(Guest Lecture): September 2023 - The Path to New Insights

Thanks so much for having me. So I'm Katherine. I'm an assistant professor in the Division of Population Sciences. I'm a cancer epidemiologist by training who works closely with the clinical colleagues in the CPAP program. A lot of my work is on the prevention side, so early detection of myeloma at its precursor conditions. And the title of my talk is The Path to New Insights, the Importance of Observational Research. And so part of my goal is to talk to you about some of the observational studies that we're it's important to do patient-powered and patient-engaged research like our P-crowd studies that Dr. Gabriel talked about, like contributing information to the Health Tree Foundation like Jenny had presented. What we're doing with that data, I'm going to go a little bit more slowly in terms of some of the studies than Dr. Gabriel tends to do to just highlight some examples of how we use that information. And so the first study is one that Dr. Gabriel already mentioned. It's called the P-crowd study. And this has been really fundamental in building our research program in myeloma prevention for the last almost decade. And so P-crowd is this patient-powered crowdsourcing initiative whereby we enroll patients with myeloma precursor conditions or really a precursor condition to any blood cancer. We engage patients nationwide through an online portal. So patients can go to pcrowdstudy.org or enroll at P-crowd, find the study, sign up for the portal and then when they see their regular physician for their standard myeloma workup, we send them a blood collection kit so they can collect additional research samples in conjunction with that clinical appointment so there's no extra sticks with a needle. They send us their medical information, if and when they have it, and we can assist with that. And then they complete some questionnaires, things like how they're thinking, how they're feeling, things about their lifestyle habits. And the overarching goal is to collect as much information as we can about patients so that that can lead to new insights and how to prevent multiple myeloma at its earliest stages and help people feel their best. And to date, as Dr. Gribble mentioned, we have over 3,000 patients that are enrolled in P-crowd that we follow over time collecting information. This information has contributed to a lot of really remarkable research discoveries that are coming out every day. And this slide just provides a brief snapshot of some of these. For example, we've done a number of studies looking at the impact of COVID-19 on patients with multiple myeloma and its precursor conditions, which helped inform how to monitor patients, how a vaccination schedule. Irene Goebbriel's lab, as she mentioned, also does a lot of work on sequencing blood and bone marrow samples from patients. And she does a lot of work to understand how the cells in the bone marrow are talking to each other and what that means in terms of who might be more likely to progress to multiple myeloma who might have more aggressive disease, as well as provide insights into new treatments, what might work for some patients. We also are very interested in understanding how to better predict which patients have more aggressive disease and need treatment right now and which patients we might be able to offer some reassurance to, have a more indolent condition. And so one project in particular that's been really central in terms of helping us better predict risk is a study called the Pangea Initiative. And this was a large international collaboration that was led out of Dana-Farber that tried to collate as much patient information as we could. And so we included as much information as we had on patients with myeloma precursor conditions at Dana-Farber, but we also reached out to our clinical colleagues in Germany, Greece, the Czech Republic, and the UK. And the idea is to pool as much information as we can about patients that are as diverse as possible so that we can ask really fundamental big picture questions, including ones that clinicians and patients care most about, which is how can we improve upon how we're currently predicting a patient's personalized risk. And when we think about risk prediction, specifically in the precursor disease space, we often refer to a model called the 2220 criteria or 2220 model. This is the most commonly used risk stratification algorithm for patients with smoldering myeloma here in the US. And it's based on classifying individuals based on three features. This includes greater than 20% bone marrow plasma cells or cancer cells in the bone marrow, a M protein or M spike of greater than two grams per deciliter, and a serum-free light chain ratio of greater than 20. And so we use these three 2220 features to stratify patients into one of three categories. The first one is a low-risk group, and it's somebody that does not have any of these higher-risk features. And these are patients that we see when we follow them over time tend to have about a 6.2% probability of progressing in two years. The second group is one that we classify as intermediate risk, which is the orange graph. These individuals have one of these 2220 risk factors and have a probability of progressing to myeloma within two years of 17.9%. And the third category is individuals that we consider highest risk, which have two or more of these 2220 features. These individuals have about a 44% risk of progressing to myeloma in the next two years and are generally individuals who we consider for treatment right away. And so these risk stratification algorithms right now are a tool that we have to understand how to evaluate a patient, how to monitor a patient, as well as evaluate them for therapies. But there's a lot of reason to believe that we can improve upon these existing models. And that's because these models look at patients, a snapshot of a patient at one point in time. But we know that your biomarkers, your lab values are changing over time. And that rate of change can provide a lot of insight into what's going to happen in the future. And so why can't we use that information to enhance our risk prediction? The other point is one that Dr. Gobriel mentioned, that when we do, you know, we're relying on these dichotomous variables of 22 and 20 to stratify patients. But a lot of times, clinically, it's more complicated than that. Irene mentioned a bone marrow might be patchy and it might be reported as a percentage. So what do we do with that range of information in terms of this risk classification? The other one is that it does rely on a bone marrow to appropriately risk classify individuals. Many patients don't have it, don't want to get it. And when they do, they don't want to get it very often, which we can't blame. So what would it look like if we just threw out the bone marrow data, if we didn't use that to classify an individual's risk? And then finally, we tend to think of MGUS and smoldering myeloma when we're thinking about risk classification as completely separate entities. But we know that many patients with MGUS look very similar to smoldering myeloma, and it's really a continuum of the same condition. So why are we thinking of them separately? And so we tried to leverage all of this patient information to create a better model. And we built a model using serially clinical data, so data that was collected every three to six months on patients over time. And we developed a model using data that we had from the P. crowd study at Dana-Farber. But then it was really important to us to go to other cohorts, these international cohorts to validate the model to make sure that it's not just generalizable to the unique group that we're seeing at our center in the U.S. But it really can be generalizable to most patients. And so we ended up identifying a core group of standard lab values that a patient gets at every clinic visit that seemed to enhance our ability to predict risk. And this included an individual's serum-free light chain ratio, M protein concentration, creatinine, hemoglobin levels, intuitively age also matters. And we ran the model with and without the addition of bone marrow data to say, can it perform just as well without that bone marrow information? And what we're able to find, these are Kaplan-Meier curves which show the progression-free survival. And essentially what you're looking at is the lines that are flat near the top, the black curve are patients that do best, and the blue that slant downwards are the patients that do worse. What we're able to see is that these are able to separate patients based on their probability of progression. And when we compare models with bone marrow on top and without bone marrow on the bottom, they do just as well. And the other interesting thing to note is that as we move left to right on the slide, as we contribute more and more clinical information, so update information with each clinic visit, the models get better at stratifying patients. The other important feature is that when we compare this new model, what we currently use in the clinic to predict risk, our model does better at predicting outcomes for patients. Does that when we look at kind of fancy statistical tests to compare models called the C-statistic. But it also does this when we look at how we're defining risk among the snapshot of patients who we know progressed to myeloma over follow-up. Our model does a much better job at accurately classifying or capturing those that are truly high risk from the existing models. And so Irene briefly made a nod to this calculator. We thought that this information, this ability to develop this personalized risk score for precursor patients was really valuable. And so it's something under the leadership of Irene Goebbriel that we made available to patients and clinicians at pangeamodels.org. So anyone can go to this website, enter their personal information. The more time points you have, the better. So you can go back in history and enter multiple visit dates to develop their own personalized risk score. And so I'd encourage everyone to check out this and share it with patients you might know. Let's see. Oh, I have a different talk. So the next study that I'm going to talk about is the PROMIS study. And this was a nationwide screening study for patients with at risk, at elevated risk for developing a blood cancer within their lifetime. And the premise of the PROMIS study is that we can learn more about multiple myeloma and how to prevent it if we study it from its earliest stages. And so as Dr. Goebbriel mentioned, we're specifically focusing on in the PROMIS study of enrolling individuals who are over the age of 40 who meet one of our two criteria. This includes individuals who self-identify as being black or African-American, as well as those with a strong family history of blood cancer or blood cancer precursor condition like MGUS and smoldering multiple myeloma. And so how does the PROMIS study work? So patients that are interested can visit our website and sign up for the study online. They take a very brief eligibility survey. And then they sign a study consent. And once they're consented, we ship them a blood kit, which they can take to any one of our contracted labs. We have a contract with Quest Labs, which has a number of locations throughout the US. They schedule that blood draw appointment. We analyze their data. And we provide them with results. If they do screen positive for a monoclonal protein or monoclonal gammopathy, they are connected with a clinician who can provide education, resources, and support. We ensure that they get the appropriate follow-up care that a patient needs. And we follow this cohort over time. Among all the patients in the cohort, we perform traditional screening tests for multiple myeloma such as the serum protein electrophoresis test, which is very common. S-PEP, you might hear it referred to in the clinic. This is a standard myeloma lab that's been around since the 1960s. It does a pretty good job of detecting M proteins. But we also use this novel, highly sensitive mass spec-based approach, which is much more sensitive and allows us the opportunity to detect monoclonal proteins in patients even earlier, at its earlier stages. And we are particularly interested in using a mass spec-based approach because it is gaining enthusiasm as a tool in the clinic for monitoring patients with myeloma, for signs of minimal residual disease. But it's also starting to be offered at different clinics throughout the country as a screening test. So what does this mean to patients? What does this mean about our understanding of the precursor disease space? And so we analyzed the first batch of patients that we screened as part of the PROMIS study. This was the first about 2,200 patients in the PROMIS study. And we supplemented it with a cohort that was from an identical population from one of our biobanks at Harvard, the Mass General Brigham Biobank. The majority of individuals met our same high-risk criteria of being black or African-American and having a family history of myeloma or another blood cancer. And what did we find? So when we looked at monoclonal protein concentrations using this very sensitive measure, we found a lot of patients that had M-spikes below the current level of detection of serum protein electrophoresis. And so we used serial dilution testing to identify that the lower limit of detection of an S-PEP is 0.2 grams per deciliter. That's what we currently think of as M-GUS. Anything below that, we don't know much about. And so to distinguish an entity that we know from one that we don't know much about, we're calling every monoclonal protein that's below that limit of detection of S-PEP an entity that we know about as monoclonal gammopathy of indeterminate potential to distinguish it from M-GUS. And so when we looked at the prevalence data from this cohort, we identified that if we use an S-PEP to look at the prevalence of monoclonal gammopathy or M-GUS in a general population over 40, we expect the prevalence to be about 3%. But in this high-risk group, so those that are black or African-American or have a family history, we see double the prevalence. So we see about 6%, which is consistent with what we thought from other large cohort studies like the Olmstead County, the Minnesota study that Dr. Gobiel referenced. When we use mass spectrometry, even using that same cut point for what we think of as M-GUS, we actually see double that. We see in this high-risk cohort, we're getting about 13% of individuals over 40 have M-GUS as defined by mass spectrometry. And when we look at the distribution by ages, we see that when we're looking at mass spec defined M-GUS, the prevalence continues to increase with age, which you see on the left, which is what we understand. But when we look at any monoclonal gammopathy, so when we incorporate these very low-concentration monoclonal gammopathy, it's a huge portion of the population. It's a huge portion of over adults. 40% of adults over 50 have some low-level monoclonal gammopathy. And so this leads to a lot of really interesting questions in terms of next steps. If many people have a low-level monoclonal gammopathy, what is driving some people to progress, to what we think of as M-GUS and smoldering, and many don't? How stable are these early monoclonal gammopathies? Is this something that's going to disappear over time? Is this something that should we even be screening for? Irene mentioned one patient who had this very tiny monoclonal gammopathy who thought that we should not be worrying, but we actually should follow those patients. So who are the ones that we follow? And who do we not need to? Cancer screening saves lives, but cancer screening can often lead to anxiety and overdiagnosis. And so that's something that we need to be mindful of, particularly when we're using these really sensitive tests. One thing you'll hear from Dr. O'Donnell about later today is the question that most patients ask us. Is there anything in my lifestyle that I can do if I have myeloma to improve my outcomes or if I have precursor myeloma to improve outcomes related to these very low-level, early early M-GUS? Does lifestyle matter in that context? We're doing a number of clinical trials to start to address those questions of are there lifestyle modifications that can help patients? And then a question is should we be looking for these early M-GUS stages in the clinical setting? Right now in the PROMIS, this is purely in the research setting to start to answer these questions about what is reasonable, what's in a patient's best interest. And so these are some of the fundamental questions that we're asking. So to conclude, I hope that I've given you a glimpse of why contributing to research studies, particularly these patient-powered research studies like PROMIS, PCROWD, contributing data to Health Tree is so important. We are continuing to leverage that data to enhance our ability to predict risk, to enhance our ability to predict how patients will do on different treatment regimens. We're seeing that there's a much broader continuum of monoclinical mopidies than we previously thought and it impacts a much larger percentage of the population. And so we are of the conviction that studying M-GUS and myeloma at its early stages can provide new opportunities for prevention. And so just one more plug for these patient-powered research opportunities, we have the PCROWD and PROMIS study, which are enrolling patients nationwide. PCROWD is for patients with a known precursor condition. And the PROMIS study is for patients that might have a family member that's diagnosed with multiple myeloma or its precursor condition. So definitely check those out and thank you so much for your time. Thank you. Thank you.