A harmonization of 'big data' in examining long-term outcomes with contemporary treatment for pediatric and adult Hodgkin lymphoma patients is yielding a new approach to inform future clinical trial design and enhance clinical decision-making. Results are being shared as part of a poster presentation at the virtual American Society for Clinical Oncology Annual Meeting taking place this week.
Investigator Andrew M. Evens, DO, MSc, FACP, associate director for clinical services and director of the Lymphoma Program at Rutgers Cancer Institute of New Jersey and medical director of oncology services at RWJ Barnabas Health, shares more about the findings and the development of a new international consortium known as HoLISTIC that contributed to the basis of the work.
Why is this an important topic to study?
Although Hodgkin lymphoma (HL) has good rates of overall disease control, its treatment comes at considerable human cost, including a significantly increased risk of treatment-related late effects (e.g., secondary malignant neoplasms and cardiovascular disease), compromised health-related quality of life, and the potential loss of young lives. In fact, HL has the second highest "cost per death" or lost productivity cost of any cancer (Hanly P et al. Int J Cancer 2015).
Due in part to the above factors, substantial debate remains on optimal treatment for individual HL patients in the modern era, including type and amount of chemotherapy, use of radiation therapy, use of interim positron emission tomography (PET) to guide response-adapted therapy, and integration of novel therapeutics (e.g., antibody-based, immunotherapy). Helping clinicians and patients assess the individualized treatment tradeoffs of acute and long-term outcomes at the time of initial diagnosis is vitally needed.
Why form a new international collaboration in Hodgkin lymphoma?
Our overall goal is to improve clinical decision-making for pediatric and adult cancer patients, given expanding treatment options, but imperfect future knowledge about long-term outcomes with contemporary treatment (i.e., 20 to 30 plus years after therapy).
To accomplish this, we are assembling the best available evidence, including detailed information from prominent HL clinical trials, "real world" HL registries, and established HL survivorship cohorts.
Dr. Susan Parsons (from Tufts Medical Center in Boston, Massachusetts) and I spearheaded the formation of an international consortium, HoLISTIC (Hodgkin Lymphoma International STudy for Individual Care), currently consisting of 55 worldwide HL pediatric and adult oncology experts, decision science experts, statisticians, epidemiologists, and patient foundations and advocates.
We are harmonizing individual patient data from 16 prominent, contemporary clinical trials of newly diagnosed HL and six large institutional/regional HL registries, the latter that are highly enriched with long-term follow-up and late effects data.
How did you approach the data and what did you find?
In this acute analysis, we performed successful linkage of the harmonized HoLISTIC dataset of individual patient data on HL patients younger than 30 years old with simulated estimates of late effects from the large St. Jude LIFE pediatric HL cohort (Bhakta N, et al. Lancet 2017), based on cumulative treatment exposures.
Patient-level data on cumulative chemotherapy doses were extracted from the harmonized HoLISTIC database for 982 HL patients (mean age 19 years) from five clinical trials treated with chemotherapy only and survived at least five years progression-free.
A microsimulation model was set up, by incorporating mortality and incidence of late effects as functions of diagnosis age, sex, race, treatment exposures and attained age estimated from 5,522 adult 10-year survivors of childhood cancers in the St. Jude LIFE cohort.
Despite differences in acute progression-free survival, reflecting in part the variation in risk/stage status, the projected long-term outcomes were similar.
What are the implications of these findings?
This innovative project capitalizes on an unparalleled international, multi-disciplinary team (pediatric and adult) dedicated to harmonizing big data together from clinical trials, "real world" registries, and survivorship cohorts to improve decision making, which is greatly needed.
The model will allow for personalization and incorporation of evolving new therapies and knowledge to emerge and will help providers identify the options best aligned with patient values and preferences.
This project's results will be significant as they are expected to enhance individualized patient care, (economic) value, and ultimately outcomes. In addition, this approach will inform the development of similar methods, treatment models and paradigms for other cancers.
What are next steps for this work and for other research with your colleagues through the HoLISTIC consortium?
We will continue to harmonize individual patient data from 16 large, contemporary clinical trials, conducted in North America and Europe across the age continuum and disease spectrum (i.e., pediatrics and adults; and early and advanced stages) from more than 12,000 HL patients.
We are augmenting this with data of over 6,000 HL patients captured in six large pediatric and adult institutional/regional registries, survivorship cohorts, and a "real world" large community-based practice. All of these are enriched with long-term follow-up and detailed information on late effects. We employed principles of data science, establishing the HoLISTIC data repository that facilitates incorporation of future data.
We have also assembled "real world" data on more than 600 HL patients in a large community oncology system (i.e., Kaiser Permanente Southern California (KPSC)) containing detailed information about upfront and salvage treatment and also created a unified data dictionary of the KPSC data with comparable data from the NCI-funded Molecular Epidemiology Resource from the Mayo Clinic and University of Iowa that also have enriched data on late effects and treatment exposures after disease recurrence.
Additionally, with funding from the Leukemia & Lymphoma Society, we have implemented a stakeholder engagement study with patients and providers, utilizing web-based surveys; in-depth interviews; and adapted, conjoint analysis.
Altogether, using data already acquired through data sharing and international collaboration across pediatric and adult specialties, we will continue to create robust, nimble, and innovative decision models to guide HL patients and their families, alongside their providers, to provide granular, objective data to enhance and optimize the difficult decisions that affect acute and long-term outcomes.
The models will also allow for incorporation of updated information as new therapies and knowledge emerge. Finally, the decision model will be converted to a web-based platform to test and evaluate among HL providers and patients at the point of care in order to help providers identify the options best aligned with patient values and preferences.