Re-emerging Philadelphia chromosome-positive acute leukaemia more than 20 years after allogeneic haematopoietic stem cell transplantation
A recent study showed that intensive chemotherapy in combination with imatinib without allogeneic haematopoietic stem cell transplantation (allo-HSCT) improved the outcome of children with Philadelphia chromosome positive acute lymphoblastic leukaemia (Ph+ALL) (Schultz et al, 2009), although allo-HSCT is still an important treatment modality for Ph+ALL (Biondi et al, 2012). A few studies have reported the long-term outcome of allo-HSCT for children with Ph+ALL (Burke et al, 2009; Fagioli et al, 2012). However, it is unclear when clinicians can safely consider children as cured after allo-HSCT for Ph+ALL.
We retrospectively analysed 22 children with Ph+ acute leukaemia (AL) who consecutively received allo-HSCT at the Department of Paediatrics, National Kyushu Cancer Centre between 1987 and 2011. Characteristics of the patients are shown in Table 1. Twenty-one children were diagnosed with ALL. One child had acute myeloid leukaemia (AML) and because his AML cells had the B cell immunophenotype, we considered that his Ph+AML was biologically similar to Ph+ALL and included him in this analysis (Okamura et al, 1988).
Table 1. Patient characteristics and outcome
|Median age (range), years||6 (2–16)|
|Median WBC at diagnosis (range), × 109/l||33·8 (4·2–388·2)|
|Disease status at HSCT|
|Donor and stem cell source|
|TBI + CA + CPA||15|
|TBI + others||7|
|CsA ± MTX||11|
|Tac + MTX||10|
|aGVHD (Grade)|| |
Table 1 shows the outcome and prevalence of graft-versus-host disease (GVHD) in this cohort. Two children failed to engraft because of early death due to sepsis and autologous haematological recovery. The overall cumulative incidence of transplantation-related mortality was 27% (95% confidential interval [CI], 6–45). The median follow-up time after allo-HSCT for all 22 patients was 134 months (range, 8–305 months). The 10-year overall survival and event free survival probability was 52·4% (95% CI, 29·3–75·5) and 51·5% (95% CI, 29·6–73·5), respectively. Unexpectedly, we experienced two patients who relapsed more than 20 years after allo-HSCT, as described below (Table 2).
Table 2. Clinical course of 2 cases
|Cytogenetic analysis||46, XY, 9p-, t(9q+; 22q-)||46, XX, t(9q+; 22q-)|
|Molecular analysis (RT-PCR)||minor BCR-ABL1||minor BCR-ABL1|
|pre-HSCT Treatment||Chemo, CrSpXRT for CNS relapse at 2 months||Chemo, CrXRT|
|Months from diagnosis||3||8|
|Disease status at HSCT||2nd CR||1st CR|
|Donor||Matched sibling||Matched sibling|
|Stem cell source||BM||BM|
|Conditioning||TBI + CA + CPA + L-asp||TBI + CPA|
|GVHD acute/chronic||grade IV/none||none/none|
|Relapse after HSCT (months from HSCT)||BM (245)||BM (264)|
|Outcome after relapse (months from 1st HSCT)||died of leukaemia (262)||2nd CR (280)|
A 4-year-old boy was diagnosed with Ph+AML with B-precursor phenotype (positive for CD10, CD13, CD19, and TdT) because cytogenetic analysis showed 46, XY, 9q-, t(9q+; 22q-). Eight weeks after the diagnosis he had an overt central nervous system relapse, which was successfully treated with intrathecal therapy and craniospinal radiation therapy. Three months after the diagnosis, he underwent bone marrow transplantation (BMT) from his human leucocyte antigen (HLA) identical brother when in second complete remission (CR). The conditioning regimen consisted of total body irradiation (TBI) (10 Gy), cytarabine (12 g/m2), L-asparaginase (12 000 μ/m2) and cyclophosphamide (120 mg/kg). Following BMT, he had no major complications until 245 months after allo-HSCT, when he presented with general fatigue. A BM aspiration revealed leukaemic cells, which were positive for CD13 and CD33, and negative for CD10 and CD19. Chimerism analysis showed leukaemic cells of recipient origin. Cytogenetic analysis showed complicated chromosome abnormalities including t(9;22)(q34;q11). The patient was diagnosed with recurrent Ph+AML. He died of progressive disease despite undergoing a second BMT.
A 13-year-old girl was diagnosed with Ph+ALL because of 46, XX, t(9q+; 22q-). Six months after the diagnosis of ALL, she underwent BMT from her HLA-identical sister. The conditioning regimen consisted of TBI (13·2 Gy) and cyclophosphamide (120 mg/kg). Although she maintained haematological remission, reverse transcription polymerase chain reaction (RT-PCR) of BM cells continued to be positive for the minor BCR/ABL1 transcript for 156 months after allo-HSCT. At 28 years old (164 months after allo-HSCT), she gave birth to a healthy girl without complications. After delivery, minor BCR/ABL1 in the BM was not detected by RT-PCR at 170 months and 196 months after HSCT. At 264 months after allo-HSCT, she had anaemia and BM aspiration showed leukaemic cells, which were positive for the Ph-chromosome. Chimerism analysis showed completely recipient type. She was diagnosed with relapsed Ph+ALL. She has remained in CR since undergoing peripheral blood stem cell transplantation from the same donor. The daughter of the patient was growing normally at 10 years old.
We retrospectively analysed the long-term outcome for children with Ph+AL who underwent allo-HSCT at a single institution and found that our results were similar to a recent report (Fagioli et al, 2012). Unexpectedly, we experienced two patients with a considerably late relapse beyond 20 years after allo-HSCT. There are no reports of a late relapse more than 5 years after allo-HSCT (Laport et al, 2008). Our experience suggests that the observation for longer than 20 years is not sufficient for concluding that patients with Ph+ALL are cured.
We considered that our patients had a relapse of Ph+AL based on chimerism and cytogenetic analyses. Mullighan et al (2008) performed lesion-specific backtracking studies, which showed that the relapse clone exists as a minor subclone within the diagnostic sample before the initiation of therapy. Therefore, as observed in Case 1, the surface marker and chromosomal abnormality might be different at diagnosis and relapse. However, we could not identify the BCR-ABL1 genomic breakpoint region by molecular techniques, including genomic fusion sequencing (Isoda et al, 2009). Genomic breakpoints need to be estimated for distinguishing a late relapse from the emergence of new leukaemia.
In Case 2, minor BCR/ABL1 continued to be detected by RT-PCR until 176 months and then disappeared. This phenomenon must be due to a graft-versus-leukaemia (GVL) effect. It was reported that GVL in Ph+ALL was more important than that in other types of ALL (Espérou et al, 2003). Therefore, GVL might take 176 months to eradicate minimal residual leukaemic cells. However, if some triggers void the GVL effect, a haematological relapse might consequently be induced.
Case 2 gave a birth to a healthy girl when she had a detectable minor BCR/ABL1 transcript in BM. Recently, genomic fusion sequencing was used to demonstrate the materno-fetal transfer of Ph+leukaemic cells (Isoda et al, 2009). Because of immunosurveillance by the allorecognition of HLA, materno-fetal transmission of leukaemia is generally extremely rare. However, we need to observe our patient's daughter carefully because leukaemic cells might have already been transmitted to the daughter's body owing to the loss of HLA in leukaemic cells (Isoda et al, 2009).
In summary, we experienced two patients who relapsed more than 20 years after allo-HSCT. Therefore, we cannot conclude that patients with Ph+ALL are cured, even when they maintain a CR for a long time after allo-HSCT.
YK, JO and JI analysed the data, and wrote the manuscript; RF, KN, NI, MN, TS, and YA analysed and interpreted data. And all authors reviewed the final manuscript.
Conflict of interest disclosure
The authors declare no competing financial interests.