Commentary (Marcus): Long-Term Survival of Children with Brain Tumors

For all the vast literature concerning survival after treatment for cancer, few articles provide long-term follow-up. At best, most provide 5-year survival statistics, the majority being actuarial, which means that many of the patients included do not even have 5-year follow-up. A few articles give 10-year data and call these long-term. But 5 or even 10 years is not a very long time for a child or young adult hoping for a normal life span, or for clinicians wishing to be able to predict a patient's prognosis after treatment.

What are the final odds of a patient's dying of cancer? What if the patient is cured--what can be expected to happen to the patient 10, 20, or 30 years after treatment? In an effort to answer these questions, Dr. Jenkin reviewed the 30-year follow-up data of children irradiated for brain tumors at the Princess Margaret Hospital, providing a unique perspective.

In considering his data, be aware that this article includes only patients who received radiotherapy as all or part of their treatment; it does not include those who received only surgery or chemotherapy or both. Therefore, survival usually cannot be predicted accurately according to histology, particularly for low-grade histologies, because an unknown number of patients were not reviewed. Regardless, as follow-up information for the results of one treatment modality, the data are quite useful.

The most disappointing aspect of these data is that the overall survival curve never appears to plateau; it declines to 38% at 20 years and 30% at 30 years. With longer follow-up, secondary malignant neoplasms become more and more important. With the maximum of 30 years of follow-up, secondary tumors developed in 31 patients, for an actuarial rate of 13% at 20 years. This appears to be quite high, but it is important to note that of the 546 deaths in the study population, 514 were caused by "disease", as opposed to only 12 deaths by secondary neoplasms.

The Major Problem

Even in the 20- to 30-year range of follow-up, the number of patients in whom relapse occurred is approximately equal to the number of patients in whom secondary neoplasms developed. It is clear that tumor relapse is still the major problem after treatment, since large numbers of patients do not survive to be at risk for secondary tumors. In addition, although the cumulative risk analysis may indicate an actuarial probability of 13% at 20 years and even higher with further follow-up, some statisticians suggest that such analyses may overestimate the true incidence of infrequent events [1].

Furthermore, before all the secondary neoplasms are attributed to irradiation, it is important to remember that increasing evidence indicates that children with one tumor have a higher tendency to develop a second one, with or without radiation as part of their therapy.

A Case in Point

For example, in 1981, a 6-month-old girl with an extremely large medulloblastoma was brought to the University of Florida. She was treated with eight cycles of MOPP chemotherapy and went into remission. Her mother refused radiotherapy, and the patient had follow-up only. In 1995, a malignant fibrous histiocytoma developed in the left frontal lobe, apparently arising from the brain parenchyma. Had this patient been irradiated (as part of the usual treatment for a medulloblastoma), the second tumor would have fit all the criteria for a radiation-induced sarcoma, although obviously it was not. This emphasizes the fact that many children with tumors have an underlying predisposition to develop malignancies, even if no specific genetic defect has been determined.

In summary, these data show that vast improvements are needed to enhance the overall survival rates of many children with brain tumors. Although we work under the assumption that progress is being made, it is discouraging that the survival rates from 1975 to date do not appear to be superior to those from the early part of the study.

References:

1. Caplan RJ, Pajak RF, Cox JD: Analysis of the probability and risk of cause-specific failure. Int J Radiat Oncol Biol Phys 29(5):1183-1186, 1994.