How Do CMML Risk Scores Differ? CPSS-Mol, Mayo, iCPSS

There is no single “best” or universally adopted risk model for chronic myelomonocytic leukemia (CMML) ^[1]. Doctors use several different scoring systems—such as CPSS-Mol, the Mayo Molecular Model, iCPSS, and GFM—because they were developed using different patient groups and predict slightly different outcomes ^[2]. Which model your doctor uses often depends on the specific blood tests, bone marrow results, and genetic data available at your diagnosis. Because applying different models to the same person can sometimes result in different risk assignments, doctors may calculate more than one score to expose uncertainty and see how your disease is classified under different systems, rather than to get a single “perfect” prediction ^[1][2].

Why Are There Multiple Risk Scores?

Over the years, our understanding of CMML has evolved. Early models relied solely on clinical factors like blood counts and bone marrow blasts (immature leukemia cells in the marrow) ^[3][4]. As genetic testing improved, researchers discovered that specific acquired DNA mutations—which are changes that happen in the leukemia cells, not inherited traits you pass to children—play a major role in how the disease behaves. This led to the creation of new models that combine clinical features with molecular (genetic) data ^[3][5].

The available models differ in how they were created and what they measure:

• Study populations: Some models were built using data from a few hundred patients at a single institution, while newer ones use data from thousands of patients worldwide ^[3][5].
• Endpoints: While models like the Mayo Molecular Model and GFM focus heavily on estimating overall survival ^[6][7], models like CPSS-Mol and iCPSS also provide validated estimates for the risk of the disease transforming into acute myeloid leukemia (AML) ^[3][5].

It is important to remember that these models provide population-level statistics—estimates based on large groups of people—not a precise guarantee of what will happen to you individually.

Comparing the Major Risk Scores

Each scoring system weighs different clinical, hematologic (blood-related), cytogenetic (chromosome abnormalities in the cancer cells), and molecular variables.

• CPSS-Mol (Clinical/Molecular CMML-Specific Prognostic Scoring System): This model combines clinical factors—specifically red blood cell transfusion dependence (needing regular blood transfusions), white blood cell count, and bone marrow blast percentage—with cytogenetic risk and mutations in four specific genes: RUNX1, NRAS, SETBP1, and ASXL1 ^[3].
• Mayo Molecular Model (MMM): A straightforward point-based system. It assigns one risk point for each of five factors: a specific type of ASXL1 mutation (a truncating mutation, which cuts the protein short), an absolute monocyte count (a type of white blood cell) above 10 × 10⁹/L, hemoglobin below 10 g/dL, platelets below 100 × 10⁹/L, and the presence of circulating immature myeloid cells (early blood cells that shouldn’t normally be in the bloodstream) ^[6][8].
• GFM (Groupe Francophone des Myélodysplasies): Looks at age (over 65 years), high white blood cell count (over 15 × 10⁹/L), anemia, low platelets (under 100 × 10⁹/L), and ASXL1 mutation status ^[7][9].
• iCPSS (International CMML Prognostic Scoring System): One of the newest models, developed using data from over 3,000 patients ^[5]. It incorporates mutations from nine genes alongside blood counts and chromosome data, separating patients into five distinct risk groups. In its validation study, it reassigned 55% of patients to a different risk category compared to older models ^[5].

When researchers compare these models head-to-head, results vary. In a 340-patient study, CPSS-Mol was best at predicting leukemia evolution, while other large datasets show that multiple models offer broadly comparable performance ^[10][11].

The Challenge of Missing or Changing Molecular Data

Because newer models require advanced genetic testing (such as a next-generation sequencing panel), your doctor cannot calculate an accurate CPSS-Mol or iCPSS score if that genetic data is missing. If your molecular data is unavailable, doctors will rely on established clinical models (like the original clinical CPSS), which remain useful for predicting outcomes based purely on your blood counts, bone marrow blasts, and chromosome analysis ^[4].

Furthermore, a risk score is a snapshot in time. The genetic mutations in your bone marrow cells can evolve—especially if the disease starts progressing toward AML ^[12][13]. However, this does not mean you need a new bone marrow biopsy every time your symptoms change. Repeat testing is individualized and generally considered only when your blood counts worsen significantly, progression to AML is suspected, or a major change in treatment or transplant planning requires it ^[13].

Using Risk Scores for Treatment and Transplant Decisions

Risk scores do not automatically dictate your treatment. Instead, they guide conversations. Treatment decisions are made by combining your prognostic score with your age, symptoms, disease burden, and personal goals ^[14].

For patients considering an allogeneic hematopoietic stem-cell transplant—the only established treatment with curative potential for CMML—risk scores help weigh the substantial risks of the procedure against the risks of the disease ^[14][15]. While baseline scores like CPSS-Mol or iCPSS can help identify who might benefit from a transplant evaluation ^[16][5], standard CMML models do not always perfectly predict outcomes after a transplant is completed ^[17]. Transplant candidacy also depends heavily on your age, fitness, other medical conditions, and donor options. For this reason, transplant teams often use specialized post-transplant risk assessments to evaluate your specific situation ^[18][19].