IPSS-M and IPSS-R Tools for Risk Stratification of MDS

Amer Zeidan, MBBS, and Rami Komrokji, MD, compare the available IPSS-M and IPSS-R tools for stratifying risk and assessing prognosis in MDS.


Amer Zeidan, MBBS: Multiple other abstracts have looked at this issue. The other group of abstracts looked at the Molecular International Prognostic Scoring System (IPSS-M), which is the other biggest news in MDS [myelodysplastic syndromes] in 2022 in terms of prognostication. We used molecular data in a nonsystematic fashion for MDS for more than 10 years, but the molecular IPSS was one of the largest because there have been other important efforts to integrate molecular data.

The molecular IPSS was published earlier in the year in the New England Journal of Medicine Evidence, and multiple groups have tried to validate that. There have been at least 6 or 7 different abstracts that looked at this, and several of them were presented in one of the oral sessions. The message from all of them is that the IPSS-M certainly seems to be valid. It seems to have some improvement over the traditional nonmolecular-based risk groups, such as the IPSS-R.

The biggest challenge continues to be regarding how many of those genetic alterations you really need to have in the model. There are concerns that you are using a very large number of genetic alterations in the current model and it's not clear that those are easily available in every setting, especially in community or global settings, but how much the accuracy of the system is affected if you are using a lower number of genes. One of the interesting validations that looked at this compared the molecular IPSS with 2 other systems that were developed by the clinical research consortium led by Aziz Nazha, MD, and another from the European groups, and it was interesting that while the IPSS-M continues to perform well, it didn't seem necessarily much better than those other systems.

Molecular incorporation in prognostication is certainly very important, but we still need to do more work in figuring out the minimum number of genes that have to be incorporated and go beyond prognostication more into prediction models, because this is what's going to eventually become very important for our patients to be able to select therapies that seem to particularly work for them rather than just focusing on the idea of the prognostication, or classification. Before I ask about your sense on lower-risk MDS, I wanted to see if you wanted to add any more to that issue.

Rami Komrokji, MD: Thank you for this review. I totally agree. We have 2 classifications that are confusing, and we need to come to a consensus for practicing oncologists to use. There is no doubt that the advantages of the new classifications are introducing molecularly defined subsets defining the morphology, but we need to come to a consensus. I agree with you. IPSS-M seems that they're finding the prognosis. One of the issues that keeps coming [up] is looking at the mutations that are either favorable, like the SF3B1, which we are learning impact of comutations, other cytogenetic abnormalities, but then also the story of the TP53 mutation. This is universally a bad mutation to have in any disease, particularly hematologic malignancies. It remains as one of our unmet challenges that the IPSS-M incorporate.

We did a presentation looking at the value of monoallelic or 1-hit p53, or 2 hits. If patients have a variant allele frequency of more than 50%, or if they have loss of 17 or loss of heterozygosity, that's called multi-hit, or if they have 2 mutations as well. Clearly, 2-hit mutations in p53 has a worse outcome. And in our abstract, if patients had a complex karyotype, a monoallelic was bad. It seems that there is some correlation with the blasts, so as the blasts go higher, the outcome with p53 is worse. The International Consensus Classification (ICC) actually incorporates that, saying if the myeloblasts are more than 10%, whether it's monoallelic or biallelic, it's regarded the same. Our group had a cutoff of 5% or more. The German group also presented some data on the p53. Is it prime time to use IPSS-M in practice? Are we telling our community oncologists to look at that?

Amer Zeidan, MBBS: Yes, I think so. It's certainly the time. Because for the most part, we have been applying some molecular data offhand. For example, our expectation is that in a younger, lower-risk patient who has a high-risk molecular alteration, such as TP53, we certainly would be contemplating things we generally do for higher risk, such as potentially bone marrow transplant. However, doing this in a systematic fashion within the IPSS-M makes more sense.

At the same time, people have to be cognizant of the limitations of those systems. I always joke that computers are not likely to replace doctors in the near future because it's not always a simple “OK, this is where you fall on the risk score, and you should do this or that.” You have to account for the particular patient situation, their social setting, and their goals. All of those factors should factor in. But there is no doubt that adding molecular data to the models seem to improve, although it's not as much as you would think. In many settings where people don't have access to molecular data, especially in maybe lower-income countries, going by the revised IPSS still has significant value and should continue to be used.

Transcript edited for clarity.

Related Videos
Related Content