By LaChelle Weeks, MD, PhD
What is clonal hematopoiesis and why is it a clinical significance? Going back to around 2014, folks in our lab fully characterized clonal hematopoiesis of indeterminate potential, or CHIP. And this is a precursor to cancers like leukemia or classic syndromes—blood cancers that are very hard to treat once they're identified. And so this really opened up a new potential for early interception and early prevention in acute leukemias, which is something we haven't had before.
We have this in the solid malignancies, of course, with early detection of colon polyps to prevent colorectal cancer, for example. And so what we did, was we know that we can detect clonal hematopoiesis, or CHIP, in about 15 to 20% of people who are over the age of 65. And this frequency of detection of CHIP is much higher in patients who have undergone chemotherapy or radiation for other malignancies, so other solid malignancies.
But we also know that there's a wide range of variation, of risk, of progression. So just like every single colon polyp won't progress to colorectal cancer, everyone with CHIP is not going to develop acute leukemia ( myeloproliferative neoplasms). And so what we really did was sit down and say what are the actual factors that influences one's increased risk of progression and how can we put them into a model that's really easy to use at the bedside with our patients so that we can inform them if they're high risk factors for progression or low risk factors for progression?
Currently the standard of care for precursors of acute leukemias is just to monitor. Going back five years ago, we didn't have the technology really to detect these precursors, such as CHIP or CHIP or CCUS (CHIP/CCUS) . In our patients and the standard of care still simply is to just monitor folks, watch their blood counts and make sure that we're catching any signs of early progression.
We don't right now have any interventions that have been proven to prevent transformation. Of CHIP to acute leukemia (myeloproliferative neoplasms). And so therefore, it is not recommended that individuals are screened for these entities. Most of the patients that I see this has been identified on other sequencing panels.
So they get sequencing for their solid cancer, and they find a DNMT3A mutation or a mutation that could be CHIP. And then they're referred to me in a hematology clinic to evaluate and talk about their risk for progression. And so what this risk model does is it actually helps us to understand who are the patients, what is the population of people who are the highest risk for progression, because those are the people in which we'll target our early intervention clinical trials.
Those are the people who are more likely to derive some sort of risk benefit from some sort of intervention experimental interventions that. Think of using in the future. And we want to spare those individuals who are lowest risk, which is the vast majority of individuals, about 90%, from any sort of early intervention.
And we also want to spare them from the anxiety and the emotional distress of thinking. They have this precursor that's going to develop into leukemia.
So we use a cohort of around 500,000 individuals from the UK Biobank. And those individuals had exome sequencing performed, and we analyzed their exomes and made calls of whether or not they had clonal hematopoiesis, or CHIP. We ended up with about 30,000 people who had clonal hematopoiesis.
And we took those individuals and we split them into two different cohorts, or sub cohorts. One was a derivation cohort, and the other was for validation. In our derivation cohort, we then looked to see what are the features that are most predictive of progression or probability of developing myeloid malignancy (hematologic malignancies) over a period of around 10 years.
And then we identified around eight features using a combination of recursive partition analytics as well as multivariable Cox regression. And then we took the weights of the statistical weights of those features and combined them algorithmically into a model or calculator that we can use to actually determine one's absolute risk of progression.
Extremely prevalent is clonal hematopoiesis (CH), a precursor to myeloid (cells) malignancies (hematologic malignancy). Formally established CH subtypes, classified by somatic mutations in genes linked with myeloid neoplasia (MN) at a variable allele fraction (VAF) 0.02, include clonal hematopoiesis of indeterminate potential (CHIP) in the absence of cytopenia and clonal cytopenia of unknown importance (CCUS).
A recursive partitioning analysis identified groups of CHIP/CCUS cases with 10-year probability of incident MN ranging from 0.0078 to 0.65, indicating substantial variation in MN risk. Genetic characteristics (single DNMT3A mutation, high risk mutations, > 1 mutation, and VAF 0.2), patient age, CCUS versus CHIP, and high red blood cell indices were identified as partitioning variables (red cell distribution width and mean corpuscular volume).
While the UKB is a population cohort, CHIP/CCUS patients examined at hematology clinics may have a higher chance of getting MN. Consequently, we examined the CHRS in two independent cohorts of CHIP/CCUS patients evaluated with clinical NGS in hematology clinics at tertiary referral hospitals.
246 incident MN episodes were recorded during a median follow-up period of 11.50 years. A recursive partitioning analysis identified groups of CHIP/CCUS cases with 10-year probability of incident MN ranging from 0.0078 to 0.65, indicating substantial variation in MN risk.
The CHRS provides a straightforward and reliable predictive approach for CH, stratifying CHIP/CCUS patients at the population and patient levels.
Read and Share the Article Here: https://oncologytube.com/v/41487
So I think the most salient feature of our research is that most clonal, hematopoiesis, or most CHIP, actually is very low risk for progression. So using our risk model, which we're calling the clonal hemopoiesis risk score, or CHRS, we have observed that in a population of individuals with clonal hematosis, 90% of them are low risk, meaning they have a less than 1% chance of progressing to myeloid malignancy.
In a 10-year observation period on the contrast of minority of individuals, around 1% are high-risk, meaning their risk for progression to a myeloid malignancy is over 50%. And so the big thing, the big takeaway from this is that high risk population is really the population of people that we should be targeting for any sort of excessive surveillance strategies or intervention strategies.
And the low risk population of individuals is really quite low risk for developing a myeloid malignancy. And we can now tell that to our patients, which will put a lot of people at ease.
We derive this risk model in a population cohort. So these are individuals who we incidentally discovered that they had clonal hematopoiesis. So one of the biggest questions is, can you use this model? Can you apply it to a population of typical hematology patients? With the idea that someone who is referred to a hematologist is more likely, or the pre-test probability of them progressing to a malignancy is much higher because they already have a symptomatic anemia or a symptomatic other cytopenia. And so we validated our model actually in two separate or two independent clinical cohorts. One from the Dana-Farber Brigham Cancer Center, and the other from the University of Pavia.
And in both of those settings, we saw that our risk model performed very well. Had excellent discriminating capability able to identify all of the people essentially who progress as high risk and the individuals who did not progress were predominantly in our low and intermediate risk categories.
I think the biggest takeaway here is that we have made significant strides in oncology in our ability to do next generation sequencing. And as our technology improves, we will be able to detect. Signs of pre-cancer earlier and earlier. And I think that what our study does is it provides a bit of perspective that though we can detect even TP53 mutations in individuals in their blood cells that we, all of those mutations are not very high risk, and that we can temper the recommendations and even the information that we provide to our patients about their risk of progression based on data and the observations that we've made in our study. And I think that. It does take a little bit of time; these CHIP mutations were first identified in 2014, as I mentioned, and now it's 2022, and we're just really starting to have clinical tools that we can use to stratify risk. But that time we've learned quite a bit. About the biology of clonal hematopoiesis. And there's absolutely more to come in the future.
Dr. Weeks is an Assistant Professor of Medicine at Harvard Medical School and a physician-scientist in the adult leukemia department at the Dana-Farber Cancer Institute. Dr. Weeks got her MD and PhD from the School of Medicine at Case Western Reserve University. She completed her internal medicine residency at Brigham and Women's Hospital and her fellowship in hematology/oncology at the Dana-Farber/Mass General Brigham Cancer Care Program. The clinic of Dr. Week specializes in the treatment of patients with precursors of myeloid malignancies (MDS myelodysplastic syndromes), such as clonal hematopoiesis of indeterminate potential (CHIP) and clonal cytopenia of uncertain importance (CCUS). Her study focuses on identifying characteristics of CHIP and CCUS that influence the evolution of myeloid illnesses such myelodysplastic syndromes (MDS) and acute leukemia (a hematological malignancy).
ASH Publications - Prediction of Risk for Myeloid Malignancy in Clonal Hematopoiesis. ASH Publications, November 15, 2022