PMF with cytopenias ddx MDS with fibrosis?

[elburrito]

I've encountered several times in practice where I see first time diagnostic marrow biopsies morphologically consistent with PMF with marked reticulin fibrosis, but if CBC demonstrates cytopenias, and if splenomegaly is absent, the clinician questions whether this might be MDS with fibrosis. Although studies have shown there is interobserver variability in the evaluation of megakaryocyte atypia, I think the morphologic features of atypical megas in PMF or MPNs is specific enough that it can be reasonably distinguishable from MDS dysmegakaryopoiesis. I don't believe there is much of a basis to consider MDS with fibrosis as a significant differential diagnostic consideration when you have a marrow biopsy that looks like advanced fibrotic phase PMF with cytopenias. I was wondering if anyone else might have any opinions on this? Basically, if JAK2, CALR, MPL are negative, I will still top line as, "MPN, morphologically compatible with PMF" but lately I've felt the need to include a sentence in the comment, "MDS with fibrosis could be considered, although the cytologic features identified in this case would be unusual." Obviously I'm referring to cases where cytogenetic studies are negative for MDS defining abnormalties.

---

Members don't see this ad.

 

[cmz]

I honestly don't know how to answer your question. Given the clinical picture that you painted above (e.g. cytopenias, absent hepatosplenomegaly, etc), I suppose MDS with fibrosis could be considered in the differential; however, I wouldn't give that as my diagnosis. Were the megakaryocytes small, hypolobated or more bizarre/atypical forms? Any bony changes? All JAK2 mutations tested for? Have you considered next-gen sequencing? 😉

I am sure you could write a nice comment that would heavily favor calling the marrow MPN, c/w PMF. You can paint the picture any way you want as long as it sounds reasonable. The clinicians just want something to relay to their patient because there is a slight difference in overall survival. If this is the first time diagnostic marrow... *shrug*. As they say, sometimes you just have to grab your balls and make a diagnosis.

Do you have any pictures of the case you could post?

 

[cmz]

On a side note, maybe I am practicing in a place where no one does adequate physical exams, but I hardly see a comment about splenomegaly/organomegaly in any of the progress notes/HPIs that I read. Unless someone ordered an ultrasound or CT abdomen, that kind of stuff isn't mentioned. It's nice to know this bit of information when you're dealing with MDS, MPN, etc (even mantle cell).

 

[elburrito]

Thanks for the response. I am always skeptical of negative physical exam findings of splenomegaly, but if there is positive palpable hepatosplenomegaly I factor that in. If ultrasound or CT imaging studies are available for correlation then that is most reliable. The megakaryocyte atypia I think of for MPNs including PMF are hypercondensed, hyperchromatic, or bizarre, versus myelodysplastic forms that demonstrate separated nuclear lobes.

You bring up an interesting point about NGS. Do many clinicians order and request it in your area for myeloid malignancies? I just had a case the other day that had a non-classical CALR mutation in a case that is morphologically compatible with ET. The mutation was only reported by the reference lab as being "not a type 1 or type 2" but the report didn't say what it was specifically. I'm curious about the role of NGS in myeloid malignancies since next year's WHO I've head is not going to expand much on sequence variations except for RUNX1 and TP53. Thus, if anyone draws conclusions about variants identified in an NGS panel, e.g. IDH1 or DNMT3A which are known to be highly recurrent, we would end up practicing outside of the WHO guidelines. I had hoped that NGS would emerge as a critical diagnostic adjunct for morphologically subtle, cytogenetically normal MDS but I don't foresee that happening soon.

 

[cmz]

Thanks for the response. I am always skeptical of negative physical exam findings of splenomegaly, but if there is positive palpable hepatosplenomegaly I factor that in. If ultrasound or CT imaging studies are available for correlation then that is most reliable. The megakaryocyte atypia I think of for MPNs including PMF are hypercondensed, hyperchromatic, or bizarre, versus myelodysplastic forms that demonstrate separated nuclear lobes.

You bring up an interesting point about NGS. Do many clinicians order and request it in your area for myeloid malignancies? I just had a case the other day that had a non-classical CALR mutation in a case that is morphologically compatible with ET. The mutation was only reported by the reference lab as being "not a type 1 or type 2" but the report didn't say what it was specifically. I'm curious about the role of NGS in myeloid malignancies since next year's WHO I've head is not going to expand much on sequence variations except for RUNX1 and TP53. Thus, if anyone draws conclusions about variants identified in an NGS panel, e.g. IDH1 or DNMT3A which are known to be highly recurrent, we would end up practicing outside of the WHO guidelines. I had hoped that NGS would emerge as a critical diagnostic adjunct for morphologically subtle, cytogenetically normal MDS but I don't foresee that happening soon.

I do community path, so pretty much no one requests NGS (yet). Maybe someone who works exclusively with heme could comment on this. I am pretty sure that there has to be an expanded section in the AML chapter in the next WHO discussing some of the genes you've touched upon.

http://www.nejm.org/doi/full/10.1056/nejmoa1112304. Figure 1 is pretty amazing.

 

[elburrito]

The Patel paper is one of the most heavily cited papers in the literature for the past few years on utility of extended NGS panel evaluation in enhancing prognostication in AML. A critical flaw with this paper is they set up their comparison between risk stratification based on NGS and based on classical cytogenetic profiling, showing they could further substratify and re-assign some intermediate risk AML patients into favorable (hence no transplant) and unfavorable risk groups. A number of those patients who were re-assigned from intermediate to favorable (hence transplant to no transplant) in this study would have been appropriately classified as favorable using the already established 1-off testing for NPM1 and FLT3 based on current WHO defined algorithms. That being said, Patel et al did demonstrate value for intensified induction chemo based on presence of DNMT3A mutations. There was another recent paper from a group in Asia that evaluated 373 AML cytogenetically normal patients and showed that extended mutational profiling could re-assign some patients from favorable to unfavorable risk (hence no transplant to transplant) based on identification of TET2 mutations; NPM1+, FLT3-, TET2+ patients actually did worse than NPM1+, FLT3-, TET2- patients (p=<0.05) (Int J Hematol. 2014 Jul;100(1):96-104). I think it will take years before enough patients are studied and reported for consensus to be achieved in the literature on what are the most relevant combinations of mutations that are prognostically significant and can guide transplantation and management. There are too many potential combinations once extensive gene panels are evaluated (e.g. what would be the combined significane of a FLT3ITD+, DNMT3A+, IDH1+, NPM1+ AML)? Its already hard enough using just FLT3, NPM1, CEBPA and cytogenetics and it took years for the literature to settle on the significant combinations using just those variables. Plus, another layer of complexity is that different mutations in given genes carry different weight (e.g. FLT3 ITD vs. D835). The literature often fails to recognize this and patients with mutations in specific genes are often lumped together in the analyses. I think the complexities for prognostication are limiting, and greatest promise of NGS may lie in its application as a biomarker for directing targeted therapy or MRD evaluation (e.g. Mod Pathol. 2014 Nov;27(11):1438-46) - NGS for NPM1 was demonstrated in that study to be more sensitive than flow cytometry by an order of magnitude for MRD detection. I was really hoping there would be more convergence in the field on utility of NGS as a diagnostic adjunct to help in early identification of morphologically subtle MDS with normal cytogenetics. Plus, God help us figure out who is going to pay for NGS panels when CMS billing code establishes reimbursement as only $90 for NGS panels >5 genes.