A study led by Smita Bhatia, M.D. (left), and Noha Sharafeldin, M.D., Ph.D. (right), has found that genetic variations associated with cognitive decline after blood or marrow transplantation can identify patients at high risk for cognitive decline more accurately than current methods. Blood or marrow transplantation is a potentially lifesaving treatment for patients with leukemia, myeloma, lymphoma and other hematologic malignancies. But BMT also can bring long-lasting side-effects, including cognitive impairment that leaves patients unable to resume work or normal activities.
In a new study, the first of its kind, UAB researchers identified a combination of genetic factors associated with cognitive impairment after BMT. The researchers used the information to pinpoint patients at highest risk for deficits in a large patient cohort, significantly enhancing risk predictions compared with demographic or clinical characteristics alone. The findings were published in the Journal of Clinical Oncology Feb. 21.
More than a third suffer from ‘chemo brain’
Hematologic malignancies, or blood cancers, account for about 10% of all new cancer diagnoses, according to the Leukemia and Lymphoma Society. The society estimates that 179,000 people in the United States were diagnosed with a hematologic malignancy in 2019.
A 2018 study by the same research team found that patients who have allogeneic BMT, in which they receive cells transplanted from others, have a significantly higher risk of cognitive impairment. “We found that up to 36% of patients who had allogeneic transplants had cognitive issues up to three years after the transplant,” said Noha Sharafeldin, M.D., Ph.D., assistant professor in the UAB Division of Hematology & Oncology and first author of the 2018 paper and the new study. Sharafeldin is a researcher in UAB’s Institute for Cancer Outcomes and Survivorship, which is led by Smita Bhatia, M.D., who was awarded a $6.38 million grant from the National Cancer Institute in 2019 to study the long-term impacts of BMT on survivors’ health. Bhatia is senior author on both papers.
“Patients use the phrase ‘chemo brain,’” said Sharafeldin, who is also an associate scientist with the O’Neal Comprehensive Cancer Center at UAB. “It’s not dementia, but mild to moderate cognitive impairment. They feel foggy, become forgetful. They say they’re not up to keeping up with appointments and medications. It hinders their ability to function in the workplace.”
Understanding accelerated aging
The combination of cancer and high-intensity treatments such as BMT, which involves chemotherapy and radiation therapy, leads to accelerated aging in many patients, Bhatia explained. “One of the hallmarks of accelerated aging is premature occurrence of the chronic health conditions that are typically seen in older populations,” she said.
Games for brainsFor patients in the fog, the question is, “what intervention strategies can we offer?” Noha Sharafeldin said. "So far the evidence is pointing toward cognitive training — essentially, playing brain games.” Learn about her ongoing brain-game research in this story. |
Little is known about the relationship between genetics and cognitive ability in cancer patients, even though the link has been demonstrated in several studies outside the oncology realm, including in Alzheimer’s disease. “There is limited data in breast-cancer survivors implicating genes involved in DNA repair and oxidative stress,” Sharafeldin said. In the current study, Sharafeldin, Bhatia and their colleagues aimed to identify single nucleotide polymorphisms, known as SNPs, that could help to explain the reason some patients are affected by cognitive impairment while others are not. SNPs are the most common type of genetic variation at the population level that can help explain inter-individual variability in susceptibility to disease.
The researchers started with a list of “biologically plausible” SNPs, Sharafeldin said. They hypothesized that chemotherapy and/or radiation induce DNA damage and shortening of telomeres, structures on the end of chromosomes that are reduced as cells age. These changes could result in neurodegeneration and cognitive impairment, as could failures of DNA repair genes and genes involved in maintaining the blood-brain barrier, which pump toxic elements out of brain cells.
Machine learning
The researchers identified nearly 1,000 SNPs in 68 candidate genes among the 277 patients in their discovery cohort — all of whom had undergone allogeneic or autologous BMT between 2005 and 2011. Because these patients had contributed pre-BMT samples, the researchers were able to test their model by checking its predictions against the patients’ results on cognitive tests up to three years after treatment.
| “A better understanding of post-BMT health care needs could result in the deployment of targeted strategies that yield better quality of survival and reduced utilization of health care resources,” said Smita Bhatia. “That is what drives all of our efforts in the Institute for Cancer Outcomes and Survivorship.” |
Using machine-learning techniques, the researchers built a risk-prediction model featuring those SNP and gene-level signals that were most strongly correlated with cognitive impairment. These included SNPs on DNA repair genes, SNPs on genes linked to blood-brain barrier maintenance and genes linked to telomere homeostasis.
The researchers then tested their model by predicting outcomes in nearly 550 patients from the BMT Survivor Study, which is examining outcomes for patients who received BMT between 1974 and 2014. The study, which is supported by Bhatia’s $6.38 million NCI grant, includes information on self-reported post-BMT learning and memory problems as identified by the patients’ health care providers.
A clinical model, which took into account information such as a patient’s treatment history, fatigue levels and cognitive reserve (as measured by IQ), performed significantly better than a basic model that only used demographic information such as age, sex, education and income level. But a combined model, adding genetic information to the other two predictive groups, performed significantly better than the clinical model. The researchers measured model performance using the area under the curve (AUC) metric, in which 0.5 indicates that a model did no better than chance, while a 1 indicates perfect discrimination and prediction. “Values of 0.80 or greater are considered excellent,” Sharafeldin said. “The combined model with genetic markers in our study had an AUC of 0.89.” Removing the genetic variants dropped the model to an AUC value of 0.77.
The cost of SNP-testing technologies is declining steadily, Sharafeldin noted. “What we envision is a custom array with a few selected SNPs that can help us in the clinic in addition to the other information we already collect,” she said. “Incorporating SNPs in a risk-prediction model can guide risk stratification and enable a more informed clinical decision-making process.”
Identifying patients at highest risk for cognitive decline after BMT enables clinicians to consider other treatment options or to proactively begin interventions aimed at enhancing cognition. “A better understanding of post-BMT health care needs could result in the deployment of targeted strategies that yield better quality of survival and reduced utilization of health care resources,” Bhatia added. “That is what drives all of our efforts in the Institute for Cancer Outcomes and Survivorship.”
—
This research was supported in part by the Leukemia and Lymphoma Society. In addition to Sharafeldin and Bhatia, authors include Joshua Richman, M.D., Ph.D., Yanjun Chen, Purnima Singh, Ph.D., and Liton Francisco from UAB; Alysia Bosworth, Sunita Patel, Ph.D., F. Lennie Wong, Ph.D., and Stephen Forman, M.D., from City of Hope in Duarte, California; and Xuexia Wang, Ph.D., from the University of North Texas.
