High-Risk MDS: Classification, Bone Marrow, and Prognosis

High-risk myelodysplastic syndrome (MDS) is a bone marrow cancer in which abnormal blood-forming stem cells fail to produce healthy blood cells, leading to dangerously low blood counts and a substantial chance of progression to acute myeloid leukemia (AML). Patients classified as high-risk have a median survival measured in months rather than years, though that timeline varies dramatically depending on specific genetic features and available treatment. The distinction between high-risk and lower-risk MDS is not arbitrary; it rests on a scoring system built from bone marrow findings, chromosome abnormalities, and increasingly, individual gene mutations that reshape both prognosis and treatment decisions.

What MDS Actually Is

MDS is a group of blood cancers rooted in the bone marrow. The marrow still tries to make red cells, white cells, and platelets, but the production line is defective. Cells look abnormal under the microscope (a feature called dysplasia), and many die before they mature, a process called ineffective hematopoiesis. The result is one or more persistently low blood counts: anemia, low neutrophils that raise infection risk, or low platelets that impair clotting. MDS is considered a neoplastic myeloid disorder and is separated from AML by a blast threshold, traditionally set at 20 percent blasts in the bone marrow or blood.¹

That 20 percent cutoff has practical consequences. A patient with 19 percent blasts is labeled MDS; a patient with 21 percent has AML. The biology on either side of the line can be virtually identical, and researchers have questioned whether a fixed blast percentage is the best way to draw the boundary.² For now, the distinction still determines which clinical trials a patient can enter and which drugs are approved for their disease, making it clinically important even if biologically imperfect.

How Risk Categories Are Assigned

When hematologists say someone has “high-risk MDS,” they are referring to a score generated by a prognostic model. The most established is the Revised International Prognostic Scoring System (IPSS-R), which was developed from data on over 7,000 patients. It uses three pillars: bone marrow cytogenetics (chromosome changes visible under a microscope), the percentage of immature blast cells in the marrow, and the depth of cytopenias in the blood. Those inputs produce a numerical score that slots patients into one of five categories: very low, low, intermediate, high, and very high risk.³

In recent years, a newer model called the IPSS-M (Molecular IPSS) has added gene-level data to the mix. It incorporates mutations in dozens of genes alongside the traditional cytogenetic and clinical features. Compared with the IPSS-R, the IPSS-M reclassified roughly 46 percent of patients into a different risk group, shifting a meaningful number of people up or down.⁴ In a transplant-focused study, about 30 percent of patients were reclassified to a higher risk category and 14 percent to a lower one when the IPSS-M was used instead of the IPSS-R. The IPSS-M was better at identifying patients at very high risk of relapse after a stem cell transplant, while the IPSS-R groups showed similar survival after transplant and could not distinguish the worst outcomes as clearly.⁵

The practical upshot: if your hematologist uses only the IPSS-R, your risk assignment might change when molecular data are factored in. Institutions are increasingly adopting the IPSS-M to guide treatment, especially transplant decisions, but access to the necessary gene sequencing panels is not yet universal.

What Is Happening Inside the Bone Marrow

A bone marrow biopsy is the cornerstone of MDS diagnosis. Under the microscope, pathologists look for dysplasia in the three cell lines (red cells, white cells, and platelets), count the percentage of blasts, and note structural features like the presence of ring sideroblasts (iron-laden red cell precursors) or fibrosis. Current classification schemes also recognize specific genetic subtypes defined by an isolated deletion on chromosome 5, an SF3B1 mutation, or a TP53 mutation.¹

Bone marrow fibrosis, the scarring of marrow tissue, deserves particular attention. Fibrosis has been identified as an independent risk factor for poor survival in MDS patients regardless of their IPSS-R score. Its presence can tip the balance toward recommending a transplant, especially for patients who otherwise fall into the intermediate-risk zone where the decision is not straightforward.⁶

Beyond what pathologists can see on a slide, the bone marrow microenvironment itself is altered in MDS. The niche of supporting cells, blood vessels, and signaling molecules that normally sustains healthy blood-forming stem cells gets reprogrammed. Early on, the microenvironment may actually work against the abnormal clone, but as the disease progresses, it shifts to provide the dysplastic cells with a survival advantage.⁷ This remodeling of the marrow niche is one reason MDS becomes increasingly difficult to treat over time.

Gene Mutations That Shape Prognosis

Next-generation sequencing has found recurrent genetic abnormalities in up to about 90 percent of MDS patients, making it an essential diagnostic and prognostic tool.⁸ Not all mutations carry the same weight, however, and a few stand out for their impact on high-risk disease.

TP53 mutations are arguably the most consequential in high-risk MDS. TP53 is a tumor suppressor gene, and when both copies are damaged (“biallelic” or “multi-hit” TP53), the disease behaves aggressively. Patients with multi-hit TP53 have extremely high rates of complex chromosome abnormalities (above 90 percent in one large study) and poor responses to most treatments.⁹ Whether TP53 is mutated in one copy or both has become a major factor in transplant planning, as discussed in the treatment section below.

On the opposite end, SF3B1 mutations are generally favorable. A large meta-analysis found that patients with SF3B1-mutated MDS had substantially better overall survival and a lower chance of leukemic transformation compared with those without the mutation.¹⁰ SF3B1 mutations are so prognostically distinct that they are now recognized as a defining entity in MDS classification.

Other splicing-gene mutations tell a different story. Mutations in SRSF2 and U2AF1 are each linked to significantly worse overall survival and a higher risk of progressing to leukemia. The same meta-analysis found that both SRSF2 and U2AF1 mutations were associated with roughly 60 percent higher mortality compared with unmutated patients, while U2AF1 mutations nearly tripled the risk of leukemic transformation.¹⁰ These genes all affect the same cellular machinery (RNA splicing), yet their clinical meaning diverges sharply, which is part of why MDS can be so unpredictable at the individual level.

The Diagnostic Workup Beyond the Biopsy

Bone marrow aspiration and biopsy remain the mandatory starting point, but the diagnostic picture has expanded well beyond morphology. Flow cytometry, which measures surface proteins on individual cells, has become nearly universal in clinical practice for evaluating suspected MDS, even though it is not strictly required by international guidelines.¹¹ The technique is especially useful for picking up subtle dysplasia in white-cell and red-cell precursors that might not be obvious on a slide, helping to distinguish MDS from non-cancerous causes of low blood counts.¹²

Gene sequencing panels have rapidly gone from research tools to clinical necessities. Because the IPSS-M depends on mutational data, and because specific mutations like TP53 or SF3B1 now define classification subtypes, sequencing has moved from a “nice to have” to something that directly changes treatment recommendations. Not every community oncology practice has in-house sequencing capability, so samples are often sent to reference laboratories. The turnaround time can be a source of anxiety when patients are waiting for a risk category that will guide their next steps.

Survival and Causes of Death

The prognosis for high-risk MDS is sobering. A single-institution study found that median overall survival in the high-risk group was about 11 months, compared with roughly 3.7 years for low-risk patients. Over 90 percent of the high-risk patients had died at the time of analysis.¹³

A much larger study of nearly 2,900 patients with documented causes of death confirmed the broad pattern. About 83 percent of deaths were directly disease-related: progression to AML (nearly half of all deaths), infections (about a quarter), and bleeding (roughly 10 percent). The remaining deaths were from unrelated causes, most commonly heart failure. As risk category climbed, so did the proportion of disease-related deaths.¹⁴ In the single-institution study, only about 7 percent of high-risk patient deaths were unrelated to MDS, compared with roughly a third in the low-risk group.¹³

Transformation to AML is the feared endpoint. Research analyzing MDS at the molecular level has shown that distinct genetic abnormalities detectable at earlier stages of MDS alter disease biology and predict which patients will progress.¹⁵ The MDS-to-AML continuum is increasingly viewed as a single spectrum rather than two distinct diseases, a perspective that is reshaping clinical trial design and drug development.

Transfusion Dependency and Iron Overload

Many patients with high-risk MDS need regular red blood cell transfusions to manage severe anemia. Transfusion dependency is not just a symptom of failing marrow; it is an independent prognostic factor. Patients who are transfusion-dependent at diagnosis have significantly shorter survival than those who are not, and the more transfusions required, the worse the outlook.¹⁶

Repeated transfusions cause iron overload because the body has no efficient mechanism for excreting excess iron. In a large analysis, the development of iron overload was the strongest independent predictor of both overall survival and the risk of leukemic transformation, outperforming even established scoring systems like the IPSS. Transfusion dependency was the second strongest predictor. The study also suggested that iron chelation therapy, which removes excess iron, could improve survival and reduce AML risk, though this remains an area of active study.¹⁷ For lower-risk patients who may live long enough to accumulate dangerous iron levels, chelation is more clearly worthwhile. For high-risk patients with limited expected survival, the calculus is less clear-cut, and treatment priorities typically focus on disease-modifying therapies.

Treatment for High-Risk MDS

The only treatment with curative potential in high-risk MDS is an allogeneic stem cell transplant (sometimes called a bone marrow transplant), in which healthy donor cells replace the patient’s diseased marrow. For eligible patients, transplant is the standard of care. However, the majority of MDS patients will never be candidates, because of advanced age, other health conditions, or the lack of a suitable donor.¹⁸

Transplant outcomes have improved over time. In a large retrospective study of over 2,000 higher-risk MDS patients, 427 underwent transplant. Median post-transplant survival was about two years, with a five-year relapse rate of 37 percent and a five-year treatment-related mortality rate of 25 percent. Survival after transplant improved markedly across eras: median survival was about 12 months for transplants performed between 2000 and 2010, roughly 28 months for 2011 to 2016, and about 40 months for 2017 to 2023. The strongest predictor of transplant success was TP53 status. Patients without TP53 mutations had a five-year survival of 69 percent after transplant. Patients with TP53 mutations fared far worse: median survival was about 9 months for those with a single-hit mutation and under 7 months for biallelic TP53.¹⁹

A separate study using a biological assignment design (comparing patients who had a donor available versus those who did not) confirmed that transplant improved overall survival even in patients with TP53 mutations, regardless of whether one or both copies were affected. Patients in the IPSS-M very-high-risk group who lacked TP53 mutations also had favorable outcomes when a donor was available.²⁰ These findings illustrate why TP53 testing is so central to transplant discussions: it defines expectations and can shape the timing and urgency of donor searches.

Hypomethylating Agents

For patients who are not transplant candidates, the backbone of treatment is hypomethylating agents (HMAs), primarily azacitidine and decitabine. These drugs work by reversing some of the abnormal gene-silencing that occurs in MDS cells. Azacitidine has shown overall response rates between 40 and 60 percent in randomized trials, improved quality of life, a reduced risk of AML transformation, and a definite survival advantage over supportive care or low-dose chemotherapy.²¹ Both azacitidine and decitabine are widely used, and the choice between them depends on clinical factors and institutional preference.²²

While HMAs extend survival and improve blood counts for many patients, they are not cures. Responses are often temporary, and once a patient stops responding, the outlook worsens sharply. This reality has driven intense interest in combination approaches.

Venetoclax Combinations and Emerging Approaches

Venetoclax, a drug that blocks a survival protein called BCL-2, has transformed treatment in AML and is being studied in high-risk MDS. In a phase I/II trial combining venetoclax with azacitidine in treatment-naive high-risk MDS patients, the overall response rate was 93 percent among evaluable patients. Responses came quickly, within one to two treatment cycles. Median progression-free survival was about 8 months and median overall survival was 13 months.²³ This combination is also being explored as maintenance therapy after transplant, using low weekly doses of decitabine with venetoclax to reduce the risk of relapse with manageable side effects.²⁴

These early results are encouraging, though the studies are small and ongoing. The field is moving quickly, with multiple trials testing new combinations, immune-based therapies, and agents targeting specific mutations. For patients in the high-risk category, the practical advice is to ask about clinical trial eligibility at every treatment decision point.

Therapy-Related MDS

A subset of high-risk MDS cases develop as a late consequence of prior cancer treatment. Patients who received chemotherapy or radiation for breast cancer, lung cancer, or other malignancies can develop MDS years later. These therapy-related cases tend to carry a heavier burden of high-risk genetic features and generally respond less well to standard treatments like HMAs. They have also historically been excluded from many of the major clinical trials that established standard-of-care regimens, which means the evidence base for treating them is thinner than for de novo (newly arising) MDS. Awareness of this subgroup matters for survivors of other cancers who develop new blood count abnormalities.

Fatigue, Symptom Burden, and How They Predict Outcomes

Fatigue in MDS is not ordinary tiredness. In a study comparing MDS patients with those who had other bone marrow failure conditions, the average fatigue score fell in the “severe” range, and quality-of-life scores were meaningfully impaired.²⁵ The fatigue is driven partly by anemia but also by the disease biology itself, including inflammatory signaling from the abnormal marrow.

What makes this clinically interesting is that self-reported fatigue has been shown to independently predict survival in higher-risk MDS, beyond what the IPSS-R score captures. A Canadian registry analysis validated a model that incorporates patient-reported fatigue alongside the IPSS-R, stratifying higher-risk patients into subgroups with meaningfully different survival curves. Even a single-item fatigue rating scale, which takes seconds to complete, added prognostic power.²⁶ The implication is straightforward: when a patient with high-risk MDS says “I’m exhausted,” that is not merely a quality-of-life complaint. It carries prognostic information that clinicians should take seriously, and it is increasingly being incorporated into clinical trial design as a formal endpoint.²⁷

Where the MDS-AML Line Is Headed

The traditional 20 percent blast cutoff that separates MDS from AML has been under scrutiny for years. Biologically, a patient with 18 percent blasts who carries TP53 biallelic mutations and complex chromosomal changes may have a worse prognosis than someone with 22 percent blasts and a favorable mutation profile, yet the first patient is labeled MDS and the second AML. This mismatch has practical consequences: it determines drug access, trial eligibility, and insurance coverage.²

Newer classification proposals from the World Health Organization and the International Consensus Classification have begun to acknowledge genetic features as boundary-defining rather than relying solely on blast counts. This is not just academic hairsplitting. If the boundary softens, patients with high-risk MDS and patients with oligoblastic AML (AML with relatively low blast counts) could gain access to the same clinical trials and treatment algorithms, which would better match treatment intensity to disease biology rather than to an arbitrary numerical threshold. For now, though, the 20 percent line still governs most clinical decision-making worldwide, and patients near that border often find themselves caught between two treatment paradigms.