). However, gaps in understanding remain for each of these risk factors, particularly for patients receiving HCT in the current era due to substantial advances in clinical transplantation practices. Additionally, the influence of non-transplant-related risk factors (e.g., germline genetic susceptibility, oncogenic viruses, lifestyle factors) is poorly understood. Clarification of the magnitude of SN risks and identification of etiologic factors will require large-scale, long-term, systematic follow-up of HCT survivors with detailed clinical data. Most investigations of the mechanisms of SN pathogenesis after HCT have focused on immune drivers. Expansion of our understanding in this area will require interdisciplinary laboratory collaborations utilizing measures of immune function and availability of archival tissue from SN diagnoses. Consensus-based recommendations for optimal preventative, screening, and therapeutic approaches have been developed for certain SN after HCT, whereas for other SN, general population guidelines are recommended. Further evidence is needed to specifically tailor preventative, screening, and therapeutic guidelines for SN after HCT, particularly for unique patient populations. Accomplishment of this broad research agenda will require increased investment in systematic data collection with engagement from patients, clinicians, and interdisciplinary scientists in order to reduce the burden of SN in the rapidly growing population of HCT survivors.

BACKGROUND AND PURPOSE

Hematopoietic cell transplantation (HCT) is an important component of treatment for many malignant and nonmalignant diseases. Over the last several decades, the number of HCTs performed annually has increased, surpassing 18,000 in the United States in 2014.¹ With improvements in clinical approaches to HCT and supportive care, survival following HCT has improved dramatically.^2–4 The current population of >100,000 survivors in the United States is expected to increase five-fold by 2030, with 14% of the population aged <18 years and 25% aged ≥60 years at transplant.⁵

Subsequent neoplasms (SN) are a leading cause of non-relapse mortality among HCT survivors.^3,6–8 SN are typically grouped into three categories: B- and T-cell malignancies (including post-transplant lymphoproliferative disorder [PTLD]), myelodysplastic syndrome/acute myeloid leukemia (MDS/AML), and solid tumors. The time course for development of SN after HCT is variable, with lymphomas and leukemias developing relatively early, whereas solid tumors tend to have a longer latency.^6,9–14

Here we summarize what is currently known about the magnitude of risk for SN and their pathogenesis, transplant- and non-transplant-related risk factors, and outcomes, as well as current guidelines for prevention and screening. When appropriate, autologous and allogeneic HCT are considered separately. To summarize existing knowledge, we searched the databases of PubMed and Embase spanning articles published during 1995–2015 using medical subject headings including “Neoplasms, Second Primary,” “Bone Marrow Transplantation,” and “Hematopoietic Stem Cell Transplantation,” as well as additional headings relevant for specific topics (e.g., “Transplantation Conditioning,” “Graft-versus-Host Disease [GVHD]”). A working group consensus over a 12-month period then defined significant gaps in understanding and developed recommendations for future research regarding SN after HCT. The highest priority recommendations are summarized in the text box. Additional information on research gaps and recommendations are found within each section below.

MAGNITUDE OF RISK

We identified 62 publications that quantified SN risk (Supplementary Table 1). About half (N=34) of the studies included patients undergoing HCT for different indications,^15–48 whereas the remainder focused on specific diseases, including myeloid malignancies (N=1),¹¹ lymphoid malignancies (N=15),^49–63 plasma cell disorders (N=3),^64–66 and benign hematologic disorders such as aplastic anemia and Fanconi anemia (FA) (N=9).^67–76 Only three were prospective.^34,36,51

Few studies included a comparison of SN occurrence to that expected in the general population. The largest study quantified risks of solid tumors by anatomic site in an international collaborative effort including 28,874 patients who received allogeneic HCT during 1964–1994.⁹ Overall, SN occurred more than twice as often as expected in the general population. Those SN with the greatest excess risks included lip (standardized incidence ratio [SIR]=26.8), salivary gland (SIR=14.2), tongue (SIR=13.3), bone (SIR=8.5), soft tissue (SIR=6.5), and liver (SIR=6.3). The SIRs for most solid tumors increased with increasing time since transplantation, with the highest SIRs observed for tumors occurring ≥10 years after transplantation. The most notable exception to this pattern was for melanomas and thyroid cancers, where risks were persistently elevated from the first year through ≥10 years after transplantation. Similar findings were observed in an Australian cohort of 3273 adults who received allogeneic HCT during 1992–2007.⁷⁷ Risks for several types of skin cancer are also strikingly elevated after HCT,^78,79 as reported in the Danish study of 3302 HCT recipients during 1999–2014 that leveraged the comprehensive reporting of all cutaneous malignancies [including basal and squamous cell carcinomas (SCC)] to general population cancer registries.⁷⁷

Quantifying hematologic SN risks after HCT is more complicated than for solid tumors for two reasons. First, hematologic malignancies are the most common indication for HCT, thus relapsed disease must be distinguished from SN of donor origin. Distinguishing a donor-derived SN ideally utilizes sensitive molecular genetic analysis,³⁷ though in the case of an opposite sex donor, cytogenetic analysis has been used. Second, certain hematologic SN such as PTLD are not directly comparable to lymphoid malignancies that occur in the general population. Nevertheless, several large-scale studies have quantified risk of hematologic SN after HCT. In an international study of 26,901 allogeneic HCT recipients during 1964–1994, the observed to expected ratio of PTLD was 29.7.^13,80 MDS/AML risk after allogeneic HCT is thought to be comparable to that of other cancer survivors who have received particular cytotoxic chemotherapeutic agents and radiotherapy, with such risks often exceeding 10-fold.⁸¹ There is also some evidence that MDS/AML risks may be even higher after autologous HCT because of the substantial cumulative doses received during pre-transplant chemotherapy and radiotherapy, chemotherapy-based stem cell mobilization and HCT conditioning.^62,82,83

Cumulative incidence is particularly useful for quantifying the clinical burden of SNs. Although most studies reported the cumulative incidence of SN, substantial differences in the populations studied in terms of primary disease, age, and HCT approach, as well as the wide range in sample size and duration of follow-up, result in significant heterogeneity in the reported incidence, hampering comparison of studies and interpretation of results (Figure 1, Supplementary Table 1).

Figure 1.

Cumulative incidence of SN (percent) after HCT from reported studies demonstrates wide variability.

Research Gaps and Recommendations

• Large-scale, long-term (>10 years), systematic follow-up of HCT recipients for SN occurrence is essential to better define the magnitude of risks for specific SN types. The full spectrum of SN should be considered.

  • Substantial efforts will be required to minimize loss-to-follow-up and ensure comprehensive ascertainment of SN diagnoses.

  • Detailed patient and clinical data should be collected to investigate modifiers of SN risks, including information on patient characteristics, primary disease, pre-transplant treatment exposures, conditioning regimen, transplant-related factors, and other clinical factors. Standardized reporting of data elements would facilitate future research, as multi-institution collaboration will be essential to achieve statistical power for certain analyses. For certain data elements, longitudinal data collection is critical (e.g., clinical factors such as GVHD).

• When possible, studies should consider SN risks relative to the general population and absolute risks to quantify the clinical burden of SNs.

• Interdisciplinary laboratory studies are needed to better characterize donor-derived SN.

PATHOGENESIS

Substantial advances in technologies related to genomics and assessment of immune function have provided insight into tumor pathogenesis in the general population. In particular, comprehensive molecular profiling of tumors has identified new disease subtypes and signatures reflective of particular exposures.⁸⁴ However, much less is known about pathogenesis of SN. Most previous research has focused on immune drivers because the post-HCT milieu is associated with unique alterations that may promote the development of SN. After both autologous and allogeneic HCT, but especially after allogeneic HCT, a period of lymphopenia and cell-mediated immune deficiency affecting T- and NK-cells and antibody production occurs. This period can persist for months and is exacerbated by well-defined factors, including acute and chronic GVHD (aGVHD, cGVHD) as well as immunosuppressive therapy. The surge of interleukin (IL)-15 after HCT promotes rapid T-cell clonal expansion leading to a restricted T-cell repertoire.^85,86 Expanding donor T-cells often have the phenotype of large granular lymphocytes, but occasionally a true clonal expansion of large granular lymphocyte disease can develop.^87–92 Defective T-cell immunity permits reactivation of Epstein-Barr virus (EBV) and other oncogenic viruses, but post-HCT immune deficiency may increase the risk of other SN, as demonstrated in solid organ transplant recipients.^93,94 Solid tumors also may arise due to the DNA damaging effects of cytotoxic chemotherapy and radiotherapy used as pre-transplant therapies or in conditioning regimens, although the exact carcinogenic mechanisms are less well understood. Additionally, post-allogeneic HCT treatment with voriconazole to prevent infectious complications has been associated with increased risk of cutaneous SCC, paralleling observations among solid organ transplant recipients.⁹⁵

Hematologic malignancies may arise through accidental transplantation from an affected donor or develop in donated cells after transplantation. Donor-derived hematologic malignancies have particularly been reported following umbilical cord blood (UCB) transplants,^37,96–101 with support for the likely importance of the post-HCT milieu in promoting neoplasm development. Despite a rate of up to 6% donor-derived MDS/AML in the recipient, there have been no instances of leukemia in the UCB donor.¹⁰² Similarly, although the occurrence of multiple myeloma after allogeneic HCT may be due to transfer of malignant cells with the transplant,^103–106 the absence of reports of disease in the donor suggests that the recipient milieu is especially favorable to multiple myeloma. There are several reports from adult and UCB transplants that donor-derived leukemias do not always develop even with the transfer of hematopoietic progenitors bearing gene mutations.^107–111 Finally, research into the allogeneic graft-versus-leukemia effect reveals donor T-cell responses to tumor antigens and minor antigens, and that disease relapse can be provoked by immunosuppressive therapy.¹¹²

Donor-derived leukemia may be driven by telomere shortening consequent upon the rapid and intense expansion of the graft.¹⁰⁸ Following allogeneic HCT, the reduction in telomere length of donor hematopoietic cells is equivalent to approximately 15 years, and in some instances 40–60 years, of aging.¹¹³ In addition, telomere erosion is well recognized to cause significant genomic instability, leading to malignant transformation.¹¹⁴

The reports of solid tumors containing donor cells indicate that donor cells may either influence malignant change or more inexplicably transform into cancer cells.^115–118 In favor of the transfer of non-hematopoietic cells in the graft are several publications describing donor cells participating in non-hematopoietic tissues.^119,120 Techniques for identifying donor derived cancer cells relying on in situ hybridization have led to both positive and negative findings in animals¹²¹ and in humans.¹²²

Research Gaps and Recommendations

• Investigations of the pathogenesis of SN will require interdisciplinary laboratory collaboration, utilizing banked cryopreserved donor and recipient blood and marrow cells and SN tissues. Samples should be annotated with detailed patient and clinical data (e.g., stem cell source to investigate why UCB transplants have such a high propensity to develop donor malignancy).

• Longitudinal biomarker studies are needed to investigate the interaction of the donor’s immune system with donor hematopoiesis or against recipient malignancies arising in non-hematopoietic tissue. Unanswered questions include mechanisms underlying viral induced SN, factors that control the development of T-large granular lymphocytes, which changes in the recipient milieu and tissue niches promote SN and how, and whether donor cells found in non-hematologic malignancies cause the malignancy or in some way promote its development.

DISEASE- AND TRANSPLANT-RELATED RISK FACTORS

Primary disease and disease status

Several reports in both autologous and allogeneic HCT have suggested that SN risk may vary by primary disease type or status. Among 674 pediatric patients undergoing allogeneic HCT, a primary diagnosis of non-Hodgkin lymphoma (NHL) was associated with a four-fold increase in the risk of solid cancers.¹⁷ Among 28,874 patients undergoing allogeneic HCT, those transplanted for chronic myelogenous leukemia (CML) had a lower risk of solid cancers than patients with acute leukemia,³⁹ but this has not been confirmed in other studies.^37,38 Two studies reported increased SN risks with more advanced disease, including adults with lymphoid malignances undergoing autologous HCT (N=300)⁵⁹ and patients with FA undergoing allogeneic HCT (N=795).⁶⁹ However, disease status was not associated with SN risk in 1487 pediatric autologous HCT recipients with a variety of primary diagnoses.²⁴

Pre-HCT therapy and conditioning regimens

Among cancer survivors who have not received HCT, radiotherapy and certain systemic therapies have been associated with SN risks.¹²³ Similar studies after HCT are complicated by the complexity of treatment exposures patients receive as part of pre-HCT therapies and conditioning regimens.

In 268 adult patients with lymphoid cancers undergoing autologous HCT, pre-HCT nitrogen mustard, oncovin (vincristine), procarbazine and prednisone (MOPP) chemotherapy was associated with a six-fold higher rate of developing any SN compared to other pre-HCT regimens.⁵⁴ An almost eight-fold increase in risk of SN was reported after etoposide stem cell priming in one study.⁶² In contrast, two other studies found no association between pre-HCT chemotherapy and risk of SN.^41,62 Pre-HCT radiotherapy has been shown to be an adverse risk factor for SN.^59,60,62 Among 133 patients with aplastic anemia undergoing allogeneic HCT, pre-HCT immunosuppressive therapy with cyclosporine (a group 1 carcinogen according to the International Agency for Research on Cancer)¹²⁴ was associated with five-fold higher rates of SN.⁷⁵

Investigations of the role of conditioning regimens have largely focused on exposure to total body irradiation (TBI). The data suggest an interaction with age, with higher risk of SN in younger recipients exposed to TBI compared with older patients.^39,125 However, the role of TBI in SN risk remains unresolved, with six publications providing strong evidence of increased SN risk from TBI exposure,^{16,17,32,46,49,60} and eight studies failing to identify an association between TBI and SN.^{22,24,33,37,38,41,59,69}

In the non-TBI setting, patients undergoing allogeneic HCT with busulfan/cyclophosphamide (Bu/Cy) conditioning demonstrated 1.4-fold increased SN risk compared with the general population.¹¹ In the autologous setting, etoposide as part of the conditioning regimen was not associated with SN among pediatric patients.²⁴

Allogeneic donor source

In recent years, allogeneic transplantation practices have expanded to include unrelated donors (URD), haplo-identical related donors and stem cell sources other than bone marrow (e.g., peripheral blood, UCB). Because these practices are relatively new, their impact on risk for SN is poorly understood. Most studies have shown no association,^22,36,38,41 but they are generally small with limited follow-up.

A comparison of 355 human leukocyte antigen (HLA)-matched sibling donor (MSD) and 108 URD HCT recipients surviving >2 years (median follow-up=10 years) did not show an effect of donor source on SN risk.¹²⁶ Pre-HCT manipulation of donor stem cells (e.g., ex-vivo T-cell depletion) has been associated with increased risk of PTLD, with particularly elevated risks when using anti-thymocyte globulin or alemtuzumab in the preparative regimen, but no difference when using MSD or URD products.^127,128 A European Society for Blood and Marrow Transplantation (EBMT) study of donor cell leukemia after 10,489 allogeneic HCTs performed during 1975–1998 found 14 cases, of which 12 were from matched related donors and two from URD.⁴⁸ However, the small sample size and the paucity of URD in the study period limit any conclusions relating donor type to SN risk, and the only other data derive from case reports. A study from EBMT of 1036 recipients surviving >5 years (median follow-up=10.7 years, range 5–22) did not find donor-recipient histocompatibility to be a significant risk factor for development of new malignancies.¹²⁹ Unpublished data from the Fred Hutchinson Cancer Research Center (FHCRC) found that the relative risk of SN in allogeneic HCT recipients, after adjustment for both aGVHD and cGVHD, was higher in those recipients who received peripheral blood stem cell (PBSC) grafts compared to bone marrow (RR=1.6, 95%CI=1.2–2.10, p=0.001). The fewer number of UCB transplants and their shorter follow-up precluded an accurate assessment of risk in UCB recipients. In a cohort of 98 adult patients who received UCB grafts, the cumulative incidence of 19% SN at two years appeared high, with PTLD being the most common.³⁷ In autologous HCT recipients, risk for MDS/AML has been reported to be three-fold higher with PBSC grafts.¹²⁸

Post-HCT maintenance

Post-autologous HCT maintenance therapy with lenalidomide for multiple myeloma has been associated in some studies with an increased risk for SN, but the magnitude of risk varied and detailed analysis of other potential risk factors was lacking.^130–133 A meta-analysis of 3254 newly diagnosed patients treated on randomized controlled trials suggested that exposure to lenalidomide plus oral melphalan significantly increased the risk for hematologic malignancies but not solid tumors.¹³⁴ Retrospective reports from single institutions are difficult to interpret because of heterogeneity in patient populations.^135,136 A Center for International Blood and Marrow Transplant Research (CIBMTR) registry analysis did not demonstrate an increased risk in new malignancies with lenalidomide.⁶⁴ There are no data for more recently utilized maintenance therapies (e.g., Flt3 inhibitors).

GVHD

Patients typically receive some type of prophylactic systemic immunosuppressive therapy or manipulated graft (e.g., CD34 selection, T-cell depletion) to prevent GVHD. Those diagnosed with GVHD are primarily treated with immunosuppressive therapy and may continue therapy for an extended time. Combinations of graft manipulation, immunosuppressive therapy, and GVHD-related immune dysregulation result in long-term immune impairment and an environment that may promote SN development.

Chronic GVHD and SN risk

A number of studies found that patients with cGVHD developed SN overall at 2–3-fold higher rates than the general population (Supplementary Table 2).^{11,19,33,46,69,137–142} Risks associated with cGVHD vary by malignancy type (Supplementary Table 3). SCC is the most commonly reported SN. One of the largest studies of SN found a nearly three-fold increased risk for SCC associated with cGVHD.¹⁰ Risks increased slightly when cGVHD was preceded by aGVHD but more strikingly – to nearly 10-fold – for severe cGVHD. Although some studies found a similar overall risk estimate for SCC after cGVHD;⁷⁹ slightly higher risk estimates of approximately five-fold also have been reported for any SCC, regardless of time after transplant.⁹ Several studies have established that cGVHD increases risk in specific organs. The most frequently reported organs include skin, buccal cavity, and esophagus.^5,10,139,140 Less is known about risks for other malignancies, though elevated risks for thyroid cancer (RR=2.94)⁴⁶ and basal cell carcinoma (RR=1.6)⁷⁹ were reported after cGVHD in allogeneic HCT recipients.

GVHD treatment and SN risk

Studies of the influence of immunosuppressive therapy for GVHD on the risk of SN after allogeneic HCT have been infrequent, largely due to their inherent need for well-annotated cohorts. Nevertheless, there are reports in which immunosuppressive therapy was associated with increased risk of SN.^10,19 Chien et al. reported a 2.5-fold increased SN risk with use of azathioprine, at one point a relatively common drug used to treat cGVHD, and 5.5-fold non-significantly increased risk with sirolimus. (Supplementary Table 4).¹⁹

Curtis et al. reported several associations of immunosuppressive therapy with an increased risk of SCC for specific SN (Supplementary Table 5).¹⁰ When all systemic immunosuppressive therapy, both for GVHD prophylaxis and treatment, was considered, longer duration of immunosuppressive therapy increasingly raised the risk of SCC (2–4 years: RR=4.75; ≥4 years: RR=6.2), with particularly elevated risks for SCC in the oral cavity and skin. These associations generally persisted when only immunosuppressive therapy for treatment of cGVHD was considered. SCC risks also were increased five-fold for recipients of azathioprine given with steroids, and the addition of cyclosporine further increased risk, especially after one year (any duration: RR=18.61; ≥12 months: RR=38.71).

Research Gaps and Recommendations

• Understanding disease- and transplant-related risk factors for SN is critical for future efforts to reduce risks as well as identify patients at highest risk who would most benefit from prevention and surveillance efforts.

• Detailed data collected at consistent time points should include pre-transplant exposures, transplant-related factors, and clinical data during long-term follow-up. Specifically:

  • Studies that investigate the role of primary disease in SN risk should consider inherent biologic susceptibility related to the natural history of the primary disease, underlying genetic susceptibility to cancer, patient-related correlates (e.g., age), and cumulative carcinogenic effect of pre-transplant therapies.

  • Transplant-related variables should include degree of HLA matching, donor and graft source, graft manipulation, and conditioning regimens.

  • Post-HCT clinical variables should be collected longitudinally, including: post-HCT maintenance; GVHD prophylaxis, therapy, type, specific organ involvement and severity; and therapy for malignancy relapse.

• Future studies that include patients who received transplants more recently are critical for defining whether the burden of SN has been altered with newer HCT and GVHD therapy approaches, even as follow-up is extended for patients transplanted in the past to identify long-term risks.

NON-TRANSPLANT-RELATED RISK FACTORS

The role of non-transplant-related risk factors in the development of SN is relatively unexplored. Below, we consider a wide range of potential risk factors, from genetic susceptibility to viruses and other medical history factors as well as lifestyle factors.

Advances in microarray and sequencing technologies have rapidly advanced our understanding of genetic susceptibility to cancer in the non-HCT setting. Allogeneic HCT is the treatment of choice for severe aplastic anemia or hematological malignancies arising in patients with variety of inherited cancer susceptibility syndromes. However, risk of SN after allogeneic HCT in individuals with rare inherited cancer predisposition syndromes is not well studied. In 157 patients with FA who survived ≥2 years after HCT, the incidence of oral cavity cancers, mainly SCC, increased with time (8% and 14% at 10 and 15 years post-HCT, respectively), and SCC accounted for 40% of deaths.⁷⁶ In another study of 262 patients with FA, the 117 HCT recipients had a four-fold increased risk of developing SCC.¹⁴³

The importance of studying genetic susceptibility outside the context of rare inherited cancer predisposition syndromes is highlighted by literature on risk for therapy-related MDS/AML (tMDS/AML) after autologous HCT or following conventional chemotherapy. Initial studies on inter-individual variability in risk of tMDS/AML using a candidate gene approach have identified common polymorphisms in low penetrance genes that regulate the availability of active drug metabolite or those responsible for DNA repair. Associations have been observed for glutathione S-transferase genes (GSTM1, GSTT1 and GSTP1), particularly among those with prior exposure to chemotherapeutic agents that are known GSTP1 substrates,¹⁴⁴ and for P-glycoprotein (encoded by MDR1), a cell membrane protein that impacts response to numerous chemotherapeutic agents.¹⁴⁵ A number of studies also have evaluated key DNA repair genes, including mismatch repair family members MLH1^146,147 and MSH2,^148,149 RAD51,¹⁵⁰ and XRCC3.^151–153 Other polymorphisms in candidate genes from DNA repair pathways include ERCC2 of the nucleotide excision repair pathway¹⁵⁴ and two common functional p53-pathway variants.¹⁵⁵

As technology for larger-scale genomic studies has advanced, Knight et al. conducted a case-control genome-wide association study, comparing patients who had developed tAML and healthy controls.¹⁵⁶ The investigators identified three polymorphisms (rs1394384, intronic to ACCN1; rs1381392, intergenic; and rs1199098, in linkage disequilibrium with IPMK) to be associated with tMDS/AML with chromosome 5 and/or 7 abnormalities. Although the investigators were able to confirm findings in an independent replication cohort, the utilization of a non-cancer healthy control group raises concerns about the case-control differences being generated by the genetics of the primary cancer versus tMDS/AML.

Telomeres, the long TTAGGG nucleotide repeats and protein complexes at the end of chromosomes, are markers of cellular replicative capacity and aging.¹⁵⁷ Mutations in telomere genes cause dyskeratosis congenita, a rare inherited marrow failure and cancer susceptibility syndrome.¹⁵⁸ In the general population, genetic variations in telomere-biology genes, namely TERT¹⁵⁹ and RTEL1,¹⁶⁰ have been associated with the development of several cancers, and short telomere length with cancer risk and mortality in large epidemiological studies.^161,162 Although no studies have investigated telomere length and cancer risk after allogeneic HCT, several studies suggest that it may be important. Telomere shortening in the transplanted hematopoietic cells appeared to accelerate post-transplant when compared with age-matched controls or matched donors (reviewed in¹⁶³). Short telomeres after HCT were more common in females and those with cGVHD,¹⁶⁴ and pre-HCT short telomere length in recipients has been associated with transplant-related mortality.^165,166 Finally, in a longitudinal study including 287 lymphoma patients, accelerated telomere shortening was associated with MDS/AML risk after autologous HCT.¹⁶⁷

For a number of specific SN that occur in excess in allogeneic HCT recipients, there is substantial understanding of the key lifestyle/medical history etiologic factors in the general population. Examples include the role of tobacco, alcohol, and human papilloma virus (HPV) infections in oral cavity cancers;^168,169 sun exposure in melanoma;¹⁷⁰ and hepatitis C virus (HCV) infections, autoimmune diseases, and sun exposure in certain NHLs.¹⁷¹ However, the role of these risk factors in SN after allogeneic HCT is largely unknown. In a large analysis of outcomes of 4161 patients with plasma cell disorders undergoing autologous transplantation, obesity was associated with a nearly two-fold increase in SN risk.⁶⁴

Investigations should also consider whether other cancer risk factors have an additive or multiplicative effect with transplant-related risk factors. For example, immune activation or chronic immune damage of tissues from GVHD could act synergistically with viruses to impact risk of cancers such as oral cavity SCC.^80,172–175 Similarly, immunosuppression could lead to the reappearance of latent HPV infection, affecting risk of cutaneous SCC or cervical cancer.^176,177

Additional studies are needed to confirm the role of oncogenic viruses in cancer risk after allogeneic HCT. Studies of cancer occurrence and cancer risk factors in other immunosuppressed populations, particularly solid organ transplant recipients and individuals with human immunodeficiency virus infection and acquired immune deficiency syndrome (HIV/AIDS), support the idea that immunosuppression contributes to cancer risk in part through loss of immune control of oncogenic viruses.^93,178 PTLD is the most common virus-related cancer in allogeneic HCT recipients, and is caused by EBV.¹⁷⁹ EBV also has been associated with gastric adenocarcinoma after allogeneic HCT.¹⁸⁰ Other virus-related cancers are rare, but observed elevations in risk for cervical cancer (HPV) and liver cancer (HCV, hepatitis B virus [HBV]) support the likely importance of oncogenic viruses in cancer risk after allogeneic HCT.^24,39,43,181 A single case reports suggests the possibility of donor T-cells acquiring human T-lymphotropic virus (HTLV1).¹⁸² Polyomaviruses also represent an emerging group of viruses requiring study because of their potential to cause epithelial malignancy.¹⁸³ The recently described Merkel cell skin cancer occurring in immune deficient states is driven by the Merkel cell polyomavirus, which is closely related to BK and JC viruses, both of which reactivate after HCT.¹⁸⁴

Research Gaps and Recommendations

• Research is needed to investigate whether established cancer risk factors in de novo malignancies exhibit similar associations for post-HCT SN, and whether established cancer risk factors interact with transplant-related risk factors.

• With advances in technology for evaluating inherited genetic variation, there is great opportunity to identify genetic risk factors for SN after allogeneic HCT. Studies should consider a range of genetic variation, from rare inherited cancer predisposition syndromes or other rare variants to common genetic variants.

• More work is also needed to understand the role telomeres play in post-HCT cancer development and the nature of donor-recipient interactions.

OUTCOMES

Although data from other studies of SN have suggested that survival may be poorer,^185–187 relatively little is known about the outcomes of SN occurring in HCT recipients. Post-HCT PTLD has been reported to have an aggressive, immunoblastic phenotype,¹⁸⁸ and can be a major contributor to death.¹³ Therapy-related MDS/AML is more likely to have adverse cytogenetic characteristics than sporadic MDS/AML,^82,189 and prognosis has been shown to be poorer even after controlling for cytogenetic abnormalities.¹⁹⁰ For solid tumors, outcomes after allogeneic HCT appear to vary by tumor type, with poorer prognosis compared with the general population for female reproductive, bone and joint, lower gastrointestinal tract, and central nervous system (CNS) tumors, but comparable outcomes for thyroid, testis, breast, oral cavity, and soft tissue tumors, as well as melanoma.¹⁹¹

Research Gaps and Recommendations

• Further research is needed to better understand the prognosis following SN in the HCT setting. Studies with systematic long-term follow-up should include comprehensive information on prior treatments, tumor subtype, stage and grade, tumor molecular characteristics, and the therapeutic approach selected for the transplant recipient. Comparisons with comparable survivors in the general population are essential.

• Studies are needed to identify optimal treatment approaches for SN after HCT. Treatment options may be limited by prior therapeutic exposures, including both pre-HCT treatments and conditioning regimens. Examples of these concerns include patients who have already received cumulative maximum anthracycline doses or radiotherapy to a particular region of the body. Optimal treatment approaches should explore most recent advances including pharmacological and cellular immunotherapy approaches (e.g., donor-derived EBV-specific lymphocytes to manage PTLD^192,193).

OPTIMAL SCREENING AND PREVENTATIVE PRACTICES

Screening

The development of screening guidelines has been the focus of several efforts by professional organizations.^194–196 The Children’s Oncology Group (COG) created long-term follow-up guidelines (www.survivorshipguidelines.org) for children, adolescent, and young adults who underwent HCT. ¹⁹⁷ Consensus-based recommendations for screening and prevention of solid cancers were adapted from those for the general population through collaboration of the CIBMTR Late Effects and Quality of Life Working Committee and the EBMT Complications and Quality of Life Working Party for children and adults.¹⁹⁸

Given the heterogeneity in the types of SN developing after HCT, cancer screening has never been adequately studied in this population to determine if it provides any benefit in terms of survival or cost effectiveness. The issue of when to initiate screening is also unclear, particularly for patients transplanted as children (the risk of SN is higher at younger ages compared to a relatively low risk in the general population), and older adults undergoing HCT (as the risk of malignancies increases with age in the general population).

Preventative practices

Several studies have described adherence to preventative health behaviors and recommended screening guidelines for SN in HCT survivors.^199–201 In a study of 1549 adult survivors of autologous HCT, those concerned about medical costs, non-white race, male gender, lower physical functioning, no chronic GHVD, longer time since HCT and lack of knowledge about recommended tests were factors associated with lower adherence to recommended preventive care. A cross-sectional study (N=662) reported that autologous HCT survivors were more likely than allogeneic to undergo screening for breast and cervical cancer but less likely to have undergone a skin examination in the previous year.²⁰⁰ The Bone Marrow Transplant Survivor Study examined health behaviors and cancer screening practices in HCT recipients who underwent HCT during 1976–1998.²⁰¹ Multivariate regression analysis revealed younger age (<35 years) at study participation and lower education (<college) to be significantly associated with high-risk health behaviors for SN. Female survivors were more likely to have had a screening mammogram when compared to gender-matched sibling controls.

Interventions

It is imperative that survivors and clinicians be educated on recommended preventative practices and screening guidelines for SN. Studies suggest that as few as 20% of providers talk to cancer survivors about health behaviors,²⁰² and only 10% of survivors report their physicians talking to them about smoking, exercise, and diet.²⁰³ Similar data are lacking for HCT survivors. Care of HCT survivors is complex, requires a multi-disciplinary approach, and may be expensive. The optimal health care delivery model for cost-effective education on preventative practices and screening for SN deserves greater study, with a particular focus on education and care delivery to socioeconomically-challenged patient populations.

Research Gaps and Recommendations

• Although the current consensus screening recommendations represent an important step forward, stronger evidence is needed to support their validity and cost-effectiveness in the HCT population. Additionally, the magnitude of risk reduction, optimal screening techniques and timing of initiation, and effect on outcomes are unknown. Future recommendations should account for host factors, underlying disease, therapeutic exposures, and donor source, as appropriate.

• As cancer prevention interventions are implemented in the general population, such as vaccination against HPV,^173,177 their efficacy in HCT survivors should be evaluated.

SUMMARY

To reduce the burden of SN and improve the long-term health of HCT survivors, it is critical to identify the magnitude of risk, risk factors and mechanisms of pathogenesis for SN and to determine optimal preventative, screening, and therapeutic approaches. Research in these areas is challenging because of the complex clinical history of many transplant survivors, the need for long-term follow-up, and diversity of timing, behavior, and origin of SN. With substantial changes in clinical transplantation practices, risks for patients receiving HCT in the current era are poorly understood. Hence, the need for establishing a robust and collaborative, multidisciplinary infrastructure for SN research has never been greater. Clinical data must be consistently and uniformly collected on patient, disease, and pre-, transplant, and post-HCT variables within the context of large-scale, long-term (>10 years) studies of HCT recipients undergoing systematic follow-up in conjunction with recommended screening practices. Resources, particularly financial and other funding mechanisms, designed to support banking of cryopreserved donor and recipient blood and marrow cells and SN tissues with detailed patient and clinical data, as well as resources that provide support for longitudinal biomarker studies will be critical to advancing our understanding of the pathogenesis of SN and identify effective approaches for prevention and management.

HIGH PRIORITY RECOMMENDATIONS.

• Conduct large-scale, long-term (>10 years), systematic follow-up of HCT recipients to better define the magnitude of risks for specific SN types, including both common and rarer tumors. The full spectrum of SN should be considered, including certain benign tumors (e.g., meningioma) as well as all types of cutaneous malignancies.

• Collect detailed patient and clinical data (e.g., primary disease, pre- and post-transplant therapies, donor source, GVHD type/severity/therapy, adherence to screening recommendations, exposure to key cancer risk factors such as tobacco) in a standardized format to investigate determinants of SN risks.

• Bank cryopreserved donor and recipient blood and marrow cells and SN tissues to enable interdisciplinary clinical and laboratory investigations of SN susceptibility and pathogenesis.