Changes in Pediatric Cancer Treatments Yield Reduced Late Mortality

Changes in Pediatric Cancer Treatments Yield Reduced Late Mortality

Efforts to reduce late effects of treatments for pediatric cancers have resulted in reductions in late mortality for children diagnosed more recently, according to new data from the Childhood Cancer Survivor Study (Abstract LBA2) presented during the Plenary Session on Sunday, May 31. Deaths because of second malignant neoplasms, cardiac toxicity, and other issues have decreased.

“In just half a century, we have witnessed a remarkable improvement in survival in childhood cancer,” said Gregory T. Armstrong, MD, of St. Jude Children’s Research Hospital, who presented the study. By 2020, there will be an estimated 500,000 survivors of childhood cancers; currently, one in 750 people in the United States is a survivor of a childhood malignancy.

However, these survivors have a substantially increased risk of late mortality compared with the general population. Dr. Armstrong said that by 30 years from diagnosis, 18% of childhood cancer survivors are deceased. Many of these individuals die in their 30s and 40s, “relatively young, by any standard,” Dr. Armstrong said.

Specific changes in the treatment of certain childhood cancers have been made with the aim of reducing late effects. For example, there have been changes to the use of cranial radiotherapy (RT) for the treatment of acute lymphoblastic leukemia (ALL), and dosage and usage of RT have decreased in the treatment of Hodgkin lymphoma. There have also been recent improvements in screening and early detection of late effects, treatment of those effects, and supportive care.

Fig. 1
 The Childhood Cancer Survivor Study is a retrospective cohort of children diagnosed with cancer from 1970-1999 at 31 participating institutions. Only patients who were alive 5 years after diagnosis were included. The current analysis included 34,033 survivors of pediatric cancer: 9,416 diagnosed in the 1970s, 13,181 in the 1980s, and 11,486 in the 1990s.

There were a total of 3,958 deaths among this cohort. The 15-year cumulative mortality rate was 10.7% for patients diagnosed in the 1970s, 7.9% for the 1980s, and 5.8% for the 1990s (p < 0.001).

Cause-specific 15-year cumulative mortality rates were also significantly different. For survivors diagnosed in the 1970s, 7.1% died as a result of recurrence or progression of the primary cancer compared with 4.9% of patients diagnosed in the 1980s and 3.4% of patients diagnosed in the 1990s (p < 0.001).

Other health-related causes of mortality included late treatment-related effects. The rate of these deaths has also decreased in the three groups, from 3.1% (1970s) to 2.4% (1980s) to 1.9% (1990s; p < 0.001). Dr. Armstrong noted particular reductions in these deaths among patients diagnosed with ALL, Hodgkin lymphoma, and Wilms tumor.

Specifically, the 15-year mortality rate from subsequent neoplasms has decreased from 1.8% for those diagnosed from 1970-1974 to 1.0% for those diagnosed from 1990-1994 (p < 0.001). Dr. Armstrong said that this reflects the changes in the use of RT, in particular cranial RT for ALL—use of this therapy dropped from 85% in patients with ALL who were diagnosed in the 1970s to 19% in those diagnosed in the 1990s. RT exposure also decreased substantially for patients with Hodgkin lymphoma or Wilms tumor.

The decrease in cardiac mortality, from 0.5% in the 1970-1974 group to 0.1% in the 1990-1994 group (p < 0.001), reflects decreasing exposure to anthracyclines, which have known late cardiotoxic effects. For example, the earliest cohort of patients with Hodgkin lymphoma had an average anthracycline exposure of 295 mg/m2, which dropped to 193 mg/m2 in the 1990s cohort.

Dr. Michael P. Link
On multivariate analysis adjusting for age at diagnosis, sex, diagnosis, and follow-up time, the era of diagnosis had a protective effect. The 5-year interval of diagnosis had a relative risk (RR) for mortality from other health-related causes of 0.86 (95% CI [0.82, 0.89]). The RR for each 5-year interval for the development of a subsequent neoplasm was 0.83 (95% CI [0.78, 0.88]), for cardiac death it was 0.77 (95% CI [0.68, 0.86]), and for pulmonary death it was 0.77 (95% CI [0.66, 0.89]).

Dr. Armstrong’s group confirmed that changes to treatments are in fact responsible for the changes in late mortality rates. When adjusting the RR analyses for the specific therapy used for ALL, Hodgkin lymphoma, and Wilms tumor, the protective effect of treatment era was largely attenuated, suggesting that the therapy used was responsible for the protective effect of the era.

“The strategy of reducing the intensity of therapy to lower the occurrence of late effects, along with promotion of early detection and improved treatment of late effects has now translated to extend the lifespan of survivors,” Dr. Armstrong said.

Discussant Michael P. Link, MD, of Stanford University School of Medicine, said that this work reflects the need to focus on the “cost of cure”—what patients endure in order to achieve a cure. The evidence from this analysis, Dr. Link said, confirms that the efforts of the oncology community in that regard are indeed bearing fruit.

“The translation of modifications in therapy designed to reduce exposures into clinically significant reductions in all-cause late mortality is a gratifying validation of 3 decades of refining our therapies to accomplish the same number of cures while lowering the cost of cure,” Dr. Link said.

There are still challenges to continued reduction of these late effects of treatment, Dr. Link said. He noted that as the use of targeted therapies increases, there is no guarantee that these treatments will not introduce a new set of late toxicities. Understanding the heterogeneity of the patients, rather than just the tumor, could also help avoid unnecessary exposure to therapy.

Watch the session: Visit the ASCO Virtual Meeting website.