Emily Matson and Dr. Ray Merrill, Department of Health Science
Although cancer is often considered a disease of old age, it kills more children ages 1 to 14 years in the United States than any other cause, except unintentional injuries . Through advances in diagnosing and treating cancer, child survival of cancer has improved in recent years [1, 2]. However, more research is needed on the risk of subsequent cancer following an initial diagnosis of cancer, of the same or another anatomical site. Associations between childhood and secondary cancers have implications for patient follow-up, by providing health professionals with better information on which subsequent cancer sites are most likely to occur, what types of screening are recommended, and what symptoms to look for based on the childhood cancer survived.
Perhaps the single most comprehensive body of information on childhood cancer and risk of subsequent primary malignancies is the Childhood Cancer Survivor Study, which began in 1970 . In the United States, the definitive source of cancer incidence and survival data comes from the Surveillance Epidemiology and End Results (SEER) Program of the National Cancer Institute [5, 6]. SEER data are currently available from 1973 through 2010 . The purpose of the current study is to provide a comprehensive assessment of the risk of subsequent primary malignancies following an initial diagnosis of several childhood cancers, utilizing SEER data and the International Classification for Childhood Cancer (ICCC), site recode ICD-O-3/WHO 2008 [1, 9]. These results will then be compared with existing study results from the literature.
SEER*Stat is a statistical software product that provides a mechanism for SEER data analysis . The software can generate multiple primary standardized incidence ratios, which represent the observed divided by the expected number of cancer cases following a first primary cancer diagnosis. Included with the ratios are 95% confidence intervals, mean age at diagnosis, and mean age when the subsequent cancer occurred. Only malignant neoplasms were considered.
To compare the results with the literature, a literature review was performed. Peer-reviewed articles on second and subsequent malignant neoplasms in childhood cancer survivors (1993-2011) were identified from electronic databases – PubMed, Wiley Online Library, ScienceDirect, American Association for Cancer Research, and CA: A Cancer Journal for Clinicians. Note: The Childhood Cancer Survivor Study defines childhood and adolescent cancers as cases diagnosed before age 21 , while the SEER program includes cases diagnosed before age 20 [1, 10-11].
The observed/expected ratios of secondary neoplasms among childhood cancer survivors with confidence intervals for specific secondary cancer types are presented in Table 1. The table also includes the differences in age between primary and secondary diagnoses.
Primary retinoblastomas had the highest ratio of observed/expected secondary neoplasms (10.3). Meanwhile, thyroid carcinoma resulted in the lowest observed/expected ratio of subsequent cancer (1.6), although all cancer types reviewed had a significant increased risk of subsequent cancer. The overall average observed/expected ratio was 4.5.
By primary cancer type, a few of the highest observed/expected secondary neoplasms involved: intrahepatic bile duct (411.2) for Wilms’ tumor; hepatic flexure (265) for neuroblastoma; intrahepatic bile duct (362.1) for renal tumors; renal pelvis (96.1) for CNS neoplasms; and nose/nasal cavity/middle ear (325.5) for retinoblastoma. Retinoblastoma was the cancer type diagnosed earliest in life on average (at 1.6 years), while Hodgkin lymphoma had the latest average age at secondary diagnosis (34.9 years old). Time between diagnoses ranged from 10.4 years (malignant melanoma) to 19.1 years (Hodgkin lymphoma), with an average of 15.1 years.
Supporting references are also presented in Table 1. For example, in the case of primary lymphoid leukemia, increased risk for each of the secondary cancer sites identified was also shown in previous studies. However, some studies identified an increased risk for cancer that we did not find. These studies are summarized in Table 2. For example, three studies found an increased risk of melanoma following a diagnosis of lymphoid leukemia, while SEER data did not show this result.