IPSS/IPSS-R Reliability in Various Disease Settings 

Amy DeZern, MD, MHS: Comorbidities, as well as the individual human’s goals, aren’t incorporated into the IPSS-R [Revised International Prognostic Scoring System] either, so that can certainly be a limitation.

The other thing is that for the IPSS-R, perhaps a woman who had breast cancer and got very aggressive cytotoxic chemotherapy for that malignancy or a gentleman who had prostate seeds, radiation, or even docetaxel for their prostate cancer can have treatment-related MDS [myelodysplastic syndromes]. I rely a little less on the IPSS-R in those patients. Is that how you think about it?

Rami Komrokji, MD: I totally agree. We’ve moved mostly to the revised IPSS-R. There are so many different models out there. For somebody in the community who is busy seeing patients, you need to adapt 1 system and move with it. The IPSS-R does a pretty good job, much better than the IPSS. We are finally moving to incorporate that even to clinical trial because for some time there was a challenge. All the trials are based on the IPSS and in real life are using the revised IPSS.

As you mentioned, the only possible dilemma in the IPSS-R is that intermediate-risk group. I like what you do. I rely on the molecular profile, transfusion burden, presence of circulating blasts, and patient age to decide how I’m going to treat the intermediate group. The other thing is that I don’t think they capture the kinetics of the dynamics of the disease. Sometimes I’ll see some patients, and they are intermediate risk. I’ll say, “Let’s see how the disease will behave over the coming 3 to 6 months.” You see those patients who will have progressive profound cytopenia, so you know the disease is acting up, and some patients will actually stay steady. There is some element of disease dynamics or kinetics that those models don’t capture.

As you mentioned in the therapy-related MDS, those models did not include patients with therapy-related MDS. We tried to look at the large cohort here of almost 500 patients, and it does predict outcome. Stage to stage, a very low-risk therapy-related MDS will do worse than the de novo MDS, but you can’t segregate patients in the therapy-related MDS on that. The challenge is that almost 70% or 80% of patients with therapy-related MDS are going to be higher, or very high risk, because it’s mostly dictated by the cytogenetics, high-risk mutations like TP53. Generally, if a patient is lower—a therapy-related MDS but doesn’t have blasts, doesn’t have bad cytogenetics—it’s reasonable to label them in the lower-risk MDS bucket and go stepwise. I deal with that event with more caution, but they do a little worse, unfortunately, than de novo MDS, even stage to stage if they are the same stage. Part of that is probably the comorbidities from down the line in the disease.

I’m obviously aware of the comorbidities; they are very important in the higher risk. You’re trying to decide transplant or not, so the comorbidities play a very important role. There’s a question I ask: How do I use them dynamically or restage those patients? It’s challenging because those models were clinically developed at diagnosis, and I don’t think they’ve been validated in terms of restaging. I don’t really do a formal restaging. I look at the blast percentages increasing—that’s going to higher risk if we acquire new cytogenetic abnormality or mutation or develop profound cytopenia. It’s important to notice that, for example, after HMA [hypomethylating agent] failure, although we label sometimes patients as lower risk, their outcome is not really good unfortunately. With amyotrophic lateral sclerosis, it’s around 18 months for overall survival. We look together, and it has been projected looking at the utility of the risk models after HMA failure, and they really don’t predict the outcome very well. The patterns are different. Maybe some patients sustain that lower-risk MDS with more profound cytopenia and bone marrow failure, and some patients progress to MDS. I don’t use a formal restaging. I think this has fallen in clinical trials after standard-therapy failure because they require restaging, and I don’t think it’s an accurate way to do it. It’s hard to develop newer models for restaging. After there’s failure, we have enough data that even after ESA [erythropoiesis stimulating agent] failure or lenalidomide failure, patients don’t do as well, and we should not be very rigid using only those risk modifications or risk certification at that time.

What I tell my fellows usually is that in any disease, you have to go through those 2 steps, because then you’re going to move to the next step, which is the management. We probably should start moving to talk about the management because we have a lot of interesting drugs to talk about. I will try to make sure the diagnosis is right, I risk stratify the patients, and factor in the patient’s related factors, comorbidities, and social things. That dictates how we manage those patients, and then I move to manage those patients. There is no doubt that in MDS, supportive care is still an important part.

Transcript edited for clarity.