On November 22, 2025, Tatiana Schlossberg, a 35-year-old journalist and the daughter of Caroline Kennedy, announced that she was dying from a rare and particularly aggressive form of blood cancer called acute myeloid leukemia (AML). In a heart-wrenching essay in The New Yorker, she described having been diagnosed with the disease just a few hours after delivering a healthy baby girl and when she was seemingly in the prime of her life. Schlossberg wrote, “I had swum a mile in the pool the day before, nine months pregnant. I was not sick. I did not feel sick. I was actually one of the healthiest people I knew.”
About 60,000 Americans die of blood cancer each year, roughly equal to one death every nine minutes. Unlike breast or colon cancer, there are no general guidelines for screening for blood cancer, and the condition is often found incidentally, when routine lab work with a primary care doctor or other specialist shows abnormal levels of white blood cells, red blood cells, or platelets, prompting a bone marrow biopsy. By then, the cancer may be advanced and little can be done to save the patient’s life. While other cancers have seen improved survival rates in recent decades, there has been a two-thirds decline in breast cancer deaths since the 1990s, outcomes for many blood cancers like leukemia remain poor, with less than a third of patients with AML still alive five years after their diagnosis.
The link between clonal hematopoiesis and blood cancer
But in recent years, doctors have discovered a strong link between some blood cancers and the condition called clonal hematopoiesis (CH), a state in which a small number of blood stem cells reproduce prolifically, creating an army of identical cells that can progress to cancer. CH represents a precursor to certain types of blood cancers called myeloid leukemias, and detecting CH can provide a window of opportunity for closer monitoring and even possible preemptive treatment. Dr. Ross Levine, chief of the Molecular Cancer Medicine Service at Memorial Sloan Kettering Cancer Center, told me that, in a similar way that mammograms and colonoscopies have impacted survival from breast and colon cancers even more than advances in medical therapies, “identifying people with CH before overt leukemia emerges is our best chance to improve outcomes for patients in the long term.” Still, myeloid leukemias are rare and the majority of people with CH do not progress to cancer, meaning that screening is unlikely to be one size fits all. Instead, if we are to go looking for CH, we will need to do so using methods that are far more nuanced than is typical in the cancer screening world.
Every one of the trillions of blood cells contained within our bodies begins its life as a hematopoietic stem cell, each of which produces an equal proportion of our total blood volume. Stem cells, like all of our cells, contain long strings of deoxyribonucleic acid (DNA) that, over time, are prone to mutations. In most cases, these genetic changes do not lead to any advantage or disadvantage for the cell. But every so often a mutation can make the cell “fitter” than its neighbors, think Darwin’s natural selection, albeit with cells rather than members of a species competing for survival. When this happens, the stem cell can far exceed its normal capacity for reproduction, and clones of the cell may end up constituting a quarter, half, or nearly all of the mature blood cells in our bodies. This situation describes CH, a condition that can be detected through simple, non-invasive genetic sequencing of the blood.
Detecting and understanding clonal hematopoiesis
In 2014, scientists discovered that mutations in just two genes account for more than half of all CH cases, resulting in a 10-fold increased risk of developing blood cancer. The findings raised alarm at large medical centers like Memorial Sloan Kettering, where CH was accidentally being discovered in many patients already diagnosed with other forms of cancer. “Patients were asking ‘what do I need to do with this information?'” Levine told me. In 2018, he spearheaded the launch of a clinic dedicated to the care of patients found to have CH, the first of its kind in the nation. Now, there are about 30 such clinics spread out across the U.S. and many more around the world.
In following these patients over time, doctors have learned that most forms of CH are low risk and require little in the way of monitoring. Certain variants of the condition, however, substantially increase the risk of blood cancer, in some cases up to 90 times that of a person without CH, and their detection may provide the chance to catch malignancy before it becomes established and difficult to treat. Knowing this, doctors have wondered: Instead of waiting to find CH by happenstance, is it possible to expertly screen for the condition amongst large groups of people and find those who are at highest risk of developing a blood cancer?
The controversy and promise of cancer screening
Screening for many non-blood cancers is a routine part of general preventive care in the United States. The U.S. Preventive Services Task Force (USPSTF), a panel of national experts in evidence-based medicine, recommends mammograms for women starting at age 40 and colonoscopies for all adults beginning at age 45 in the general population. These guidelines are widely followed by doctors, though not without the knowledge that, while such screening has the potential to catch cancer early, the risks of over-diagnosis, false positives, and unnecessary treatments still exist. Certain types of cancer screening are particularly fraught with controversy. A blood test called prostate-specific antigen (PSA) can be used to look for prostate cancer, but studies suggest that as many as half of all prostate cancers found through PSA screening are slow-growing tumors that never would have shortened the patient’s life. The USPSTF therefore recommends that “the decision to undergo PSA-based screening for prostate cancer should be an individual one” made jointly between a patient and their doctor.
Where CH screening should fit into the pantheon of cancer screening tests is not entirely clear. Unlike a mammogram, which requires an X-ray, or a colonoscopy, which is an invasive procedure, screening for CH requires nothing more than a blood test. Next generation sequencing (NGS), the DNA decoding technology used to detect CH, is widely available and affordable in the U.S. (the cost is about $250). It is true that breast, colon, and prostate cancers are much more common than blood cancers like AML, which is diagnosed in fewer than five in 100,000 people each year, but blood cancer can be acutely fatal, putting a premium on its early detection.
Predicting leukemia through large-scale screening
Dr. George Vassiliou, co-lead of the Hematological Malignancies Program at the University of Cambridge in the United Kingdom (UK), told me that he and colleagues have been developing sophisticated methods to detect high-risk CH in the general population. Vassiliou’s group has demonstrated that AML can be predicted years before its clinical onset by identifying high-risk CH and specific mutation patterns in healthy individuals. His research has established that most AML cases arise from pre-leukemic clones detectable in the blood long before diagnosis and that, using very detailed genetic testing, doctors can tell the difference between harmless age-related blood cell mutations and early warning signs that someone may be developing AML.
Recent work from Vassiliou and his colleagues has expanded this approach to large groups of patients, such as those in the UK Biobank, a research program that tracks the health, genetics, and lifestyle information of about 500,000 volunteers in the UK. Vassiliou’s group is now developing a machine-learning algorithm that can use simple data from a complete blood count, one of the most widely performed tests in primary care, to identify patients who should be screened for CH. “By getting data that is very cheap and widely available,” he told me, “we might be able to catch up to 25 percent of high-risk patients, which is much better than we are doing now.”
Developing preemptive treatments for blood cancer
On a practical level, most doctors agree that a screening test is only as good as the intervention that can be offered based on its results. When we identify a genetic mutation linked to an increased risk of a breast cancer, such as a breast cancer gene (BRCA) mutation, then it is possible to perform surgery to remove the breast tissue before a cancer takes root. But, in a patient with high-risk CH, you cannot just “cut out” blood cells at risk for cancer.
Dr. Kelly Bolton, an oncologist at Washington University School of Medicine in St. Louis, is hoping to change this situation, at least for a portion of patients who all share a history of low blood counts, CH, and a mutation in the gene called isocitrate dehydrogenase 1 (IDH1), which increases the risk of AML by about 30 times that of patients without the mutation. By using ivosidenib, an inhibitor of the IDH1 enzyme, Bolton has been able to normalize blood counts and reduce the size of the CH clones in about 20 patients in her trial. The study, and others like it, seek to replicate cancer prevention treatment strategies that already exist for other malignancies. For instance, the use of tamoxifen, an anti-hormonal agent, has been shown to halve the risk of breast cancer in women who are at high risk of developing the disease, and aspirin can reduce colon cancer risk by about 40 percent in patients with Lynch syndrome, a genetic condition tightly linked to the development of the disease.
Still, some doctors explain that we are far from any sort of widespread screening system for CH that will do more good than harm. Amy DeZern, director of the Bone Marrow Failure and Myelodysplastic Syndromes Program at Johns Hopkins Hospital, fears that screening too broadly for CH runs the risk of consigning large numbers of people to cancer purgatory, where they will endlessly worry about the onset of a disease that may never come. She also told me it is premature to conclude that treating high-risk CH can definitively prevent the onset of a future leukemia. “We have no proof that any of these interventions change anything, and we must remember that every medication has side effects,” she told me.
The challenge of silent cancers
But in speaking with many oncologists for this story, what emerged was a shared feeling that we should be doing better for all patients. The current approach to blood cancers like AML, waiting until disease declares itself, is no longer good enough. Tatiana Schlossberg understood this with devastating clarity. In telling her story, she brought rare attention to a cancer that often unfolds in silence. On December 30, 2025, Schlossberg died. Her story should not be remembered only as tragedy, but as a challenge. If AML begins earlier, in forms we can now detect, then the question is not whether we can do better but, instead, when we will do better.
Jason Liebowitz is a rheumatologist.
