Prospective Comparative Evaluation of Flow Cytometric Scores Regarding the Sensitivity and Specificity As Well As Prognosis in 616 Patients with Proven or Suspected MDS

Background:

Flow cytometry (FCM) has recently been recognized as an important supplementary tool in the diagnostic work-up of patients with MDS. Within the international MDS-flow working group of the ELN, different and also novel Flow scores have been evaluated and tested in rather small patients series.

Aim:

Currently available FCM scores have been prospectively compared concerning their potential diagnostic and prognostic impact in patients with proven or suspected MDS.

Patients and Methods:

Bone marrow (BM) samples from 616 patients (951 measurements) with proven or suspected MDS, 42 MPN patients, 30/18 MDS patients in cytomorphologic/cytogenetic CR, 22 patients with inappropriate BM aspirations as well as 48 age matched controls have been analyzed by FCM. For measurement and analysis a FACS-CantoII cytometer including FACS-DiVa software was used, performing an 8-color panel and measuring 200,000 events per sample. The following FCM scores were compared: FCSS (granulopoiesis / monopoiesis), Ogata-score (blasts), ELN-red score and Mathis score (both erythropoiesis), and the new iFScore (granulo-, mono-, erythropoiesis, and blasts). Overall survival (OS) was estimated with a median follow up of 21 month (range: 3.4-77 month) applying the Log-Rank-Test within SPSS software.

Results:

The new iFScore, evaluating granulo-, mono-, erythropoiesis and blasts, turned out to be the most comprehensive score with the best balance between sensitivity and specificity (81%; 78%). FCSS indeed showed a comparably high sensitivity (82%) but has a clearly lower specificity (60%). The Ogata-score and the red-scores showed a lower sensitivity (60%, 31%, 40%) but were very specific (92%, 97%, 89%). Interestingly, despite inappropriate bone marrow aspirations (w/o crumbs) in 22 patients the new iFScore accounted for a "MDS conformable" result in 72% of these cases.

Next, the prognostic impact of the different FCM scores was evaluated. Of note, patients with MDS conformable FCM-scores showed a significantly shorter OS compared to patients with scores being not MDS conformable, new iFS, Ogata- and Mathis-score: p = 0.001 (ELN-red: p = 0.003; FCSS: p = 0.005). This also held true evaluating only untreated patients, with Ogata-score exhibiting the highest significance (p < 0.001). Remarkably, even when analyzing the subgroup of MDS-SLD/MLD with normal karyotype and without ringsideroblasts plus non-clonal cytopenias only, a high Ogata score predicted for a significantly lower OS (p = 0.035; median OS = 40 month vs. not reached). Thus, the group of patients with MDS conformable FCM showed significantly higher IPSS-R (p = 0.023) and more mutations (p = 0.0357).

When evaluating MDS patients in cytomorphologic or even cytogenetic CR, 23% (17%) showed a MDS conformable iFS with significantly shorter OS (p = 0.025; p = 0.026).

Finally, 49% or 35% of patients with MPN showed MDS conformable iFS or Ogata scores, respectively. This translated in a shorter OS in those patients (p = 0.048; p = 0.018).

Conclusions:

Currently available FCM scores have different sensitivity and specificity in diagnosing MDS, with the novel and comprehensive iFScore, capturing granulo-, mono-, erythropoiesis and blasts together, being superior to other FCM-scores.

Additionally, iFS and Ogata score comprise significant prognostic information, even in lower risk MDS. Remarkably, FCM might be a tool to refine response evaluation in MDS and reveal significant dyspoietic features in a subset of MPN patients.