DISCUSSION
Current treatment protocols for both ALL and AML in children are effective, and the majority of children can be saved by intensive conventional chemotherapy with/or without stem cell transplantation.1,2 Indeed, the newest approaches, such as use of bispecific antibodies or cell therapy, are successful even for therapy-resistant ALL.5 It is still an open question whether immunotherapy alone can suffice to achieve long-lasting remission particularly in refractory disease.7 A number of immunotherapies for r/r AML are under investigation, but the best immunotherapy for this condition that is currently available entails attempting to control the disease and subsequently performing stem cell transplantation.8 Thus stem cell transplantation seems to be the most widely applied and proven curative procedure for children with refractory AL.9-12 It is commonly accepted that the better the disease control before HSCT,13-16 the better the expected outcome. Thus in r/r AL, the main focus is on effective reduction of blast cell load that can potentially bridge the patients to a salvage procedure: HSCT alone or cell therapy with or without subsequent HSCT. Immunotherapy was not available when we designed and performed our study, and the only possible approach to bridge the patients to HSCT at that time was intensified re-induction chemotherapy.17,18Clofarabine, both alone or in combination with other agents, is known to be effective and is recommended for relapsing AL, with acceptable toxicity. The salvage therapy is nearly always given to heavily pretreated children, including those with previous HSCT, and such patients often have pre-existing organ toxicities. We chose the combination of clofarabine, etoposide, and cyclophosphamide based on promising published data showing an acceptable toxicity profile (even in HSCT).6 However, our results suggest that the applicability of the protocol for this demanding patient group was limited by the treating physicians’ pursuit of an individualized approach, and a lack of confidence in repeating CloEC if the first course did not result in a substantial response. Two of the seven patients deviated significantly from the CloEC protocol: one of them was transplanted already after the first course, and the other one received a different second course. One patient received the third CloEC off protocol (despite not fulfilling the required response criteria) and proceeded to transplantation while not in morphological remission; this decision reflected the patient’s good clinical condition despite a high treatment burden, and the treating physician’s hope that the GvL effect after haploidentical HSCT might overcome a high tumor load at transplantation. Considering the two CloEC responders (both with BCP-ALL), one was transplanted already after the first course of CloEC despite still having a high MRD (> 1%) before transplantation. The other non-responder was in complete morphological remission but had 1 x 10-4 positive MRD after the first course and undetectable MRD after the second course; this is the only long-term survivor in our cohort. The children with AML and T-ALL did not respond and progressed, although one of them could be salvaged by changing chemotherapy from CloEC to FLADx; this patient was finally transplanted in morphological remission but died in remission of AdV hepatitis. Our results suggest that CloEC should probably not be repeated in patients showing no response after the first course, whereas repeating CloEC may further decrease the pre-transplant tumor load in responding patients.
No deaths related to CloEC toxicity were observed, and, in all children who died before being transplanted, death was due to progressive disease. However, the toxicity of the courses of CloEC was significant and mainly related to the profound immuno- and myelosuppression. Still, it is plausible that, together, both CloEC and previous treatments (including allogeneic HSCT in two patients) contributed to the immunosuppressive state of the patients. Contrary to some previous observations,6,17 we found no significant liver toxicity, neither during induction phase nor after transplantation. In all cases, the timing of the transplantation could be adjusted to the changing situations of the children. There was no unnecessary delay in transplantation due to logistical difficulties, reflecting the flexibility of this transplant modality. Ex-vivo T-cell depletion effectively prevented clinically relevant GvHD, even though the pharmacological prophylaxis was limited, MMF based, and given only during the first 4 weeks after transplantation.
In conclusion, the combination of clofarabine, etoposide, and cyclophosphamide did not seem to be a universal treatment for r/r AL: the only children responding to CloEC in our study were those diagnosed with BCP-ALL. Nevertheless, this particular group of patients may be offered salvage immunotherapy with either bi-specific T-cell engaging CD19 antibodies or chimeric antigen receptor (CAR) T-cells, an approach that seems more reasonable than exposing such patients to CloEC. Haploidentical HSCT remains a viable option in cases in which flexibility in planning the transplantation essential, although further studies are needed to assess the anti-leukemic potential of such HSCT.