- © 2001 by American Society of Clinical Oncology
Solid Cancers After Bone Marrow Transplantation
- Smita Bhatia,
- Andrew D. Louie,
- Ravi Bhatia,
- Margaret R. O’Donnell,
- Henry Fung,
- Ashwin Kashyap,
- Amrita Krishnan,
- Arturo Molina,
- Auayporn Nademanee,
- Joyce C. Niland,
- Pablo A. Parker,
- David S. Snyder,
- Ricardo Spielberger,
- Anthony Stein and
- Stephen J. Forman
- From the Divisions of Hematology and Bone Marrow Transplantation, Pediatric Oncology, Biostatistics, and Pathology, City of Hope National Medical Center, Duarte, CA.
- Address reprint requests to Smita Bhatia, MD, MPH, Division of Pediatric Oncology, City of Hope National Medical Center, 1500 E. Duarte Road, Duarte, CA 91010-3000; email sbhatia{at}smtplink .coh.org.
Abstract
PURPOSE: To evaluate the incidence and associated risk factors of solid cancers after bone marrow transplantation (BMT).
PATIENTS AND METHODS: We analyzed 2,129 patients who had undergone BMT for hematologic malignancies at the City of Hope National Medical Center between 1976 and 1998. A retrospective cohort and nested case-control study design were used to evaluate the role of pretransplantation therapeutic exposures and transplant conditioning regimens.
RESULTS: Twenty-nine patients developed solid cancers after BMT, which represents a two-fold increase in risk compared with a comparable normal population. The estimated cumulative probability (± SE) for development of a solid cancer was 6.1% ± 1.6% at 10 years. The risk was significantly elevated for liver cancer (standardized incidence ratio [SIR], 27.7; 95% confidence interval [CI], 1.9 to 57.3), cancer of the oral cavity (SIR, 17.4; 95% CI, 6.3 to 34.1), and cervical cancer (SIR, 13.3; 95% CI, 3.5 to 29.6). Each of the two patients with liver cancer had a history of chronic hepatitis C infection. All six patients with squamous cell carcinoma of the skin had chronic graft-versus-host disease. The risk was significantly higher for survivors who were younger than 34 years of age at time of BMT (SIR, 5.3; 95% CI, 2.7 to 8.6). Cancers of the thyroid gland, liver, and oral cavity occurred primarily among patients who received total-body irradiation.
CONCLUSION: The risk of radiation-associated solid tumor development after BMT is likely to increase with longer follow-up. This underscores the importance of close monitoring of patients who undergo BMT.
DURING THE PAST 25 years, the number of patients who undergo bone marrow transplantation (BMT) for a variety of malignant and nonmalignant disorders has increased steadily. Improvement in survival after BMT has resulted in a need to assess issues related to long-term complications, such as malignant neoplasms. Second neoplasms are a known complication of chemotherapy and irradiation treatment for patients with Hodgkin’s disease and non-Hodgkin’s lymphoma,1-3 and they have also been reported after BMT. Previous studies have demonstrated a low but significant risk of neoplasms for various malignant and nonmalignant disorders among recipients of BMT.4-11
It is unclear whether the subsequent malignancy is related to pretransplantation chemotherapy and radiotherapy, whether it is the result of transplant conditioning regimens, or whether it is a cumulative effect of all of these exposures. We analyzed 2,129 patients who underwent BMT for various hematologic malignancies at the City of Hope National Medical Center between 1976 and 1998. Our goal was to evaluate whether pretransplantation therapeutic exposures and/or transplant conditioning regimens were associated with an increase in the risk of subsequent nonhematopoietic malignancies.
PATIENTS AND METHODS
We used both a retrospective cohort and a nested case-control study design to evaluate the role of pretransplantation therapeutic exposures and transplant conditioning regimens in the development of subsequent nonhematopoietic malignancies.
Cohort Analysis
Between 1976 and 1998, 2,129 patients underwent BMT at the City of Hope National Medical Center for hematologic malignancies. Data were obtained from the BMT database housed in the Department of Biostatistics at City of Hope. These included the unique patient number, initial diagnosis, sex, date of birth, date of transplantation, source of stem cells (bone marrow, peripheral stem cells, or bone marrow and peripheral stem cells), and conditioning regimens. For patients who developed a subsequent nonhematopoietic malignancy or solid cancer, the date of diagnosis and morphology were recorded. If the patient had died, the date and cause of death also were recorded. Reports of pathologic findings in the post-BMT malignancies were reviewed. Patients who had not been seen within the last 18 months were identified, and efforts were made to assess their vital status and potential complications they might have developed in the interim period through contact with their primary physicians or directly with the patients. At the time data were abstracted, contact had been documented within the previous 2 years for approximately 87% of the patients.
Allogeneic BMT from HLA-matched or partially matched family member donors was performed in 1,157 patients; 213 patients received unrelated donor grafts matched for HLA-phenotype, and 759 patients received autologous marrow.
Conditioning regimens for patients with leukemia (acute and chronic myeloid leukemia and acute lymphoblastic leukemia) included total-body irradiation (TBI) with cyclophosphamide (Cy) alone (Cy/TBI) or etoposide (VP-16) alone (VP-16/TBI), a combination of TBI, VP-16, and Cy (Cy/TBI/VP-16), or busulphan and cyclophosphamide. Patients with Hodgkin’s disease or non-Hodgkin’s lymphoma received Cy/TBI/VP-16 or a combination of cyclophosphamide, VP-16, and carmustine or lomustine. Patients with severe aplastic anemia were conditioned with cyclophosphamide alone or a combination of cyclophosphamide with total lymphoid irradiation, with or without antithymocyte globulin.
To estimate the risk of new solid cancers after BMT, age- and sex-specific person-years of observation were compiled for the cohort. Cancer incidence rates (obtained from the Surveillance, Epidemiology, and End-Results registry12) were used to calculate the expected number of cancer cases. Standardized incidence ratios were calculated by obtaining a ratio of the observed and the expected number of cases. The 95% confidence intervals (CIs) were estimated using a method described by Vandenbroucke.13 The absolute risk, defined as the excess number of cancer cases per 10,000 patients per year, also was calculated. The time at risk for second neoplasms was computed from the date of BMT to the date of subsequent malignancy, the date of last contact, or the date of death according to which of these came first. Cumulative probabilities for development of subsequent nonhematopoietic malignancies over time were calculated using the Kaplan-Meier method.14 The log-rank test was used to compare the various subpopulations. Cox proportional hazards regression techniques were used to calculate relative risk estimates. Variables in the regression model included primary diagnosis, sex, age at transplantation, source of stem cells, and conditioning regimens with and without TBI. Age at transplantation (< 34 years v ≥ 34 years) was included as a categorical variable. This cohort provided us with the sampling frame from which to select controls for the case-control study.
Case-Control Analysis
For each of the subsequent malignancies, three controls were randomly selected from within the cohort according to the following matching criteria: primary disease, age at transplantation (± 2 years), type of transplant (allogeneic [related and unrelated] or autologous), and duration of follow-up (± 1 year). For the cases and controls so identified, the entire medical record was used to abstract clinical information important for assessment of the characteristics and treatment of the malignancy before the transplantation. On occasion, the medical record consisted of records from more than one hospital if portions of treatment were provided elsewhere. The following data were collected:
-
1. Disease characteristics. These data included diagnosis (histology), stage, and primary site.
-
2. Treatment before transplantation. Chemotherapy data included dates, protocols/regimens, chemotherapeutic agents and their dose schedules, and routes of administration. The total cumulative doses per square meter of body-surface area were calculated for each chemotherapeutic agent used before transplantation. The median dose for each agent was identified, and the therapeutic agents were entered as categorical variables, dichotomized as less than or greater than the median dose for that agent. Radiation therapy data included dates, total dose, field, fractions, dose per fraction, equipment, and a copy of the institutional radiation therapy summary report. Surgical and other procedure-related data included a summary of all surgical procedures, which was completed and attached to a copy of institutional surgical reports.
-
3. Priming regimens for stem-cell mobilization.
-
4. Source of stem cells.
-
5. Conditioning regimens used. These data included the use of conditioning radiation and/or chemotherapy.
Analyses of the case-control study measured the degree of association of antecedent factors (such as chemotherapy and radiation) with adverse outcome (subsequent malignancy) by estimation of odds ratios. Exact confidence limits of the odds ratios were calculated. The test for trend was used in the analysis of categorical data, ordered by degree of intensity and/or duration. Conditional logistic regression was used to investigate the simultaneous effects of several variables. The variables in the model included pretransplant exposures to the various chemotherapy agents, radiation, and the various conditioning regimens used for transplantation.
RESULTS
As of January 1999, 50% of the cohort of 2,129 patients who had undergone BMT at the City of Hope National Medical Center were alive at last contact. The median follow-up period was 3.3 years (range, 0.1 to 21.1 years). Among the 941 patients who survived for at least 1 year after transplantation, the median duration of follow-up was 3.6 years (range, 1.0 to 21.1 years). The median age at transplantation for the entire cohort was 33.9 years (range, 1.5 to 71.5 years), and the cohort had accrued 5,951 person-years of follow-up after BMT. The characteristics of the patient population are listed in Table 1.
Characteristics of the Patient Population
Among the 2,129 patients who underwent BMT, 29 new cases of invasive solid cancer were observed compared with 9.4 expected cases in the general population (ratio of observed to expected cases, 2.1; 95% CI, 1.3 to 3.0) ( Table 2). Among the 29 patients, nine were diagnosed with nonmelanoma skin cancers (three with squamous cell carcinoma and six with basal cell carcinoma of the skin); they were not included in the calculation of standardized incidence ratios. The remaining 20 patients with invasive second cancers included four patients with cancer of the cervix uteri, three patients with salivary gland tumors, and three patients with squamous cell carcinoma of the oral cavity. Two patients each had breast cancer, liver cancer, and cancer of the thyroid gland, and one patient each had astrocytoma, malignant fibrous histiocytoma, squamous cell carcinoma of the esophagus, and synovial sarcoma. The risks were significantly elevated for cancers of the liver (ratio of observed to expected cases, 27.7), cervix (13.3), and oral cavity and pharynx (17.4). The absolute risk for all cancers was 17.6 cases per 10,000 patients per year. The cumulative incidence rates (± SE) of new solid cancers 5 and 10 years after transplantation were 1.6% ± 0.5% and 6.1% ± 1.6%, respectively ( Fig 1). Cumulative incidence rates for individual cancer types are listed in Table 2.
Ratio of Observed and Expected Rates of Solid Tumors and Cumulative Incidence Rates by Site and Primary Diagnosis
Fig 1. Cumulative probability of solid cancers after BMT in 2,129 patients.
Solid Cancers by Type of Transplant
The cumulative incidence rate of new malignancies after BMT varied by type of transplant; the rates were 6.4% and 1.6% at 10 years for patients who had undergone allogeneic versus autologous transplantation, respectively. Among the 759 patients who underwent autologous BMT, only two patients developed a second solid malignancy (one cancer of the cervix and one breast cancer).
Solid Cancers by Primary Diagnosis
The risk of solid cancers varied according to the primary diagnosis (Table 2) and tended to be significantly higher in patients with a primary diagnosis of acute and chronic myeloid leukemia compared with the general population. This relationship also was reflected when the risk was evaluated as absolute excess risk. A significant relationship was found between patient age at the time of transplantation and the risk of cancer ( Table 3). The risk for patients who were younger than 34 years of age at the time of BMT was 5.3 times that of the general population (standardized incidence ratio [SIR], 5.3; 95% CI, 2.7 to 8.6). Conversely, the risk was not significantly elevated for patients who were older than 34 years of age (SIR, 1.1; 95% CI, 0.5 to 2.0). The inverse association between risk and age persisted when the risk was evaluated as absolute excess risk. Significantly elevated risks also were found for patients with acute myeloid leukemia, acute lymphoblastic leukemia, and chronic myeloid leukemia who had undergone BMT before the age of 34 years (Table 3). There seemed to be a significant relationship between sex and specific types of cancer, such as thyroid cancer and cancer of the oral cavity, as discussed below. The risk of solid cancers also varied according to the use of TBI ( Table 4), and risk tended to be significantly higher than in the general population for patients who had received TBI as part of the myeloablative therapy and who subsequently developed secondary liver, thyroid, and oral cavity cancers. Several unusual cancers were observed in this population, and they are described below.
Standardized Incidence Ratio of Solid Malignancies According to Site, Primary Diagnosis, and Age at BMT
Ratio of Observed to Expected Cases of New Solid Cancers According to Site and TBI
Cervical Cancer
Four patients in this cohort were diagnosed with cervical cancer. The primary diagnoses were chronic myeloid leukemia (n = 2), acute myeloid leukemia (n = 1), and severe aplastic anemia (n = 1). The four cervical cancers developed a median of 3.3 years after BMT (range, 1.6 to 9.7 years) among women who had undergone transplantation at a median age of 41.3 years (range, 27 to 55.6 years). There was a 13-fold increased risk for development of cervical cancers versus a comparative normal population (SIR, 13.3; 95% CI, 3.5 to 29.6). Older age at transplantation ( > 34 years; SIR, 18.5; 95% CI, 3.5 to 45.9) was associated with an increased risk of cervical cancer.
Thyroid Cancer
Thyroid cancer developed in two patients who had undergone allogeneic transplantation for acute myeloid leukemia. Their ages at transplantation were 7.9 and 33.9 years, and the second malignancy developed 7.5 and 18.0 years after BMT, respectively. Both cases of thyroid cancer in this cohort developed among women, and the female cohort had a 13-fold increased risk for development of thyroid cancer after BMT compared with the general population (SIR, 13.1; 95% CI, 1.3 to 38.2). Younger age at transplantation (< 34 years; SIR, 11.1; 95% CI, 1.1 to 31.9) and TBI treatment (SIR, 10.0; 95% CI, 1.0 to 30.2) were the other risk factors for development of secondary thyroid cancers.
Liver Cancer
Two patients developed liver cancer 6.7 and 14.9 years after BMT. The cohort had a 27.7-fold increased risk compared with the general population. The median age at transplantation was 26.2 years, and the younger cohort ( < 34 years at transplantation) was at a significantly increased risk for development of liver cancer (SIR, 200; 95% CI, 18.9 to 573.2). Both patients with liver cancer had a previous history of chronic hepatitis C infection and hepatic cirrhosis. In addition, in both patients TBI was associated with an increased risk for development of liver cancer (SIR, 20.0; 95% CI, 1.9 to 57.3).
Oral Cavity Cancer
Six patients developed cancer of the oral cavity. These included three patients with salivary gland tumors. The median age at BMT was 25.7 years, and the younger cohort had a 53.3-fold increased risk compared with the general population (SIR, 53.3; 95% CI, 13.9 to 118.4). Cancers of the oral cavity and pharynx occurred primarily among males, and the male cohort had a 25-fold increased risk of developing this cancer compared with the general population (SIR, 25.6; 95% CI, 8.1 to 53). Patients who received TBI had a 25-fold increased risk of developing cancer of the oral cavity (SIR, 25.0; 95% CI, 7.9 to 51.7). In addition, the three patients with squamous cell carcinoma of the oral cavity also had chronic graft-versus-host disease, as did the three patients with squamous cell carcinoma of the skin.
Case-Control Study
As listed in Table 5, multivariate analysis failed to reveal an association between pretransplant radiation or chemotherapy, conditioning chemotherapy or radiation therapy, presence of acute or chronic graft-versus-host disease, and the demographic variables examined and the risk for development of solid cancers after BMT.
Multivariate Analysis of Risk Factors Associated With Solid Cancers (case-control study)
DISCUSSION
Among the 2,129 patients who received BMT for hematologic malignancies at the City of Hope National Medical Center between 1976 and 1998, we found an estimated cumulative risk of a new solid cancer of 1.6% at 5 years and 6.1% at 10 years after transplantation. The 29 post-BMT malignancies represent a two-fold increased risk compared with a comparable normal population. The risk increased sharply over time, to 14.9% at 15 years, and risk was highest among patients who had undergone transplantation when they were younger than 34 years of age. Statistically significant increases in the risk were confined to cancers of the oral cavity and pharynx, cervix, and liver.
Previous studies have reported an increased risk of new malignancies after BMT, including therapy-related leukemia, lymphoma, lymphoproliferative disorders, and other nonhematologic cancers.4-11 Hematologic malignancies that occur after BMT develop relatively early in the posttransplant period, and they have been well described previously.4,15-21 Solid cancers have a longer latency period, however, and they are increasingly being described because of improved survival after BMT.4-6 Patients who have undergone BMT have a two- to three-fold increased risk for these cancers compared with the general population. It is speculated that the increased risk of solid cancers after BMT is related to pretransplantation conditioning with radiation, altered immune function in association with past infections such as human papillomavirus or hepatitis B or C, and previous treatment for the primary disease. We undertook this analysis to evaluate whether pretransplant therapeutic exposures or transplant conditioning regimens were associated with an increased risk of subsequent solid cancers.
In our study, the risk of thyroid cancers, cancer of the oral cavity and salivary gland, and liver cancer was elevated among patients who had received TBI as part of their myeloablative therapy, whereas it was not elevated among patients who did not receive TBI. An increased risk of cancers of the thyroid gland, oral cavity, and salivary glands in association with radiation has been described.22 In general, radiation-induced cancers have a long latency period, and the risk of such cancers is frequently high among patients who undergo irradiation at a young age.3,23 This was true for our cohort, in which patients with thyroid cancer, cancer of the oral cavity, and liver cancer underwent BMT at a significantly younger age than the median age for the cohort, and the median time for development of the solid tumors ranged from 7.6 years for cancer of the oral cavity to 12.7 years for thyroid cancer. Thyroid cancer has been reported in children exposed to radiation, as have salivary gland tumors, demonstrating a dose-response relationship to radiation.24-27
It has been reported that immune-suppressed patients have an increased risk of cancer at certain sites. Chronic graft-versus-host disease, with the attendant chronic inflammation, has long been suspected as a potential risk factor for the development of squamous cell cancers of the skin and the buccal cavity.28-31 In this study, all six patients with squamous cell cancer of the skin and buccal cavity had chronic graft-versus-host disease, as did the patient who developed squamous cell carcinoma of the esophagus.
In immune-suppressed patients, oncogenic viruses such as human papillomaviruses may contribute to squamous cell cancers of the skin and buccal mucosa after transplantation.32 The female survivors were at a significantly elevated risk for development of invasive squamous cell carcinoma of the cervix compared with the general population. It is speculated that past infection with human papillomavirus in association with immune deficiency may contribute to the increased risk of cervical cancer after BMT.33 The two patients who developed liver cancer after BMT were chronic carriers of hepatitis C, and the liver cancer developed in a setting of chronic hepatitis. There is a significant body of epidemiologic, clinical, and laboratory data that point to a role for hepatitis C virus infection in hepatocellular cancer, especially in immune-suppressed patients.34
An exhaustive attempt was made to collect information on all pretransplant exposures for patients with second malignancies and the identified matched controls. Analysis of exposure to chemotherapy or radiation therapy before transplantation failed to reveal any associations. However, the relatively small number of cases or the lack of heterogeneity among the exposures could have limited our ability to identify an association.
As the length of follow-up increases, an increasing number of radiation-associated solid tumors may emerge, as has been reported among long-term survivors of Hodgkin’s disease treated with conventional therapy.3,35 We conclude that patients undergoing BMT have an increased risk for development of solid cancers. The contribution of treatment received before or during transplantation remains unclear. The critical question of the actual additive contribution of the transplantation procedure, above the increased risk for second cancers borne by these patients as they enter the BMT process, remains unresolved. The risk of radiation-associated solid tumors probably will increase with longer follow-up, especially in association with certain chronic viral infections, such as hepatitis B and C and human papillomavirus. This underscores the importance of monitoring these patients on a continuous basis.
Acknowledgments
Supported in part by grants no. CA 30206 and CA 33572 from the National Cancer Institute, Bethesda, MD.
- Received May 2, 2000.
- Accepted September 7, 2000.
