Page 139 - 2022_01-Haematologica-web

Predicting response to IST plus EPAG in SAA
reported concerning response predictors, partly because of the very limited number of non-responders after such highly effective therapy. Our large retrospective study in 177 SAA patients treated with IST plus eltrombopag, incorporating 85 new subjects after the initial publica- tion,4 has now revealed that pretreatment blood counts remain strong predictors of overall response to IST plus eltrombopag, similar to their predictive value in 240 patients treated with IST alone. We also noted the rela- tionship of ALC with complete response for the first time in both groups of patients, i.e., those treated with IST plus eltrombopag and those treated with IST alone.
The overall response rate to IST plus eltrombopag exceeded 90% in patients who had an ARC of at least 10×109/L, and the complete response rate reached 60% in adolescents and young adults who had an ALC of more than 1.3×109/L prior to treatment. The addition of eltrombopag to IST markedly improved responses in patients who had an ARC in the 10-30×109/L range. In contrast, the addition of eltrombopag to IST did not sig- nificantly increase overall response in patients who had an ARC lower than 10×109/L or higher than 30×109/L, but was still beneficial as it induced significantly more complete responses compared to IST only.
Multiple studies have attempted to predict response to IST in patients with SAA because HSCT is available as a competing treatment option. In practice, only the patient’s age, in addition to the availability of an appropriate donor, is generally considered to decide initial therapy.2 This algo- rithm is based on outcomes of patients treated with IST alone,35,36 and may need to be revised in the eltrombopag era. In the short-term, patients with residual hematopoiesis might benefit more from IST plus eltrom- bopag than from HSCT, but long-term follow-up especial- ly of relapse and clonal evolution with this new combina- tion is necessary and is currently underway. The impact of pre-transplant blood counts on the outcomes after HSCT needs to be studied as well in AA.
The correlations of initial ARC and ANC with overall resposne were not surprising, as residual hematopoietic stem and progenitor cells should have an impact on hematopoietic recovery after immunosuppression, and they themselves are criteria for response. Similarly, the impact of thrombopoietin levels would be attributable to their inverse correlation with platelet production,37,38 another overall response criterion, although pretreat- ment platelet counts, affected by platelet transfusion, were not predictive of responses to IST.
Complete response was unexpectedly not discriminat- ed by the strong overall response predictors, including ARC, ANC and thrombopoietin levels; these may not represent the quantity or quality of hematopoietic stem cells needed for complete hematologic recovery, as blood counts of patients with AA are often supported by aber- rant hematopoiesis from PNH clones,24,39 HLA class I allele lacking clones,40-42 or clones associated with age- related somatic mutations.23 These are usually detected in myeloid and erythroid cells but are less frequent or absent in lymphocytes.43-45
We observed an association of ALC with complete response and further validated the correlation in rabbit ATG-based IST. ALC may be more representative of lev- els of residual hematopoietic stem cells than other blood counts. On the other hand, recent studies in HSCT indi-
cated that ATG dose adjustment based on ALC, rather than body weight, is beneficial to achieve an optimal level of immunosuppression because circulating lympho- cytes bind and clear ATG;46-49 a low pretreatment ALC leads to reduced clearance of ATG and prolonged immunesuppression, which is likely unfavorable for hematologic recovery, as observed in patients treated with rabbit ATG-based IST.32 The impact of ALC was age-dependent, as complete response was associated with lower ALC in young children aged <10 years, but with higher ALC in adolescents and adult patients. Our literature review supported the differential impact of pretreatment ALC in children and in adults. The differ- ence by age is probably associated with higher normal ALC and higher lymphoproliferative capacity in young children than in adolescents and adults;50,51 our body weight-based ATG dosage might be insufficient to sup- press proliferative lymphocytes of young children unless ALC is markedly reduced before treatment.
A tendency to a slightly higher overall response in patients with PNH clones was attributed to the correla- tions of the presence of PNH clones with higher ALC and ARC, which may represent escape hematopoiesis by PNH-phenotype hematopoietic stem cells and underesti- mation of PNH clones in very severe disease in which neutrophils are absent and red blood cells are replaced by transfusion. A lower threshold allows detection of patients having small PNH clones but also creates more difficulties in very severe disease as many more cells are needed to identify the small populations. Care must be taken to interpret biomarker studies on prediction of response to IST, as they are also potentially influenced by the blood counts. For example, clonal hematopoiesis that preferentially involves myeloid progenitors will be under- estimated in patients with low ANC; thrombopoietin lev- els show an inverse correlation with platelet produc- tion;37,38 and lymphocyte subsets also corelate with myelopoiesis.28
In summary, residual hematopoiesis, as indicated by baseline ARC, ANC and thrombopoietin, is predictive of initial overall response to eltrombopag plus IST, but does not correlate with complete response. We did, however, observe an age-dependent association between complete response and ALC. These findings should be helpful for future protocols and in decision-making regarding treat- ments for individual patients.
Disclosures
No conflicts of interest to disclose.
Contributions
YZ, PAB and NSY designed the study and wrote the first draft of the manuscript. PAB, EMG and NSY participated in the patients’ care. YZ, RS and COW analyzed the data. YZ and XF quantified thrombopoietin levels. All authors critically reviewed the draft and approved the final version for publication.
Funding
This investigation was supported by the Intramural Research Program of the NIH and the NHLBI. NCT01623167 is funded by Novartis and previously by GSK through a cooperative research and development agreement (CRADA).
Data-sharing statement
Individual patients’ information will not be disclosed.
haematologica | 2022; 107(1)
131