Discussion
We describe our ten-year institutional experience in CAYA patients with myeloid malignancies undergoing allo-HCT on our pediatric service. This includes twenty-three consecutive patients of all risk categories who were conditioned with BU-FLU-MEL and received grafts from a variety of donor and stem cell sources. Thirty-nine percent of our patients were young adults (19 to 27 years). At a median follow-up of greater than 37 months, we have a remarkable OS of 100% and PFS of 95.5%. Irrespective of age, risk index, or treatment related disease, our OS and PFS is noticeably higher that previous reports of AML/MDS in CAYA undergoing allo-HCT.
There are several indications for allo-HCT in AML, including poor cytogenetics such as the presence of high allelic FLT3/ITD ratio, monosomy 5, monosomy 7, or an MLL/KMT2A rearrangement. A pediatric DRI has been developed to risk stratify children and adolescents with leukemia prior to transplant21. For AML, stratification into low, intermediate, high, and very high-risk groups is based upon cytogenetic risk, but also age and disease status complete remission (CR) and MRD. Using this stratification, 5-year leukemia free survival (LFS) was 78%, 53%, 40%, and 25% respectively with OS of 80%, 64%, 50%, and 33%21. Applying this DRI to our AML/MDS patient cohort, we had 2 low risk patients, 15 intermediate risk, 3 high, and 1 very high for which we would anticipate OS and PFS of 64% and 54% respectively at 36 months. Four of our patients had treatment related MDS/AML. A multicenter study of 273 pediatric patients with treatment-related MDS/AML undergoing MAC allo-HCT found the 5-year OS and EFS to be 49.9% and 41.2%25. A retrospective review of pediatric patients with poor cytogenetics who underwent allo-HCT (55 MSD and 55 unrelated donor [URD]), demonstrated a 5-year OS of 46% with MSD, 50% with URD26. These results are in line with what would be predicted based on the DRI. In a review of Center for International Blood and Bone Marrow Transplant Research (CIBMTR) data comparing allo-HCT outcomes of pediatric (<15 years) and adolescent and young adult (AYA) patients (15-40 years), similar survivals were noted27. For patients receiving MSD allo-HCT in first complete remission (CR1) the 5-year OS was 69% for pediatric patients compared to 52% for AYAs. The PFS was 65% and 51% respectively. For those receiving URD allo-HCT, decreased OS was seen with 34% for pediatric patients and 41% for AYAs. More recent CIBMTR data have shown improvement in 3-year probabilities of OS at 70% for AML patients <18 years receiving MSD in CR1 and 67% in those in CR2 or greater and 61% following MUD allo-HCT.
As there is still no consensus on optimal conditioning regimen for allo-HCT in CAYA with myeloid malignancies there have been several reports examining various regimens. A large multicenter study by Pediatric Disease Working Party of the European Group for Blood and Marrow Transplantation reported on pediatric patients with AML receiving TBI-CY (n=109), BU-CY (n=389) or BU-CY-MEL (n=133)10. The addition of MEL to BU-CY resulted in a reduced incidence of relapse of 14.7% when compared to 31.5% with BU-CY and 30% with TBI-CY as well as improved OS, 76.6% vs 64% with BU-CY, and 64.5% with TBI-CY as well as PFS 74.5 vs 58% vs 61.9% respectively with comparable NRM. There have been limited studies applying BU-FLU-MEL as conditioning in pediatric transplantation. We have previously reported on six patients receiving alternative donor allo-HCT conditioned with BU-FLU-MEL that was found to be safe and effective11. No grade 4 toxicities were seen, with grade 3 mucositis being the most common adverse effect. In a single institution study, pediatric (n=17) and adult patients with AML/MDS received BU-MEL-FLU and ATG followed by a T-depleted PBSC graft. The dose of BU and MEL was lower, 8-9.6 mg/kg and 140 mg/m2 when compared to our regimen of 12.8 mg/kg and 180 mg/m2 respectively for MSD and URD transplants28. The PFS was only 47.2% for pediatric patients. The incidence of grade III-IV aGvHD and cGvHD was low as one would expect from a T-cell depleted graft at 3.9% and 5.9% respectively with a relapse rate of 28%28. In a large multicenter retrospective report of 846 adult patients compared FLU-BU to FLU-BU-MEL conditioning in myeloid and lymphoid malignancies and found a significant improvement in the 5-year OS of 34.2% vs 30.1% and a decrease in relapse 30.8% vs 38%9. Similarly, a retrospective single-center review of 39 adult patients reported on their use of FLU-BU-MEL in myeloid and lymphoid malignancies with BM, PB and UCB as stem cell sources, with a 2-year OS of 62% for their entire cohort and 44% in their high risk to very high-risk subset. The authors concluded that this was a safe and successful regimen considering the highest risk group included many chemo-resistant patients with active disease at the time of transplant29.
We observed a significantly higher incidence of grades II-IV aGvHD in MSD and MUD PBSCT compared to patients undergoing haplo-BMT(Figure 1B). As expected with PBSC grafts, we saw a significantly increased cumulative incidence of cGvHD (1 mild, 2 moderate and 1 severe) and MUD allo-HCT (1 moderate, 1 severe) when compared to the anticipated lower occurrence in haplo-BMT with PT-CY, with none of the haplo-BMT patients developing cGvHD (Figure 1D) . Similarly, none of the three unrelated UCB transplant patients showed signs of cGvHD. Consequently, GRFS was superior in haplo-BMT patients when compared to our MSD/MUD PBSCT recipients (Figure 1F) . While aGvHD, and especially cGvHD, is well-known to occur more frequently with the use of PBSC grafts, the addition of MEL has also been associated with increased aGvHD when compared to non-MEL containing conditioning regimens. In an adult report in patients with hematologic malignancies receiving MUD or MSD PBSC, BM, or UCB grafts and either tacrolimus or cyclosporine GvHD prophylaxis, an increase in grade II-IV aGvHD was seen with BU-FLU-MEL compared to FLU-BU (52.6% vs40.8%) with a analogous protection from relapse (30.8% vs 39%)9. Moreover, pediatric AML patients receiving BU-CY-MEL had a higher incidence of grade III-IV aGvHD (17.6%) compared to those conditioned BU-CY alone (9.6%) with a parallel reduction in relapse of 14.7% and 31.5% respectively. Chronic GvHD was higher in PBSCs (34.9%) vs BM recipients (25.6%) as was observed in our series.
Our OS and PFS were similar, regardless of donor and stem cell source. In a recent single institution retrospective pediatric study, the outcomes of patients with hematologic malignancies (40% AML/MDS) undergoing MSD BMT (n=31) and MUD BMT (n-47) where compared to their haplo-PBSCT with PT-CY (n=26) recipients30. The OS (80.9%, 83.3% and 80.7%), PFS (66.5%, 70.2% and 73.8%) and the GRFS (62.1%, 44.6% and 61%) were not different between MSD, MUD and haplo respectively. However, cGvHD was significantly higher in MUD BMT recipients 31.7% compared to MSD BM 10% and haplo 9.2% despite use of PBSC grafts in the latter group. We have recently updated our experience of 31 CAYA patients with hematologic malignancies receiving haplo-BMT, which compares favorably to the above study with OS 85.6%, PFS 76.1% and GRFS 58.2% with our median follow-up of 32 months compared to only 12 months in the above study13.
Our current study along with our previous reports and those of others confirm that haplo-BMT in CAYA is a viable alternative to MSD and MUD allo-HCT12-15, 30, 31. This study expands our experience with myeloablative BU-FLU-MEL in allo-HCT for myeloid malignancies. As we are a smaller pediatric HCT center, this report is limited by the relatively small number of patients, their heterogeneous transplant characteristics in terms of stem cell source, type of donor and GvHD prophylaxis regimen used and the retrospective nature of our study. Nonetheless, the favorable outcomes remain striking considering the disease risk in this ethnically and racially diverse population with myeloid malignancies undergoing allo-HCT. Of caution is the unacceptably high aGvHD and cGvHD seen in patients receiving MSD and MUD PBSC transplants with MTX + CSA GvHD prophylaxis. It remains to be determined whether the evolving application of PT-CY in the non-haplo setting will offer better control of GvHD in CAYA PBSCT recipients following BU-FLU-MEL without increasing their risk of relapse.