Allogeneic Hematopoietic Cell Transplantation with Full-Intensity Conditioning for Adult Acute Lymphoblastic Leukemia: Results from a Single Center 1998–2006

Abstract

A retrospective analysis identified 161 consecutive adults with acute lymphoblastic leukemia (ALL) who underwent allogeneic hematopoietic cell transplantation (HCT) with full-intensity (myeloablative) conditioning between 1998 and 2006. Median patient age was 36.1 years. Seventy-six patients were in 1^st remission; 85 were in ≥2^nd remission or in relapse. Fifty-nine patients had Ph⁺ ALL. One hundred fifty-nine patients received chemotherapy plus TBI for conditioning. Graft-versus-host disease (GVHD) prophylaxis included a calcineurin inhibitor plus methotrexate or mycophenolate mofetil. Sixty donors were related and 101 were unrelated. One hundred ten patients received G-CSF stimulated peripheral blood, 47 received bone marrow and four received cord blood as the stem cell source. Fifty-five patients have relapsed at a median of 231 days after transplantation. The estimated five year probabilities of relapse-free survival, relapse, and non-relapse mortality are 47%, 30%, and 29%, respectively. By multivariable analyses, transplantation in 1^st remission was the most important predictor of transplant success. Pretransplant evidence of minimal residual disease, especially as detected by flow cytometric analysis, was associated with both lower overall survival and relapse-free survival. Compared to a similar cohort of patients transplanted from 1990–1997, overall survival is similar for patients transplanted in 1^st remission, with lower non-relapse mortality being offset by higher rates of relapse in patients transplanted more recently.

INTRODUCTION

For adults with ALL, the indications for, and timing of, allogeneic hematopoietic cell transplantation (HCT) continue to be debated.[1–3] Over the past 20–25 years, refinements in conditioning regimens, GVHD prophylaxis and treatment, donor selection and supportive care have been made in an effort to improve transplant outcomes. The methodology used to define minimal residual disease (MRD) continues to evolve with sensitive flow cytometric methods and PCR monitoring of molecular markers redefining relapse.[4,5] Criteria for classifying chronic GVHD also have been revised and await validation.[6] Given these modifications over time, we performed a retrospective analysis of consecutive adult ALL patients who underwent allogeneic HCT with full-intensity conditioning between 1998 and 2006. Multivariable analyses were performed, focusing on clinical factors, as well as indicators of MRD, for their association with transplant outcome. These current results were then compared to a similar cohort of patients transplanted between 1990 and 1997, the results of which were published previously.[7]

MATERIALS AND METHODS

Patient and disease characteristics

Patients included in this analysis had a diagnosis of ALL, were ≥18 years of age, received a full-intensity conditioning regimen and underwent HCT between January 1, 1998 and December 31, 2006. Written informed consent was obtained from all patients and donors using forms approved by the Institutional Review Board of the Fred Hutchinson Cancer Research Center.

Pretransplant patient characteristics are summarized in Table 1. Median age was 36.1 (range, 18.1–61.8) yrs. Patients were classified as having B- or T-cell ALL based on immunophenotyping. B cell maturity at diagnosis was characterized as immature or mature: Immature B cells expressed TdT and/or CD34 and/or CD10; mature B cells expressed CD20, CD22, or SIg and did not express any immature markers.[8] Myeloid markers were defined as the expression of CD13 or CD33. Cytogenetic studies were performed on bone marrow (BM) or peripheral blood cells. Karyotypes were grouped as normal, Philadelphia chromosome positive (Ph⁺⁾, other unfavorable abnormalities (−7, +8, or 11q23 rearrangement) and miscellaneous abnormalities.[9] Patients were considered to be Ph⁺ if conventional cytogenetics were positive for t(9;22) or fluorescence in situ hybridization (FISH) studies were positive for the BCR/ABL rearrangement. There were too few PCR studies done pretransplant to be informative. Remission was defined as <5% blasts by morphology in marrow specimens that were >20% cellular.

Table 1.

Patient and Donor Characteristics

Characteristic	Number*	%
No. of patients	161
Median patient age, yr (range)	36.1 (18.1–61.8)
≤35 years	75	47
>35 years	86	53
Donor/recipient sex†
M/M	54	33
M/F	28	17
F/F	34	21
F/M	43	27
F+M/M	1	1
M+M/F	1	1
Immunophenotype
B cell, immature	137	85
B cell, maturity unknown	3	2
T cell	17	10
Unknown	4	2
Cytogenetics
Normal	35	22
Ph+‡	59	37
Other unfavorable§	15	9
Miscellaneous#	39	24
Unknown	13	8
Median WBC × 10−9/L at diagnosis, (range) [N=154]	16.5 (0.4–683)
Extramedullary disease**
CNS only	16	10
CNS + other sites	3	2
Testis	3	2
Other	2	1
None	137	85
Median disease duration, mo (range)	8.6 (2.4–216.3)
Disease status at transplant
1st Remission	76	47
≥2nd Remission	57	35
Relapse	28	18
MRD pretransplant (59 remission pts.)††
Karyotype	35	59
Abnormal	22	37
Normal	2	4
N.D.
Flow cytometry	26	44
Abnormal	33	56
Normal
FISH	4	7
Abnormal	21	36
Normal	34	57
N.D.
Median donor age, yr (range)§§	35 (4–65)
Donor/recipient HLA compatibility	55	34
Related/matched	5	3
Related/mismatched	62	39
Unrelated/matched	39	24
Unrelated/mismatched
Donor/recipient CMV serology
−/−	55	34
−/+	36	23
+/−	21	13
+/+	49	30

^*Number of patients unless otherwise specified

^†Of 4 cord blood transplants, 2 were single unit transplants and 2 were double unit transplants

^‡Ph⁺ includes patients with a t(9;22) only, t(9;22) plus additional clonal abnormalities, or BCR/ABL positive by FISH analysis

^§Includes −7, +8 or 11q23 rearrangement

^#Other clonal abnormalities

^**Extramedullary disease occurring before or at the time of transplant

^††MRD = minimal residual disease, N.D. = no data. Pretransplant studies = studies done proximate to conditioning regimen. Eleven of the 59 patients had two abnormal markers (flow cytometry +abnormal karyotype or abnormal FISH)

^§§N = 158 (cord blood transplant donors excluded)

Donor selection and stem cell collection

Donor characteristics are listed in Table 1. Median donor age, excluding the 4 cord blood donors, was 35 (range, 4–65) yrs. HLA compatibility was determined by medium resolution molecular methods for HLA-A, -B, -C and DQB1, and by DNA sequencing methods for DRB1. Fifty-five patients were transplanted from HLA-matched, related donors; 5 from HLA-mismatched, related donors; 62 from HLA-matched, unrelated donors; and 39 from HLA-mismatched, unrelated donors. The stem cell source was BM in 47 patients, G-CSF stimulated peripheral blood stem cells (PBSC) in 110 patients, and umbilical cord blood in 4 patients.

Transplant regimens

Table 2 summarizes the preparative regimens, stem cell doses, and posttransplant immunosuppression. Total body irradiation (TBI) was delivered by dual opposing cobalt sources for patients transplanted through mid May of 2000 (n=38) and by linear accelerator (18 MV photons) with lung shielding for subsequent patients (n=121). Male patients also received a 400 cGy testicular boost. The majority of patients (n=109) received cyclophosphamide (CY) plus 12.0 Gy fractionated TBI. Eighty-seven of the 109 patients received CY dosed at 60 mg/kg/day IV for 2 days and the remaining 22 patients received one dose of CY at 45mg/kg IV followed by a second dose that was adjusted, based on therapeutic drug monitoring after the first dose.[10] The mean second CY dose for the 22 patients was 66 mg/kg. All patients received at least 6 doses of prophylactic intrathecal methotrexate (12 mg/dose), with two doses administered pretransplant and four posttransplant, beginning on day 30 after transplant and then biweekly thereafter. Eleven patients with Ph⁺ disease participated in a posttransplant trial of imatinib. Eighty-four percent of patients received either cyclosporine (CSP) and methotrexate (MTX) (n=99) or tacrolimus (FK506) and MTX (n=37) for GVHD prophylaxis.[11–13]

Table 2.

Transplant Regimens and Outcome Data

Parameter	Number*	%
Preparative regimen†
CY, 12.0 Gy TBI	109	68
CY, 13.2 Gy TBI	45	28
VP16, 12–13.2 Gy TBI	4	2
Other	3	2
Median stem cell dose (range)‡
BM × 10−8 TNC/kg (n=42)	3.08 (0.89–9.01)
PBSC × 10−6 CD34+ cells/kg (n=109)	8.08 (2.87–32.15)
Cord blood × 10−7 TNC/kg (n=4)	0.42 (0.2 – 0.5)
GVHD prophylaxis§
CSP + MTX	99	61
FK506 + MTX	37	23
FK506 + MMF	7	4
CSP + MMF	8	5
CSP + MTX + other	4	3
MTX, sirolimus, FK506	5	3
None	1	1
AGVHD (n=152)
Grades 0-I	25	17
Grades II–IV	127	83
Median day onset Gr. II–IV aGVHD (range)	21 (6–102)
Chronic GVHD	80
Relapse (N=55)
Median day of relapse (range)	231 (19–2298)
Survival
Patients alive	63	39
Patients dead	98	61
Causes of Death# (N=98)
Relapse	52	53
DAD/ARDS	19	19
Infection	13	13
CGVHD +/− infection	9	9
AGVHD +/− infection	3	3
Graft failure + infection	2	2

^*Number of patients unless otherwise specified

^†CY = cyclophosphamide, TBI = total body irradiation, VP16 = etoposide

^‡TNC = total nucleated cells. The stem cell dose was not available for 5 patients who received BM and 1 patient who received PBSC.

^§GVHD = graft-vs.-host disease, CSP = cyclosporine, MTX = methotrexate, FK506 = tacrolimus, MMF = mycophenolate mofetil

^#DAD/ARDS = diffuse alveolar damage/adult respiratory distress syndrome

Engraftment, GVHD and Quality of Life

Time to myeloid engraftment was defined as the first of three consecutive days with an absolute neutrophil count (ANC) ≥0.5 × 10⁹/L. Platelet engraftment was defined as the first of seven consecutive days with an untransfused platelet count ≥20 × 10⁹/L. Acute GVHD (aGVHD) was diagnosed and graded according to previously published criteria.[14] All patients who engrafted (n=153) were considered evaluable for aGVHD. Chronic GVHD (cGVHD) was defined by NIH consensus criteria with all included patients requiring systemic immunosuppressive therapy.[6] The NIH criteria were applied retrospectively; subcategories of cGVHD could not be assessed due to incomplete data.[15] Karnofsky scores were obtained from patients or their referring MDs as part of yearly long-term follow-up questionnaires.[16]

Supportive care

All red cell and platelet transfusions were irradiated. All patients who were cytomegalovirus (CMV) negative prior to transplantation received CMV “safe” blood products.[17] Routine antibacterial, antifungal and antiviral prophylaxis included levofloxacin until neutrophil engraftment, fluconazole for a minimum of 75 days after transplantation, and acyclovir or valacyclovir for one year after transplantation. Trimethoprim-sulfamethoxazole or dapsone for Pneumocystis jiroveci prophylaxis was administered pretransplant and after engraftment for a minimum of six months. Surveillance testing for CMV reactivation was done weekly for a minimum of 100 days after transplantation and ganciclovir or foscarnet therapy was begun for patients who demonstrated an increase in viral copy numbers.[17] G-CSF was not used on a routine basis, but was administered to neutropenic patients based on their clinical condition. Since 2003, all patients were also treated prophylactically with ursodiol for 90 days after HCT.[18,19]

Statistical methods

Outcomes evaluated in univariate analyses included death, non-relapse mortality (NRM), relapse and relapse-free survival (RFS). Factors evaluated in univariate analyses for association with outcomes included: year of transplant, patient age at transplant, patient sex, donor/patient sex, disease parameters (WBC at diagnosis, cytogenetic markers, leukemic phenotype, presence of myeloid markers, CSF involvement, extramedullary disease), minimal residual disease (MRD) pretransplant (morphologic, cytogenetic, or flow cytometric markers), donor/recipient CMV serology, donor/recipient HLA matching, stem cell source (BM, PBSC), stem cell dose, GVHD prophylaxis, acute and chronic GVHD.

Estimates of the probability of overall and RFS were obtained using the method of Kaplan and Meier.[20] NRM and relapse were summarized using cumulative incidence estimates, where relapse was regarded as a competing risk for NRM and NRM a competing risk for relapse.[21] Cox regression models were used to examine the association of various factors with the outcomes overall mortality, NRM, relapse, and cGVHD.[22] Logistic regression was used to do the same for the endpoint grades II–IV and grades III–IV aGVHD. Multivariable regression models were fit using stepwise regression with entry and exit p-values of 0.10. All two-sided p-values resulting from regression models were estimated using the Wald test, and no adjustments were made for multiple comparisons.

RESULTS

Engraftment

One hundred fifty-three patients (95%) achieved a sustained ANC of ≥0.5 × 10⁹/L at a median time of 18 (range, 3–35) days after transplantation. The remaining eight patients died early at a median time of 16.5 (range, 8–60) days. One hundred thirty-one patients achieved self-sustained platelet counts of ≥20 × 10⁹/L at a median time of 15 (range, 8–60) days. Two patients who did not achieve this level of platelet engraftment prior to day 100 are alive 774 and 796 days after transplantation. Twenty-two patients died between days 5 and 65 without platelet engraftment and six died >day 100 without evidence of platelet engraftment prior to discharge.

Regimen-related toxicity

The cumulative incidence of NRM prior to day 100 for all 161 patients was 19%. Non-relapse mortality at five years was significantly lower (p=.07) for patients transplanted in 1^st remission (21%) compared to those transplanted in ≥2^nd remission (35%) or in relapse (46%). (Figure 1)

Figure 1.

Cumulative incidence of non-relapse mortality (NRM) for 161 adult patients with ALL according to disease status at transplant. The cumulative incidence of NRM at 5 years is 21% for patients in 1st remission, 35% for patients in ≥2nd remission, and 46% for patients in relapse.

Graft-vs.-host disease

One hundred twenty-seven of 152 patients (84%) who were graded for aGVHD developed grades II–IV aGVHD at a median time of 21 (range, 6–102) days. Eight of the nine patients who were not graded died ≤day 35; the 9^th patient who had no documented aGVHD assessment, died at day 204. The high incidence of acute GVHD at our center has been previously noted and associated with a high diagnostic sensitivity and increased awareness of upper gut GVHD.[23] Of 125 patients who were evaluable for cGVHD, 80 (64%) developed cGVHD by NIH consensus criteria.

Relapse

Fifty-five patients relapsed at a median time of 231 (range, 19–2298) days. Seventeen patients relapsed more than one year after transplant with the latest relapse occurring at 6.3 years. The cumulative incidence of relapse at 5 years for all patients was 35%. The probability of relapse was greater for those patients transplanted in relapse (40%) compared to those transplanted in 1^st remission (32%) or ≥2^nd remission (37%) (Figure 2).

Figure 2.

Cumulative incidence of relapse for 161 adult patients with ALL according to disease status at transplant. The probability of relapse at 5 years after transplant is 32% for patients in 1st remission, 37% for patients in ≥2nd remission, and 40% for patients in relapse.

Survival and Causes of Death

Sixty-three patients are alive and 98 have died. Actuarial survival at 5 years was 38% for all 161 patients. Survival and RFS were significantly associated with disease status at transplant, with patients transplanted in first remission having a 5-year RFS of 47% (Figure 3). Survival for patients with Ph⁺ ALL was comparable to that of patients with normal cytogenetics (42% versus 44% at 5 years, respectively, Figure 4). Of 32 patients who survived >5 years, self-assessment of performance was available for 28 patients at varying dates of last contact. The median Karnofsky score was 90% (range 50–100%) at a median time of 7 (range 2.1–11.6) years after HCT. MD-assigned scores were available for 25 of these 32 patients. The median MD score was 100% (range 70–100%) at a median time of 6.2 (range=1.5–11.5) years after HCT.

Figure 3.

Relapse-free survival for 161 adult patients with ALL according to disease status at transplant. RFS at 5 years after transplant is 47% for patients in 1st remission, 28% for patients in ≥2nd remission, and 13% for patients in relapse.

Figure 4.

Actuarial survival for 148 adult patients with ALL based on cytogenetic markers of disease. Survival at 5 years is 44% for patients with normal cytogenetics, 42% for Ph⁺ patients, 13% for patients with other unfavorable cytogenetics and 33% for miscellaneous clonal abnormalities.

The most frequent cause of death was recurrent ALL (n=52). Other causes of death are summarized in Table 2.

Prognostic factors

Univariate Analyses

Table S1 (available on the Biology of Blood and Marrow Transplantation website) details the results of the univariate analyses. Disease status at transplant was statistically significantly associated with all endpoints, with patients transplanted in 1^st remission having better survival and RFS, lower NRM and lower relapse rates. A more recent year of transplant was associated with both improved survival and RFS, and a lower relapse rate. There was a trend for opposite sex donor/recipient pairs to adversely affect survival and relapse rates. Patient age, sex, WBC at diagnosis (B-ALL), Ph+ cytogenetics, immunophenotype, myeloid markers (B-ALL patients), stem cell source, marrow or PBSC cell dose, donor/recipient HLA typing and GVHD prophylaxis were not statistically significantly correlated with outcome.

All markers of active disease pretransplant were strongly associated with decreased survival, and the majority of these markers correlated with decreased RFS and increased NRM. Persistent disease pretransplant, as documented by morphology, cytogenetics, or flow cytometry, also was statistically significantly associated with a higher risk of relapse after transplant with flow cytometry having the strongest association (p=0.009).

Severe (grades III–IV) aGVHD, modeled as a time-dependent covariable, was associated with a 2.6 fold increase in mortality and a >7-fold increase in NRM compared to patients who did not develop acute GVHD. While grades III–IV aGVHD was associated with a lower relapse rate, patients with severe aGVHD often die early after transplantation before they are at risk of relapse. Chronic GVHD alone or together with grades II–IV acute GVHD was not significantly associated with any of the four endpoints evaluated.

Multivariable Analyses

Results of the multivariable Cox regression analyses are seen in Table 3. Transplantation in 1^st remission remained significantly associated with increased survival (p=0.005) and RFS (p=0.06), and decreased NRM (p=0.01). A more recent date of transplant was also associated with a lower relapse rate (p=0.01) and improved RFS (p=0.05). Mismatched donor/recipient CMV serology pretransplant (−/+ or +/−) correlated with increased non-relapse mortality (p=0.005) and decreased RFS (p=0.07).

Table 3.

Multivariable analyses

Endpoint	Variable	Hazard ratio	P value	95% C. I.
Overall mortality	1st Remission	0.55	0.007	(0.36–0.85)
	Cytometric remission pretransplant	0.55	0.005	(0.36–0.83)
RFS	1st Remission	0.67	0.06	(0.43–1.02)
	Cytometric remission pretransplant	0.55	0.006	(0.37–0.84)
	Donor/recipient CMV serology mismatched pretransplant	1.46	0.07	(0.98–2.19)
	Year of transplant	0.93	0.05	(0.86–1.00)
Relapse	Cytometric remission pretransplant	0.46	0.005	(0.26–0.79)
	Year of transplant	0.88	0.01	(0.79–0.98)
Nonrelapse mortality	1st Remission	0.46	0.01	(0.25–0.84)
	Donor/recipient CMV serology mismatched pretransplant	2.28	0.005	(1.28–4.05)

Relapse was most strongly associated with MRD pretransplant as assessed by flow cytometry (p=0.005). If “remission” is re-defined as the absence of cytogenetic, FISH or flow cytometric evidence of disease, in addition to a marrow that was >20% cellular with <5% morphologically normal blasts, actuarial survival at 5 years for the 74 patients transplanted in any remission is 52% for patients without MRD and 29% for patients with MRD (P=0.08). The corresponding probabilities of relapse are 27% (no MRD) and 44% (with MRD) (p=0.06).

DISCUSSION

In 2003, we published the results of 182 consecutive adult ALL patients who underwent allogeneic HCT with full-intensity conditioning from January 1990 through December 1997.[7] In that study factors most significantly associated with improved survival and RFS were younger patient age and being in first remission at the time of transplantation. The predictive value of markers of MRD was not examined. At five years after transplantation, relapse-free survival, non-relapse mortality and the probability of relapse for patients transplanted in first remission were 43%, 42% and 15%, respectively.

There are several significant differences between the current cohort of 161 patients and our previously published data. In the current study (1) median patient age is greater (35.3 years vs. 29.4 years, p=.0004), (2) median disease duration from diagnosis to transplantation is less (8.5 months vs. 13.3 months, p=.05) and (3) the number of patients transplanted in first remission is higher (47% vs. 23%, p<.0001). The majority of the more recent transplants utilized unrelated donors, rather than HLA-matched, related donors (p<.0001) and the criteria by which these suitable alternative donors were identified have been refined.[24]

Transplant regimens have also changed over time. While the majority of our patients continued to receive a CY/TBI conditioning regimen, the method of delivering TBI has changed from dual Cobalt sources to a linear accelerator with lung shielding. The source of most stem cell infusions is now G-CSF mobilized peripheral blood rather than BM. For Ph⁺ ALL patients, tyrosine kinase inhibitor (TKI) therapy is being incorporated both pre and posttransplant.[25–27]

Supportive care has improved with the availability of more effective antiviral and antifungal agents and more sensitive techniques to detect reactivation of infection, e.g., CMV PCR assays and serum galactomannan levels.[17,28] Beginning in 2003, ursodiol was added as a liver “protectant” and has resulted in a significant decrease in posttransplant hepatic toxicity, including sinusoidal obstructive syndrome and hepatic GVHD.[18,19]

Despite these changes, 5-year RFS in our patient population transplanted in 1^st remission is similar between 1990–1997 and 1998–2006 (43 vs. 47%, respectively). While NRM has decreased (42% vs. 21%), the probability of relapse has increased (15% vs. 32%). Reasons for this increase in relapse rate are not totally understood, although decreased early deaths from non-relapse causes allows more patients to survive long enough to relapse.

Table 4 summarizes several studies of allogeneic transplantation for patients with ALL.[2,7,29–33] Most reports have focused on patients in 1^st or 2^nd CR who had HLA-identical sibling donors. Pediatric and adult patient outcomes are often reported together. Common to all studies is the use of hyperfractionated TBI, with total TBI doses of ~12–15.75 Gy, plus cyclophosphamide or etoposide. Except for our current patient cohort, BM has been the major source of stem cells. The best outcomes are those reported by the Stanford group in 2003 for patients in 1^st CR with HLA-matched, related donors and by the MRC/ECOG E2993 study in 2008 for patients in 1^st CR with “standard risk” disease and HLA-matched, related donors.[2,29] These patients all received TBI followed by etoposide, rather than cyclophosphamide. In the Stanford study, which included both pediatric and adult patients, overall survival, NRM, and relapse rates at 10 years were 64%, ~25%, and 15%, respectively. In the MRC/ECOG study, patients received uniform, intensive induction and consolidation therapy prior to transplant and survival data were calculated from the time of diagnosis. Overall survival and relapse rates for the standard risk (<35 years of age, WBC at diagnosis ≤30,000 × 10⁹/L for B-ALL, and Ph⁻) patients at five years were 62% and 24%, respectively, with a NRM at two years of 19.5%. Our patients received several different induction regimens for a varying number of cycles before referral for HCT. Only seven of our 1^st remission patients with HLA-matched siblings donors were standard risk by MRC/ECOG criteria since such patients were not routinely offered early transplantation.

Table 4.

Myeloablative, allogeneic HCT for ALL

Group	Study dates	N	Median age, y (range)	Remission Status	Donors	Conditioning Regimen	Stem cell Source	OS	DFS	NRM	Relapse
Stanford* Jamieson, et al.[29]	1987–2002	85	24 (0–48) CR1 10(3–42) CR2	CR1; CR2	MRD	13.2 Gy TBI/VP16	BM(78); PB(7)	66% 62%	64% 61%	~25% ~15%	15% 33%
CIBMTR/COH Marks, et al.[30]	1989–1998	502	<25(1–56)	CR1; CR2	MRD	Cy/TBI or VP16/TBI	BM (466); PB (36)	55–62% 40–72%	51–61% 33–62%	13–27% 10–20%	13–33% 17–46%
FHCRC Doney, et al.[7]	1990–1997	182	29 (18–57)	CR1; ≥CR2; relapse	MRD;URD;MM	Cy/12–15.75 Gy TBI	BM (158); PB (19); cord (1)		43% 23% 9%	42% 46% 9%	15% 30% 45%
9 European Ctrs. Kiehl, et al.[31]	1990–2002	221	31 (17–62)	CR1; CR2; Relapse 1° induction failure	MRD;URD	10–13.5 Gy TBI-based (82%) or Busulfan-based (18%)	BM (110); PB (111)		43% 23% 3–10% 36%
MRC/ECOG† Goldstone, et al.[2]	1993–2006	310	(15–54)	CR1, Ph−, “std” risk; CR1, Ph−, “high” risk	MRD	13.2 Gy TBI/VP16	NR	62% 41%		19.5% 35.85%	24% 37%
COH/Stanford LaPort, et al.[32]	1985–2005	79	36 (2–57)	CR1, Ph+ >CR1, Ph+	MRD	13.2 Gy TBI/VP16 (85%) or 13.2 Gy TBI/VP16/Cy (14%)	BM (43); PB (36)	54% 29%	48% 26%	54% 31%	28% 41%
CIBMTR‡ Marks, et al.[33]	1995–2006	1428	28 (16–62)	CR1; CR2	MRD:URD:MM	<13 Gy TBI (59%) ≥13 Gy TBI (29%) Cy-based (10%)	BM (817); PB(611)	51% 33%	49% 32%	33% (CR1+CR2)	21% 31%
Current study	1998–2006	161	36 (18–61)	CR1; ≥CR2; Relapse	MRD;URD;MM	Cy/12–13.2 Gy TBI	BM (47); PB (110); cord (4)		47% 28% 13%	21% 35% 46%	32% 37% 40%

^*Stanford outcome data are 10 year estimates.

Given the high relapse rates seen in our patients, regardless of disease status pretransplant, continued improvement in RFS may depend on both earlier identification and treatment of patients at high risk of relapse. Our data suggest that pretransplant flow cytometric evidence of MRD is a sensitive marker significantly associated with posttransplant relapse. Such a marker could be used to identify patients for whom additional conventional chemotherapy pretransplant or more intensive transplant preparative regimens is warranted. For Ph+ patients who are being monitored for MRD with PCR-based assays, the efficacy of adding TKI therapy during induction and consolidation therapy will help to answer this question.[25,26] For Ph⁻ patients with evidence of MRD pretransplant intensification of transplant regimens with approaches such as radiolabeled immunotherapy are being tested.[34]

Posttransplant monitoring for MRD may also prove important in guiding early posttransplant intervention for MRD, e.g., earlier manipulation of GVHD to maximize a graft-vs.-leukemia (GvL) effect. This approach has, however, been met with varying success. In 2004, Kiehl, et al., published outcome data for 264 adults with ALL who received full-intensity conditioning stem cell transplants.[31] A significant increase in disease-free survival (DFS) was seen in those patients who developed grades I–II aGVHD, compared to patients without aGVHD or with grades III–IV aGVHD. The effect of chronic GVHD was not addressed. The authors concluded there was a “strong” GvL effect. Also in 2004, Nordlander, et al., analyzed full-intensity conditioning HCT results in 199 patients with ALL, both pediatric and adult.[35] In univariate analyses, the presence of any aGVHD or chronic GVHD (not otherwise defined) was significantly associated with a decreased probability of relapse. In the corresponding multivariate analysis, the absence of chronic GVHD was associated with a 3.9 fold increased risk of relapse. In our current univariate analyses, only grades III–IV aGVHD resulted in a statistically significant decrease in the probability of relapse. This association, however, was not maintained in the multivariate analysis. Older analyses of multi-center data have also reported disparate results in the ability of acute or chronic GVHD to significantly decrease relapse rates.[36–39]

In Ph⁺ ALL patients early reinstitution of TKI therapy after transplantation is being studied as maintenance therapy with PCR-based assays used to assess response.[27] For other patients at high risk of relapse, however, options for either maintenance therapy or other therapeutic intervention at the first sign of MRD, are limited. Approaches under active investigation include re-institution of conventional chemotherapy, unconjugated or conjugated monoclonal antibodies and adoptive cellular therapies, including DLI or cytotoxic T-lymphocyte clones that recognize leukemia-associated antigens.[40] Early second transplants using nonmyeloablative regimens are also sometimes feasible.

In summary, posttransplant relapse remains the major cause of morbidity and mortality in our adult ALL population, regardless of disease status pretransplant. Historically, disease status has been defined by morphologic criteria only. With the advent of sensitive assays to detect MRD that correlate with relapse posttransplant, greater attention should be focused on both minimizing disease burden pretransplant as well as developing more aggressive approaches to treat MRD early in the posttransplant setting.

Supplementary Material

01

Table S1. Univariate analyses of factors associated with death, non-relapse mortality, relapse and relapse-free survival