Chicago (FRONTLINE MEDICAL NEWS) – Survivors of childhood cancers face several later risks from treatment, and investigators presented studies evaluating risks in three specific areas – secondary neoplasms, premature menopause, and neurocognitive function – at the annual meeting of the American Society of Clinical Oncology.

Discussant Paul Nathan, M.D., of The Hospital for Sick Children in Toronto, said “the whole purpose” of research in this area “is to start to understand the predictors and modifiers of late effects” and then to design risk assessment tools and interventions to reduce long-term toxicity. These interventions include modification of chemotherapy and radiation doses, protective strategies, and disease risk stratification to adjust intensity of therapies.

Other strategies are to use behavioral interventions directed at improving compliance with follow-up to detect problems earlier and the use of real-time monitoring, such as with smart phones or fitness trackers. He said one limitation of this sort of research and implementing interventions to reduce late toxicities is that “you need time to document long-term outcomes.” So tracking newer therapies, such as proton beam radiation, small molecule drugs, and immunotherapy, is “going to take time, perhaps decades, before you understand their impact on patients.”

Risk of secondary neoplasms reduced

Risk-stratifying of disease “has allowed us to make attempts to minimize late effects by modifying therapy over time in certain subgroups of lower-risk patients,” said Dr. Lucie Turcotte of the University of Minnesota in Minneapolis.

To study the effects of these changes, she determined the risk of certain subsequent malignant or benign neoplasms over three periods of therapeutic exposure among 23,603 5-year survivors of childhood cancers diagnosed at less than 21 years of age from 1970 to 1999, drawing from the cohort of the Childhood Cancer Survivor Study (CCSS). The CCSS represents about 20% of childhood cancer survivors in the United States for the study period.

Over the decades of 1970-1979, 1980-1989, and 1990-1999, the use of any radiation went from 77% to 58% to 41%, respectively. Cranial radiation for acute lymphoblastic leukemia (ALL) decreased from 85% to 19%, abdominal radiation for Wilms tumor from 78% to 43%, and chest radiotherapy for Hodgkin lymphoma from 87% to 61%. The proportion of children receiving alkylating agents, anthracyclines, and epipodophyllotoxins went up, but the cumulative doses went down ( N Engl J Med. 2016 Mar 3;374(9):833-42 ).

Dr. Nathan said today, almost no child gets cranial radiation for ALL. “So we’ve slowly learned that our treatments are toxic, and we’ve certainly done what we can to change them.”

But have these changes made a difference? Dr. Turcotte found that survivors remain at increased risk of a secondary neoplasm, but the risk was lower for children treated in later time periods.

Dr. Nathan pointed to Dr. Turcotte’s data showing that the incidence of subsequent malignant neoplasms decreased from 1970 to 1999 by 7% for each 5-year era (15-year risk: 2.3% to 1.6%; P = .001; number needed to treat, NNT = 143). Similarly, non-melanoma skin cancer 15-year risk decreased from 0.7% to 0.1% (P less than .001; NNT = 167). The NNT’s are “certainly important, but these are not major differences over time,” Dr. Nathan said. Knowing the impact of newer, targeted therapeutic approaches will take some time.

Predicting risk of premature menopause

Also using the CCSS data, Dr. Jennifer Levine of Columbia University Medical Center, New York, N.Y., studied the prevalence of and risk factors for nonsurgical premature menopause (NSPM), defined as cessation of menses prior to age 40 years, as well as the effect on reproductive outcomes for survivors of childhood cancers.

Dr. Nathan said when a child is first diagnosed with cancer, seldom does the issue of fertility come up early in the discussion, “but when you treat young adults who are survivors, the number one thing they talk about often is fertility. And so doing a better job in predicting who is at risk for infertility is clearly a priority for survivorship research.”

He said the development of the cyclophosphamide equivalent dose (CED) by D.M Green et al. ( Pediatr Blood Cancer. 2014 Jan;61(1):53-67 ) has been very helpful for quantifying alkylating agent exposure to make comparisons between studies. The goal is to develop a risk assessment tool to be able to tell patients and families their fertility risk based on demographics, therapy, and biomarkers.

Being able to evaluate risk is critically important because for girls, oocyte or ovarian harvesting or even transvaginal ultrasound is highly invasive, and these procedures should be recommended only if their risk for infertility is very high.

Dr. Levine studied 2,930 female cancer survivors diagnosed at a median age of 6 years between 1979 and 1986 and a median age at follow-up of 34 years, who were compared with 1,399 healthy siblings. Of the survivor cohort, 110 developed NSPM at a median age of 32 years, and the prevalence of NSPM at age 40 years for the entire cohort was 9.1%, giving a relative risk of NSPM of 10.5 compared with siblings, who had a 0.9% NSPM prevalence at age 40.

She found that exposure to alkylating agents and older age at diagnosis put childhood cancer survivors at increased risk of NSPM, which was associated with lower rates of pregnancy and live births after age 31 years. The greatest risk of NSPM occurred if the cyclophosphamide equivalent dose was greater than 6000 mg/m2 (odds ratio = 3.6 compared with no CED); if there had been any radiation to the ovaries (less than 5 Gy: OR = 4.0; 5 Gy or more: OR = 20.4); or if the age at diagnosis was greater than 14 years (OR = 2.3).

Women with NSPM, compared with survivors without NSPM, were less likely ever to be pregnant (OR = 0.41) or to have a live birth after age 30 (OR = 0.35). However, these outcomes were no different between the ages of 21 and 30. Dr. Levine said this information can assist clinicians in counseling their patients about the risk for early menopause and planning for alternative reproductive means, such as oocyte or embryo harvesting and preservation.

Neurocognitive functioning after treatment

Dr. Wei Liu of St. Jude Children’s Research Hospital, Memphis, Tenn., studied the neurocognitive function of long-term survivors of ALL.

Dr. Nathan called ALL “the paradigm for how we’ve sort of learned and adjusted how we treat patients based on late effects.” Early on, the disease was treated with craniospinal radiation and intrathecal chemotherapy, and while patients survived, it became obvious that they suffered neurocognitive and endocrine problems, growth abnormalities, and secondary malignancies. These findings forced a reevaluatuon of treatments, leading to elimination of spinal radiation, reduction of cranial radiation dose, intensification of systemic therapy, including methotrexate, and risk stratification allowing modification of therapies.

Survival was sustained, but long-term outcomes were still based on children treated with radiation. So long-term cognitive consequences in the more modern era of therapy were unknown. Only recently have adult cohorts become available who were treated in the chemotherapy-only era.

Dr. Liu studied 159 ALL survivors who had been treated with chemotherapy alone at a mean age of 9.2 years. The follow-up was at a median of 7.6 years off therapy at a mean age of 13.7 years. At the end of the chemotherapy protocol, patients completed tests of sustained attention, and parents rated survivors’ behavior on standard scales.

She found that for these childhood cancer survivors, sustained attention and behavior functioning at the end of chemotherapy predicted long-term attention and processing speed outcomes. Only exposure to chemotherapy, and not end-of-therapy function, predicted that survivors would have poor executive function of fluency and flexibility at long-term follow up.

Dr. Nathan praised the investigators for their foresight to collect data on the methotrexate area under the curve, number of triple intrathecal therapies (cytarabine, methotrexate, and hydrocortisone), and neurocognitive functioning at the end of chemotherapy. “What’s clear is that chemotherapy alone can lead to neurocognitive late effects,” he said. “But what’s also important is that not all late effects can be predicted by end of therapy assessments.” These late effects appear to evolve over time, so ongoing assessments are needed.

Dr. Turcotte, Dr. Liu, Dr. Levine, and Dr. Nathan each reported no financial disclosures.