Risk Stratification and Targeted Therapy, Abetted by Collaboration, Improve Outcomes for Children with Cancer

Get Permission

Pediatric cancer has been the poster child for what can be accomplished through collaboration, through the understanding of multidisciplinary care.

— Michael P. Link, MD

Outcomes for children with cancer have “improved over the course of the years incrementally, mostly not from the development of new drugs, because virtually all the drugs that we use now in leukemia were available in the 1970s. It is really through better understanding of the heterogeneity of the disease and therefore more appropriate risk stratification, giving more intensive therapy (and, occasionally, targeted therapy for patients at highest risk of relapse), and wherever possible, modifying therapy to reduce the side effects of therapy for patients who have a very good prognosis,” 2011–2012 ASCO President Michael P. Link, MD, told participants at the AACR/ASCO Presidential Symposium presented during the American Association for Cancer Research (AACR) Annual Meeting.1 The symposium was co-chaired by Dr. Link and 2011–2012 AACR President Judy E. Garber, MD, MPH.

“This is the path for improving the outcome in children with acute lymphoblastic leukemia (ALL)”—stratifying risk according to molecular classification and basing our therapeutic decisions on these classifications. “It is certainly true in neuroblastoma … and acute myeloid leukemia (AML) as well,” said Dr. Link, who is Professor of Pediatrics, Division of Pediatric Hematology/Oncology at Stanford University School of Medicine in Palo Alto, California. Many of the lessons learned from treating children with cancer can also be applied to treating adults, he added.

Progress in treating childhood cancers owes a lot to “a tremendous willingness to collaborate and to share,” Dr. Link acknowledged. “Pediatric cancer has been the poster child for what can be accomplished through collaboration, through the understanding of multidisciplinary care,” he said in an interview with The ASCO Post. “Collaboration is going to become more important than ever as tumor classifications become more precise in defining subsets of common cancers.”

‘One of the Great Success Stories’

“ALL is the most common malignancy in the pediatric population, and [its treatment] is really one of the great success stories in all of cancer therapies,” noted James R. Downing, MD, Scientific Director and Deputy Director of St. Jude Children’s Research Hospital in Memphis. “Today, if a child comes into the hospital with ALL, there is a greater than a 90% chance of survival,” Dr. Downing said.

“What is responsible for that? It really is the effective use of combination chemotherapy, figuring out how to use it, how to deliver intensive therapy, how to add maintenance therapy, and supportive care. A major influence, though, has been risk stratification into genetic subgroups and then, based on those, adjusting the intensity of therapy for those patients,” Dr. Downing explained.

“Over the past 5 years, using genome-wide approaches,” Dr. Downing said, “we have made a number of discoveries, as have a number of other laboratories around the world, that have provided new diagnostic and prognostic markers. What we showed was that 60% of ALL patients actually had genetic lesions in genes that regulate normal B-cell development.” These mutations appear with very high frequency in high-risk leukemias, and studies are being developed to look at how to therapeutically target them.

Pediatric Cancer Genome Project

In 2010, St. Jude Children’s Research Hospital initiated the Pediatric Cancer Genome Project as a joint project with Washington University School of Medicine in St. Louis. The goal was to sequence 600 diagnostic and 600 germline samples over a 3-year period, “to develop a foundational effort that would provide a database for all investigators working in pediatric or adult cancer,” Dr. Downing said.

“The frequency of mutations across the cancer types varies markedly,” Dr. Downing reported. “There are particular pediatric cancers, such as osteosarcoma, that have more mutations than seen in almost any adult cancer. There are other pediatric cancers, like infant acute lymphoblastic leukemia, that have exceedingly few mutations.” Data on these genomes can be explored at the project’s website: explore

“We are generating prognostic markers, … and there are therapeutic targets that are coming out of this,” Dr. ­Downing noted. “But there are a number of lessons that we need to keep in mind. First, whole-genome sequencing will be required to really identify all mutations. Clonal heterogeneity is an issue. We are going to need sensitive methods to detect the rare clones. Relapse often arises from that rare clone being selected in response to chemotherapy,” he said.

Even with whole-genome sequencing, much of the “large landscape of the genome remains uncharacterized,” Dr. Downing said. “There is going to be lots of work over the next several years to really look at those areas of the genome, to decipher exactly what mutations mean, and how they impact and transform the phenotype.”

Total Therapy Trials

To minimize morbidities and improve quality of life for children being treated for ALL, the Total Therapy XV protocol at St. Jude uses immunologic and molecular assays to measure the level of minimal residual disease after induction of remission to increase the precision of risk-directed therapy. “Final risk classifications are really based on the response to the treatment,” Ching-Hon Pui, MD, said. “We know the depth of remission affects the ultimate treatment outcome.” Dr. Pui is Chair of the Department of Oncology and Co-Leader of the Hematological Malignancies Program at St. Jude.

With the whole-genome analysis, 100% of the ALL population can now be classified into prognostically and therapeutically relevant subgroups, Dr. Pui said. “We found it is important to individualize the drug dosages based on pharmacokinetics, pharmacogenetics, and pharmacodynamics,” he noted.

Although prophylactic cranial irradiation had been standard treatment for high-risk leukemia, “up to 20% of patients will eventually develop radiation-induced second cancers, many brain tumors, after 20 years or more,” Dr. Pui said. “So in Total Therapy XV, we totally omitted the use of prophylactic cranial irradiation regardless of the presenting features.” Triple intrathecal chemotherapy with methotrexate, hydrocortisone, and cytarabine was used, with patients at high risk of CNS relapse receiving more intensified intrathecal therapy, Dr. Pui explained. The protocol also limited the use of cyclophosphamide because of its negative effect on fertility, and of anthracyline because of its link to cardiomyopathy, and omitted etoposide entirely because of safety concerns.

“Using this approach,” Dr. Pui said, “we are achieving 10-year event-free survival of 86% and 10-year survival of 91%,” but 2% of patients develop isolated CNS relapse. Following salvage therapy, patients with CNS relapse “have all been in second remission for 7 to 10 years. In all likelihood, they are cured. In the past 12 years at St. Jude, we have not lost a single patient from isolated CNS relapse although none of the patients received prophylatic cranial irradiation,” Dr. Pui said.

Under a new protocol—the Total Therapy Study XVI—triple intrathecal therapy is used upfront (and intensified for high-risk patients who can be clearly identified), because based on the results of Total Therapy XV, patients at high risk of CNS relapse can be identified, Dr. Pui said. “In the past 5 years, we have treated over 260 patients and we have not had a single patient with CNS relapse,” he added. “So we can push event-free survival to 94% today.”

Long-term Survival Issues

“One in 900 individuals is a childhood cancer survivor, but this comes at the price of a high morbidity,”  according to Smita Bhatia, MD, MPH, Director of the Center for Cancer Survivorship and of Outcomes Research at the City of Hope in Duarte, California. Data from survivors followed for 30 years show that “close to 80% are going to have long-term toxicity, and close to 40% of them have severe or life-threatening long-term treatment-related complications,” she said. These conditions include steroid-related osteonecrosis; radiation-related musculoskeletal deformities, lung cancer, breast cancer, brain tumors, and pulmonary dysfunction; and chemotherapy-related leukemia and cardiac complications.

In the 1980s, when these late events and complications became apparent, “we stopped throwing the kitchen sink at our patients and started tailoring therapy to risk factors,” Dr. Bhatia said. “In the 1990s, we started understanding the relationship of dose of radiation and chemotherapeutic agents to these late effects, and we started initiating efforts to track and educate our survivors,” she continued. The emphasis is now on the etiology and pathogenesis of these treatment-related adverse events, identifying patients at highest risk, and intervening to reduce morbidity and mortality.

Second Cancers

Second cancers can result from the combination of inefficient detoxification of genotoxic insults that can be caused by chemotherapy and inefficient repair of DNA damaged by these agents, Dr. Bhatia explained. A small proof-of-principle study found that “patients either repair their DNA damage or have apoptosis. If there is DNA repair, which is complete, you have recovery. If it is really bad DNA repair, you have apoptosis, but sometimes in between you develop therapy-related leukemia,” she said.

Another study looked at gene-expression changes in CD34-positive cells among patients who were undergoing autologous transplant for lymphoma. In patients who would eventually develop therapy-related leukemia, we saw changes in gene expression that are associated with the development of leukemia, but they could be identified as early as the time the peripheral blood stem cells were procured, Dr. Bhatia said. “Then we looked at the time of development of therapy-related leukemia and, again, … the changes in gene expression associated with the development of therapy-related leukemia were present in the bone marrow stem cells of those patients,” she explained.

“This speaks to the personalized therapy that Dr. Pui is talking about,” Dr. Bhatia noted, “because if we can discern that at the time of transplant that a group of patients is at a very high risk of therapy-related leukemia, instead of offering them autologous transplant, we can offer them allogeneic transplant or other methods of treatment.”

In a study of genomic instability, as measured by telomere length, “we found that patients who eventually developed therapy-related leukemia showed attrition in telomere lengths, as opposed to those who eventually did not develop therapy-related leukemia,” Dr. Bhatia reported.

Reducing Morbidity and Mortality

“We need to start working on interventions to reduce morbidity and mortality,” Dr. Bhatia said. “There is a clear dose-response relationship—even at low doses—between anthracycline and cardiomyopathy,” she noted, that “is begging for a targeted intervention for those at increased risk of cardiomyopathy.” One approach being tried is the use of beta-blockers to reduce the risk of anthracycline-related congestive heart failure.

“We know clearly that radiation for unrelated cancer increases the risk of breast cancer. Among girls who received radiation, by the time they are 40 years of age, 20% of them are going to develop breast cancer,” Dr. Bhatia said. “We have developed a low-dose tamoxifen trial that we hope will reduce the risk of radiation-related breast cancer.”

The risk of adverse outcomes with therapeutic exposures is modified by the patient’s age and gender, viral infections, lifestyle exposures, and genetic predisposition.

“These factors can be used to identify those who have the highest risk of developing adverse outcomes. We can modify our therapies, provide personalized therapies, or screen those at highest risk,” she concluded.

‘Boatload’ of Information

The future practice of oncology and truly personalized medicine will “not only have to consider the tumor genome, but will have to consider the host genome as well,” taking into account a “boatload” of information, Dr. Link said. “Sophisticated informatics will be crucial if we are going to be able to utilize all of this information to deliver personalized medicine” while continuing to take into account patient needs and desires, he noted.

There are also implications for clinical trials. “We need highly annotated tissues,” Dr. Link said. “We can identify very small subsets of patients within a cancer type that are better defined by sequencing, but these smaller subgroups make robust clinical trials much more difficult to accomplish. We need broader trials. In pediatrics, international collaboration is the rule in order to recruit sufficient numbers of patients. We need new regulatory considerations for licensing new agents. We need novel endpoints,” he added. “And we need incentives for developing agents that will be applicable for a diminished target population.” ■

Disclosure: Drs. Link, Bhatia, Downing, and Pui reported no potential conflicts of interest.


1. Pediatric Cancer in the Age of Genomics—Lessons from Our Children. American Association for Cancer Research Annual Meeting. AACR/ASCO Presidential Symposium. Presented April 2, 2012.