Impact of Eltrombopag on Clonal Evolution in Refractory Aplastic Anemia

Despite documented success of IST in tx of aplastic anemia (AA), a significant minority of pts remain refractory to tx and most clinical responses are incomplete. Due to comorbidities, HSCT may not be available for all refractory pts and until recently anabolic steroids were the mainstay of salvage tx. With the approval of the novel cMpl agonist eltrombopag (E), first in refractory and now in the upfront setting for SAA, the tx paradigm has changed.

Clonal evolution from AA to MDS remains a serious complication. With the advent of E, a valid question proposed is whether the growing use of E would improve recovery rates thereby decreasing the rate of clonal evolution or if stimulation of growth factor receptors by E would potentially increase progression to sMDS.

At our institution, 308 AA pts were managed from 1998-2018. Of these, 35 (median age 67 yrs, f/u 42 mo, F: M 17:18) were tx with E for IST-refractory AA (50-150 mg PO QD), with median tx duration of 19 mos; in contrast 202 pts tx in the non-E era received ~2 tx lines (median age 40 yrs, f/u 60 mo, F: M 99:103). Response was evaluated by sustained improvement in blood counts and transfusion independence after 12 wks of tx. This was attained in 66% (23/35). 14% (5/35) had stable disease w/intermittent transfusions (1 underwent HSCT). Among non-responders 1 developed PNH prompting tx w/eculizumab, 1 received HSCT and 2 progressed to MDS/AML. In non-E SAA pts, ORR to 1st tx was 61%. 18 progressed to hemolytic PNH, 20 to MDS/AML and 49 to HSCT. The MDS/AML evolution rate from SAA in pre-E era has been~15% in 10yrs(Socié, G et al. 2000). Our analysis was confirmatory, with median time to progression of 4 yrs. This implies that no firm estimates can be made on the impact of E on MDS progression on clinical grounds alone.

Somatic mutations (SM) typical of MDS have been detected in AA/PNH. Studies show that SM present at evolution to MDS can be detected at AA dx in 30% pts. Clearly, some of the SM seen in AA constitute a progression risk, while others likely reflect clonal hematopoiesis. We studied clonal kinetics using deep NGS to identify early signs of clonal evolution/eradication and subsequently predict MDS risk. Reported is our retrospective analysis using a case control approach.

35 E-tx pts were matched with 32 non-E tx pts using response rates and clinical parameters. Serial samples were obtained at start and end of tx for both groups. 4/35 E pts had a chromosomal abnormality at dx(t[10;18],+Y,+12 and del11q, i(7)(q10); latter 3 resolved on f/u). In non-E pts, 3 had cytogenetic abnormality at dx (del16q,+15,+Y; latter two acquired + 6 and -7). Due to close PNH association we evaluated E related size of PNH clone. PNH granulocyte expansion was seen in 11/35 after E and in 5/32 non-E pts. In the E group, 2/35 progressed to MDS/AML compared to 7/32 in the non-E group (p=.08). Similarly there was no difference in PNH progression (p=.36).

Serial NGS analysis for both groups revealed a ≥1 SM per pt (BCOR, PIGA, ASXL1, DNMT3A, etc.) in 14 E- pts vs 4 in control group (U2AF1, NF1,PHF6,SAMD9L). Average VAF of BCOR was 19% and PIGA was 24% in E and it was 34% in non E. At the end of observation period, 22/35 (E) had ≥1 SM vs 12/32 in control. NGS post tx noted that 5/35 w/single SM (CEBPA, EZH2, SAMD9L, U2AF1, TP53) had minimal expansion of their clone. Only the pt w/EZH2 acquired 2 additional clones of RUNX1 and U2AF1. The pt w/TP53 developed MDS. 6 pts expanded their original clone (BCOR, ASXL1/U2AF2, PIGA, VHL, NRAS, PIGA) of which the latter 4 acquired 1 additional SM (ZRSR2, PIGA, ASXL1, BCORL1). 1 pt had regression of the RIT1 clone, but acquired a BCOR. 2 others had transient clones of BCOR and DNMT3A at dx w/acquisition of ASXL1. Of those w/baseline SM, 8/21 acquired new SM, of which 1 developed AML w/ gain of DNMT3A, ASXL1 and U2AF1. In the controls, 2 pts w/NF1 and SAMD9L expanded their original clone and acquired ASXL1/SETBP1 and TP53 respectively, w/both having MDS/AML progression. 1 pt had a decrease in U2AF1 and another had a transient PHF6 clone. Pts without baseline SM acquired ~2 SM during the disease course. Many of these are the same SM classically found in MDS/AML.

To date, there are no data to support an increased clonal evolution risk with use of E in AA pts either on clinical grounds or by molecular testing. However, presence of certain somatic mutations (CBL, SETBP1, DTA andRUNX1) strongly associated w/MDS/AML progression may warrant caution in the application of E or at least close molecular monitoring.