Using bilaterality and positive family history as a surrogate for the presence of an inherited germline RB1 mutation, we found that bilateral Rb survivors with a presumed inherited germline mutation had a borderline significant 37% increased risk of an SC compared with survivors who had a presumed de novo germline mutation. This is the first report, to our knowledge, to estimate the risks for SCs by Rb mutation classification (inherited or de novo) taking treatment and other factors into account. As expected, SIRs compared with the general population were elevated for three of the four Rb survivor groups (bilateral with and without family history and unilateral with family history). We noted that the SIR for all SCs combined for bilateral survivors with a family history was significantly greater than the risk for bilateral survivors without a family history of Rb. Restricting the analysis to an internal comparison of SC risk in bilateral survivors allowed us to evaluate the influence of family history in the presence of other risk factors that have been linked to SCs, including radiotherapy, chemotherapy, young age at Rb diagnosis, and attained age, and to simultaneously control for genetic susceptibility.
Among the three most common SCs occurring after bilateral Rb (bone and soft tissue cancer and melanoma), family history of Rb was consistently significantly associated with risk for only melanoma in both the SIR and the multivariate analyses. Melanoma likely accounted for some of the significant excess risk of SCs in bilateral survivors with a family history of Rb in the SIR analysis. Treatment for Rb exerted a stronger effect than family history for the risks of bone and soft tissue sarcomas (ie, radiation for soft tissue sarcoma and radiation and chemotherapy for bone). An increased incidence of bone cancer after Rb has been previously been associated in a dose-dependent manner with chemotherapy, in addition to a dose-related increased risk with radiotherapy,19,20
whereas a radiation dose-related increased risk of soft tissue sarcomas after Rb has been observed.10
In this cohort, approximately 70% of the bone and soft tissue sarcomas occurred in the radiation field.10,17
Not unexpectedly, melanoma risk was significantly related to older attained age and earlier calendar year of Rb diagnosis, because melanoma risk usually starts to increase at age 20 years in the general population.21
Two of the 23 survivors reported multiple melanomas, but family history of melanoma was not specifically queried. The site distribution of the melanomas was consistent with the distribution in the general population.21
Ultraviolet exposure is a strong risk factor for melanoma; however, we did not have sufficient data to evaluate this exposure. Additionally, the cumulative incidence for melanoma was related to both family history of Rb and time since Rb diagnosis. Taken together, these findings indicate that having an inherited germline mutation is more likely than a de novo germline mutation to predispose to melanoma. An explanatory mechanism is not obvious but may be related to other shared genetic changes.
In contrast, the cumulative incidence for bone and soft tissue cancer did not differ by family history; however, the cumulative risk for soft tissue sarcomas increased three-fold from 20 to 50 years after Rb, consistent with a previous report of increased risk for leiomyosarcomas between ages 30 and 50 years in this cohort.17
The cumulative risk of bone cancer did not differ appreciably between 20 and 50 years after Rb because the majority of bone cancers usually occur by age 29.9
The cumulative risk for bone cancer at 20 years in this cohort (7% for no family history and 8% for family history) agrees with a report by Hawkins et al19
of a cumulative risk of 7.2% at 20 years among 3-year survivors of heritable Rb.
Several other cancers also occurred in excess in the bilateral survivors, but only the SIR for pineoblastoma was greater in bilateral survivors with a family history. However, the small number of cases precluded any meaningful analysis. On the basis of the higher incidence of pineoblastomas from a meta-analysis, additional pineoblastomas may have occurred in this cohort and been missed because they were attributed to metastatic Rb.22
Not surprisingly, the risks for SCs differed among survivors with unilateral Rb by family history. Although we had only 36 unilateral survivors with a family history of Rb who developed three subsequent cancers, it was clear that the risk for a SC was approximately five times greater compared with unilateral survivors without a family history. Despite this increased risk, we did not combine these unilateral with bilateral survivors, because those with unilateral Rb were phenotypically different and their germline mutations may represent mosaicism or mutations with lower penetrance. We do not have the data to adequately address this issue. Interestingly, the proportion of mosaics has been reported to differ only slightly among bilateral survivors (5.5%) and unilateral survivors (3.8%).5
A weakness of the present study is that we did not confirm the presence of germline mutations, but relied on reports of family history from medical records. For a proportion of bilateral and unilateral survivors, family history was unknown, and it is possible that some of the bilateral and unilateral survivors were misclassified and actually had a family history of Rb. The proportion of survivors with a positive family history is consistent with some population-based studies23–25
and other clinical or register-based series26,27
; however, the generalizability of these results is limited to other hospital-based series. Although there are survivors older than 50 years of age, the cumulative incidence at 50 years is based on sparse data.
These data provide new clinical information for survivors, their parents, and their health care providers on the risk of SCs according to family history of Rb and presumed classification of germline mutation, which has not been available previously. The risk for melanoma seems to be higher for bilateral survivors with a family history of Rb, or an inherited germline mutation, compared with survivors with a de novo mutation in this large cohort of long-term survivors. All bilateral survivors, especially those with a family history of Rb, and their affected family members should be alert to the risk of melanoma, especially that posed by excessive sun exposure. Genotyping bilateral survivors with a melanoma should further elucidate the genetic link between these two cancers. Unilateral survivors with a family history of Rb should be screened for SCs according to the same guidelines as bilateral survivors.