Reduced Intensity Conditioning Followed by Related and Unrelated Allografts for Hematologic Malignancies: Expanded Analysis and Long Term Follow-Up

Abstract

Reduced intensity conditioning (RIC) extends the curative potential of allogeneic hematopoietic cell transplantation (HCT) to patients with hematologic malignancies unable to withstand myeloablative conditioning. We prospectively analyzed the outcomes of 292 consecutive patients, median age of 58 (range 19–75), with hematologic malignancies treated with a uniform RIC regimen of cyclophosphamide, fludarabine, and total body irradiation (TBI) (200 cGy) with or without anti-hymocyte globulin (ATG) and cyclosporine and mycophenolate mofetil graft versus host disease prophylaxis followed by allogeneic HCT at the University of Minnesota from 2002–2016. Probability of 5-year year overall survival (OS) was 78% for patients with indolent NHL, 53% for CML, 55% for Hodgkin Lymphoma, 40% for AML, 37% for MDS, 29% for myeloma, and only 14% for those with MPN. Corresponding outcomes for relapse were 0%, 13%, 53%, 37%, 39%, 75%, and 29% respectively. Disease risk index (DRI) predicted both survival and relapse with superior survival (64%) and lowest relapse (16%) in those with low risk score compared to 24% survival and 57% relapse in those with high/very high-risk scores. Recipient CMV positive serostatus was protective from relapse with the lowest rates in those also receiving a CMV positive donor graft (29%).The cumulative incidence of 2-year non-relapse mortality (NRM) was 26% and was lowest in those receiving a matched sibling graft at 21%, with low (21%) or intermediate (18%) HCT-CI and similar across age groups. The incidence of grade II-IV acute graft-versus-host disease (aGVHD) was 43% and grades III-IV was 27%: highest rates in those receiving a URD PBSC donor at 50%. Chronic GVHD at 1 year was 36%.

Future approaches incorporating alternative GVHD prophylaxis, particularly for URD PBSC grafts and targeted post-transplant anti-neoplastic therapies for those with high DRI risk are indicated to improve these outcomes.

INTRODUCTION

Allogeneic hematopoietic cell transplants (HCT) are potentially curative therapy for a wide range of hematologic malignancies. Advanced age, medical comorbidities, and prior treatment history can preclude the use of more toxic myeloablative conditioning and limit the applicability of this valuable therapy. Consequently, the development of reduced intensity conditioning (RIC) regimens expands the use of HCT and potentially limits the non-relapse mortality (NRM). Prior publications have shown that underlying disease type, disease stage at transplantation, comorbidity and the degree of conditioning intensity reduction impacts transplant outcomes.^1–3

We previously reported outcomes on the first 123 patients with hematologic malignancies treated at our institution with a consistent RIC platform of cyclophosphamide, fludarabine, and low dose total body irradiation (TBI) using sibling donors.³ We described a well-tolerated platform in an older patient population and noted superior outcomes in indolent lymphoma. We have completed protocol enrollment and now report outcomes on 292 patients with advanced hematologic malignancies treated in a prospective trial using both sibling and unrelated donors.

PATIENTS AND METHODS

Patients were treated at the University of Minnesota on this protocol between 2002–2017 according to protocol NCT00303719. Those transplanted through 2016 were were included in the analysis to allow for sufficient follow-up. Criteria for undergoing RIC allogeneic HCT from both related and unrelated donors and eligibility criteria were previously reported.³ In summary, eligible patients were ≤ 75 years old and received a transplant from a 5–6/6 HLA-matched related donor or an 8/8 allele level HLA matched unrelated donors; mismatched unrelated donors were excluded from the analysis due to low numbers (n=3). Patients < 55 years old were eligible if they had evidence of organ dysfunction, were heavily pre-treated, or had a recent fungal infection precluding myeloablative conditioning. Eligible diseases included acute leukemia in complete remission, CML (non-blast crisis), chemotherapy sensitive lymphoma, chronic lymphocytic leukemia or myeloma; MDS with < 5% blasts, and myeloproliferative neoplasms.

The conditioning regimen consisted of fludarabine, cyclophosphamide, TBI +/− ATG. Fludarabine was dosed 40mg/m² intravenously (IV) day −6 through day −2 until October 2009. A change to fixed dose fludarabine 30 mg/m2 daily on days −6 through days −2 thereafter was based on prior institutional pharmacokinetic analyses linking increased F-ara-A (fludarabine metabolite) levels with higher NRM and diminished survival. .⁴ Cyclophosphamide 50 mg/kg was given on Day −6 and a single dose of 200 cGy of total body radiation (TBI) on Day −1. Equine anti-thymocyte globulin (ATG) (dosed at 15mg/kg IV every 12 hours for six doses on days −6, −5, and −4 with methylprednisolone 1mg/kg as a premedication) was administered to those not exposed to combination chemotherapy within the preceding 6 months for related donors or 3 months for unrelated donors.

The primary graft source for related donors was mobilized peripheral blood stem cell (PBSC), whereas bone marrow was the preferred source for matched unrelated donors. Graft versus host disease (GVHD) prophylaxis included cyclosporine (CSA), targeting a trough level 200–400 ng/ml, and mycophenolate mofetil (MMF), beginning Day −3 until Day +30.³ MMF dosing transitioned from 2 grams/day to 3 grams/day in 2011 based on institutional data highlighting a need for higher MMF dosing to achieve an adequate AUC. The higher AUC translated into less aGVHD in cord blood transplants.^5–7 CSA continued through Day +100 and tapered at a rate of 10% per week if no evidence of GVHD. G-CSF 5 micrograms/kg was administered beginning Day +1 and continued until absolute neutrophil count (ANC) was > 2.5 × 10⁹/L for 2 consecutive days. Infectious prophylaxis included antibacterial, antifungal, and anti-viral therapies per institutional guidelines.

Disease risk index (DRI) scores were calculated and assigned retrospectively⁸ using the Center for International Bone Marrow Transplant Registry (CIBMTR) on-line calculator. Two independent reviewers (EW and NB) scored all patients and any discrepancies were resolved. Hematopoietic comorbidity scores (HCT-CI)⁹ were calculated and assigned prospectively for more recent transplants and were calculated retrospectively for earlier transplants.

This trial was a prospective clinical study reviewed and approved by the Masonic Cancer Center Protocol Review Committee and Human Subjects Institutional Review Board (IRB) at the University of Minnesota. All patients signed IRB approved informed consent in accordance with the Declaration of Helsinki.

Study Endpoints and Statistical Analysis:

We followed all patients longitudinally until death or last follow-up at the University of Minnesota using standardized collection procedures. The endpoints included neutrophil and platelet recovery defined the first day of an absolute neutrophil count (ANC) >0.5 × 10⁹/L for at least 3 consecutive days and a platelet count > 20 × 10⁹/L with no transfusion for at least 7 days, overall survival (OS), relapse, non-relapse mortality (NRM), grades II-IV and III-IV acute GVHD and chronic GVHD. Stopping rules were in place and defined as the SAE of >30% NRM at Day +100. Continuous monitoring for this endpoint ensued through the life of the trial.

Unadjusted estimates and 95% confidence intervals were calculated using Kaplan-Meier curves for OS.¹⁰ Statistical comparisons were completed by the log-rank test. The cumulative incidence function was used to estimate the probability for all other outcomes, considering non-event death as a competing risk for relapse, GVHD and engraftment, and relapse as a competing risk for NRM ¹¹ Statistical comparisons were completed by Gray’s test.

Cox regression was used to examine the independent effect of factors for OS.¹² Fine and Gray regression was used to examine the independent effect of factors on relapse, NRM, GVHD and engraftment.¹³ Factors considered in regression analyses were Fludarabine and MMF dosing combinations (Fludarabine 40 mg/m² +MMF 2 gm versus 30 mg/m² + 2 gm versus 30 mg/m²+ 3 gm), age (<50 versus 50–60 versus > 60), donor type (matched sibling versus mismatched sibling versus matched URD (marrow) versus matched URD (pbsc), use of ATG (yes versus no), HCT-CI (low risk versus intermediate risk versus high risk), gender (male versus female), prior autologous transplant (yes versus no), diagnosis, disease risk index (low versus intermediate versus high/very high), year of transplant and CMV serostatus (donor+/recipient+ versus donor+/recipient- versus donor-/recipient+ versus donor-/recipient-).

All reported P-values were 2-sided. All analyses were performed using SAS 9.4 (SAS Institute, Inc., Cary, NC) and R version 3.3.1. Outcomes and covariates were all pre-specified. All statistical tests were reported and there was no adjustment for multiple testing.

Results:

Patients:

A total 292 of 349 patients transplanted on our RIC protocol using related and unrelated donors from 2002 to 2016 were included in this analysis. Fifty-seven patients were excluded from the analysis for the following reasons: prior allogeneic transplant (n=8), transplanted during post chemotherapy marrow aplasia or post radio-labeled antibody (n=13), patients with prior natural killer cell therapeutic infusion (n=12), patients with mismatched unrelated donors (n=3), non-malignant patients (n=4), renal cell cancer (n=4), fraternal twin or cousin donor (n=4), pediatric age (n=2), and transplants in 2017 with insufficient follow up (n=7).

Table 1 summarizes the patient and donor characteristics, diseases, and transplant characteristics. The median age was 58 years (range 19–75 years) with a median follow-up of 8 years (interquartile range (IQR) 4–12 years). Median time from diagnosis to transplant was 13 months (IQR 2– 198 months), 61% were male, and 61% had a HCT-CI of 0–2. Eighteen percent had a prior autologous transplant and 30% received ATG with their conditioning. Sibling donors accounted for the majority of the transplants (HLA matched bone marrow 1%, matched PBSC 76%, mismatched PBSC 4%) with HLA matched unrelated marrow (13%) and matched unrelated PBSC (6%) the remainder. DRI was intermediate risk (65%) for the majority of patients followed by low risk (20%) and high/very high risk (14%).

Table 1.

Demographics/Characteristics.

	All Groups
N	292
Gender: Male	178(61%)
Age at Transplant
Median(Range), (IQR)	58(19–75), (51–63)
< 50	61
50–60	123
61–70	99
70+	9
Fludarabine: MMF Dosing Groups
Flu 40 mg/m2 + MMF 2 gram/Day	136 (47%)
Flu 30 mg/m2 + MMF 2 gram/Day	24 (8%)
Flu 30 mg/m2 + MMF 3 gram/Day	132 (45%)
Year of Transplant
2002–2005	70(24%)
2006–2010	88(30%)
2011–2016	134(46%)
Donor Type
Sibling Match (marrow)	2(1%)
Sibling Match (pbsc)	223(76%)
Sibling MM (pbsc)	12(4%)
URD Match (marrow)	37(13%)
URD Match (pbsc)	18(6%)
Prior Autologous:	54(18%)
ATG with Prep	89(30%)
Diagnosis Group
AML	81(28%)
MDS	47(16%)
Aggressive NHL	29(10%)
Indolent NHL	23(8%)
Hodgkins	21(7%)
CML	15(5%)
CLL	22(8%)
Multiple Myeloma	24(8%)
MP Neoplasms	14(5%)
ALL	16(5%)
Months from DX to TX
Median(Range), (IQR)	13 (2–198), (5–39)
DRI
Unknown	3(1%)
Low	57(20%)
Intermediate	190(65%)
High	39(13%)
Very High	3(1%)
Comorbidity (HCT-CI)
low risk	82(28%)
intermediate risk	95(33%)
high risk	112(38%)
Missing Data	3(1%)
Karnofsky
≤80	54(18%)
90	159(54%)
100	79(27%)
CMV Serostatus
D+R+	86(29%)
D+R−	35(12%)
D−R+	83(28%)
D-R−	88(30%)
Follow-Up (Reverse K-M)
Median (IQR)	8 years (4–12)

IQR-Interquartile range; MM-HLA mismatch; pbsc-peripheral blood stem cell; URD-unrelated donor; RIC-reduced intensity conditioning; Cy-Cytoxan; Flu-Fludarabine; TBI-Total body irradiation; ATG-antithymocyte globulin; CSA-Cyclosporine; MMF-Mycophenolate mofetil; AML-acute myeloid leukemia; MDS-Myelodysplasia; NHL-Non-hodgkin’s Lymphoma; CML-Chronic Myeloid Leukemia; CLL-Chronic Lymphoid Leukemia; MP Neoplasms – Myeloproliferative Neoplasms (3 CMML, 3 MDS/MPN, 5 myelofibrosis, 3 myeloproliferative Disease NOS) HCT-CI-Hematopoietic Cell Transplant Comorbidity Index; CMV-Cytomegalovirus; D-Donor; R-Recipient; K-M-Kaplan-Meier. Aggressive Lymphoma (Diffuse Large B Cell, T cell lymphoma, Mantle Cell Lymphoma, Burkitt Lymphoma, Lymphoblastic Lymphoma). Indolent Lymphoma (Follicular lymphoma, Waldenstroms).

Engraftment

Two hundred and eighty-nine out of 292 patients achieved neutrophil engraftment of > 500/mcL by Day +42 for an incidence of 99% (95% CI 98–100%). The median time to engraftment was 9 days (IQR 7–11 days). Platelet engraftment of 20 × 10⁹/L by 6 months was 91% (95% CI 84–98%) for the entire cohort. The median time to platelet recovery was 16 days (IQR 0–19 days). We observed a slightly lower incidence of platelet engraftment by 6 months in those with CML (67%, 95% CI 40–94%) and MPN (79%, 95% CI, 52–100%) compared to over 87% in all other disease groups. At Day +28 the median donor chimerism/engraftment was 98% (IQR 91–100) and 100% by Day +100 (IQR 96–100). There was no difference in chimerism based on DRI or donor source.

Survival

After a median follow-up of 8 years (range 0.8–15 years), 130 patients survived for a five year overall survival of 42% (95% CI, 36–48%). Five year survival was superior in those receiving matched sibling donor transplants (46%, 95% CI 39–52%) and was inferior in those using a matched URD PBSC (15%, 95% CI, 3–36%) or mismatched sibling PBSC (25%, 95% CI, 6–50%) p<0.01. The vast majority of deaths occurred within the first 2 years. Underlying diagnosis and DRI significantly influenced overall survival with superior survival in those with indolent NHL (78%, 95% CI 55–90%), HL (55%, 95% CI 31–74%) or CML (53%, 95% CI 25–74%) and those with low risk DRI (64%, 95% CI, 50–75%). Younger age was associated with superior survival with the best 5 year survival in those younger than 50 (56%, 95% CI, 43–68%) with similar survival in those aged 50–60 (41%, 95% CI 32–50%) and aged > 60 (34%, 95% CI, 25–44) p=0.04. Figure 1 details overall survival stratified by donor, DRI, and age.

Figure 1:

Survival by Donor Type Adjusted for DRI and Age

Cox regression analysis confirmed the independent predictive value of donor type, DRI, and age on survival outcomes. Conditioning/GVHD prophylaxis combinations (fludarabine/MMF dosing) had no effect on survival. In comparison to those transplanted with a matched sibling, those with a mismatched sibling PBSC donor (n=12) (HR 3.1 (95% CI, 1.5–6.2) or a matched PBSC URD (n=18) (HR 2.3 (95% CI, 1.3–4.0) ) or matched marrow URD (n=37) (HR 1.7 (95% CI, 1.0–2.6) had inferior outcomes as did those with an intermediate risk (n=190) (HR 1.9 (95% CI, 1.2–3.1) and high/very high risk DRI (n=42) (HR 2.9 (95% CI, 1.6–5.1) when compared to low risk patients. Table 2 and Table 3

Table 2.

Univariate Analysis.

		N	5 Year Relapse			5 Year Survival			2 Year NRM			aGVHD Grade II– IV
Factor	Strata		Estimat e	CI 95 %	P	Estimat e	CI 95 %	P	Estimat e	CI 95 %	P	Estimat e	CI 95 %	P
All Patients		29 2	38%	32– 44 %		42%	36– 48 %		26 %	20– 31 %		43%	37– 49 %
Donor Type	Sibling Match	22 5	39%	32– 46 %	0.75	46%	39– 52 %	<0.0 1	21 %	16– 27 %	<0.0 1	39%	32– 46 %	<0.0 1
	Sibling MM	12	33%	8– 41 %		25%	6– 50 %		50%	27– 78 %		45%	17– 74 %
	URD Match (Marrow)	37	38%	22– 55 %		38%	23– 54 %		30%	15– 46 %		54%	36– 72 %
	URD Match (PBSC)	18	22%	3– 41 %		15%	3– 36 %		56%	30– 81 %		67%	42– 92 %
Diagnosis	AML	81	37%	25– 48 %	<0.0 1	40%	28– 51 %	0.02	25%	15– 35 %	0.09	36%	25– 46 %	0.12
	MDS	47	39%	24– 54 %		37%	22– 51 %		28%	15– 42 %		40%	26– 55 %
	Aggressive NHL	29	35%	17– 53 %		41%	22– 59 %		18%	4– 32 %		48%	29–67 %
	Indolent NHL	23	0%			78%	55– 90 %		17%	2– 33 %		32%	13– 50 %
	Hodgkin	21	53%	30– 76 %		55%	31– 74 %		14%	0– 29 %		33%	13– 53 %
	CML	15	13%	0– 29 %		53%	25– 74 %		41%	16– 66 %		47%	21– 73 %
	CLL	22	42%	20– 64 %		44%	23– 64 %		27%	9– 46 %		68%	46– 91 %
	Multiple Myeloma	24	75%	52– 98 %		29%	13– 48 %		17%	2– 31 %		54%	33– 75 %
	MP Neoplasms	14	29%	6– 51 %		14%	2– 37 %		57%	29– 85 %		29%	5– 52 %
	ALL	16	46%	20– 73 %		33%	11– 57 %		31%	9– 53 %		63%	36– 89 %
HCT-CI	0 (Low)	82	38%	27– 49 %	0.68	46%	33– 57 %	0.46	21%	12– 30 %	<0.01	44%	33– 55 %	0.08
	1–2 (Intermediat e)	95	44%	33– 55 %		45%	34– 55 %		18%	10– 25 %		36%	26– 46 %
	3+ (High)	11 2	34%	24– 43 %		36%	27– 45 %		37%	27– 46 %		49%	39– 59 %
DRI	Low	57	16%	6– 25 %	<0.0 1	64%	50– 75 %	<0.0 1	23%	12– 34 %	0.26	36%	23– 49 %	0.70
	Intermediate	19 0	41%	33– 49 %		38%	31– 46 %		27%	21– 34 %		45%	37– 52 %
	High/Very High	42	57%	40– 74 %		24%	12– 38 %		24%	11– 37 %		43%	27– 58 %
Age	<50	61	43%	30– 57 %	0.56	56%	43– 68 %	0.04	18%	8– 28 %	0.23	52%	39– 66 %	0.20
	50–60	12 3	40%	31– 50 %		41%	32– 50 %		25%	17– 33 %		41%	31– 51 %
	>60	10 8	31%	22– 40 %		34%	25– 44 %		30%	21–40 %		40%	31– 49 %
Fludarabine:MM F Dosing	40 mg/m2: 2 gram/day	13 6	32%	21– 41 %	0.13	45%	46– 53 %	0.5	26%	19– 34 %	0.86	42%	34– 51 %	0.79
	30 mg/m2: 2 gram/day	24	50%	28– 72 %		46%	26– 64 %		21%	5– 37 %		50%	29– 71 %
	30 mg/m2: 3 gram/day	13 2	41%	32– 50 %		38%	28– 47 %		26%	18– 34 %		42%	34– 51 %

PBSC = peripheral blood stem cell; DRI = disease risk index; CMV = cytomegalovirus; HCT-CI = hematopoietic cell transplant comorbidity index; MMF = Mycophenolate Mofetil; MM = Mismatch; MP Neoplasm = Myloproliferative Neoplasm; NHL = Non-hodgkin lymphoma; AML = Acute Myeloid Leukemia; MDS = Myelodyspastic Syndrome; URD Unrelated Donor; CML = Chronic Myeloid Leukemia; CLL = Chronic lymphocytic leukemia; ALL = acute lymphocytic leukemia.

Table 3.

Multivariate Regression Analysis.

Outcome	Factor	HR (95% CI)	P value
5 Year OS	Donor Type*
	Sibling Matched	1.0 (ref)
	Sibling Mismatched	3.1 (1.5–6.2)	<0.01
	URD Matched (marrow)	1.7 (1.0–2.6)	0.03
	URD Matched (pbsc)	2.3 (1.3–4.0)	<0.01
	*-only sibling mismatched and URD are statistically significant in comparison to sibling matched. No other pairwise comparisons are significant
	DRI
	Low	1.0 (ref)
	Intermediate	1.9 (1.2–3.1)	0.01
	High/Very High	2.9 (1.6–5.1)	<0.01
	Age
	<50	1.0 (ref)
	50–60	1.7 (1.1–2.7)	0.02
	>60	1.8 (1.1–2.9)	0.02
5 Year Relapse	DRI*
	Low	1.0 (ref)
	Intermediate	2.9 (1.5–5.8)	<0.01
	High/Very High	4.9 (2.3–10.8)	<0.01
	*-high risk is statistically significantly different from Intermediate risk at the 0.05 level
	Recipient CMV
	Negative	1.0 (ref)
	Positive	0.6 (0.4–0.8)	<0.01
	Flu/MMF
	40 mg/m2, 2 gm/day	1.0 (ref)
	30 mg/m2, 2 gm/day	1.7 (1.0–3.1)	0.07
	30 mg/m2, 3 gm/day	1.5 (1.0–2.2)	0.056
2 Year NRM	Donor Type*
	Sibling Matched	1.0 (ref)
	Sibling Mismatched	3.9 (1.6–9.6)	<0.01
	URD Matched (marrow)	2.2 (1.0–4.7)	0.04
	URD Matched (pbsc)	3.0 (1.5–6.2)	<0.01
	*-only sibling mismatched and URD matched (pbsc) are statistically significant in comparison to sibling matched. No other pairwise comparisons are significant
	DRI		NS (0.81)
	Low	1.0 (ref)
	Intermediate	1.3 (0.6–2.5)
	High/Very High	1.1 (0.5–2.7)
	Age
	<50	1.0 (ref)
	50–60	1.9 (1.0–4.7)	0.06
	>60	2.2 (1.0–4.7)	0.04
	Comorbidity (HCT-CI)
	Low/Intermediate Risk	1.0 (ref)
	High Risk	2.4 (1.5–3.9)	<0.01
	Recipient CMV
	Negative	1.0 (ref)
	Positive	1.1 (0.7–1.8)	NS (0.65)
	Flu/MMF		NS (0.45)
	40 mg/m2, 2 gm/day	1.0 (ref)
	30 mg/m2, 2 gm/day	0.8 (0.3–2.3)
	30 mg/m2, 3 gm/day	0.7 (0.4–1.2)

OS = Overall Survival; PBSC = peripheral blood stem cell; HR = hazard ratio; DRI = disease risk index; CMV = cytomegalovirus; HCT-CI = hematopoietic cell transplant comorbidity index

Relapse

The cumulative incidence of relapse at 5 years for the entire cohort was 38% (95% CI, 32–44%). Donor type, DRI, disease subtype, and recipient CMV serostatus best predicted relapse. Risk of relapse was lowest in those with indolent NHL (0%) and CML 13% (95% CI, 0–29%) and highest in patients with multiple myeloma at 75% (95% CI, 52–98%) p <0.01. Likewise, relapse was lowest in those with low risk DRI (16%, 95% CI, 6–25%) compared to 57% (95% CI, 40–74%) in those with high/very high-risk DRI p < 0.01 (Figure 2). Recipient CMV status correlated with relapse risk with the lowest risk in patients who were CMV seropositive. (Figure 3)

Figure 2:

Relapse by DRI

Figure 3:

Relapse by CMV Serostatus

In regression analysis, DRI and recipient CMV serostatus were independently significant predictors of relapse. Fludarabine/MMF dosing did not impact relapse. Compared to low risk DRI, the risk of relapse for those with an intermediate risk DRI was 2.9 (95% CI, 1.5–5.8) and 4.9 (95% CI, 2.3–10.8) in the high/very high DRI risk group. Positive recipient CMV serostatus was associated with significantly less risk of relapse with a HR of 0.6 (95% CI, 0.4–0.8) p<0.01. (Table 3) Rates of CMV reactivation at Day +100 highly correlated with recipient CMV serostatus (D+/R+ (40%; 95% CI, 30–40%); D+/R− (9%, 95% CI, 1–18%); D−/R+ (49%; 95% CI, 38–60% ); D-R− (3%, 95% CI, 1–6%). Interestingly, CMV reactivation itself did not significantly impact relapse with HR of 1.2 (0.8–1.8, 95% CI) for those with reactivation.

Non-relapse mortality (NRM)

NRM at 2 years was 26% (95% CI, 20–31%) for the entire cohort. Univariate analysis revealed that donor type and HCT-CI influenced outcomes. Two year NRM was lowest in those with matched sibling donors (21%; 95% CI 16–27%) and low (21%; 95%CI 12–30%) and Intermediate HCT-CI (18%, 95% CI; 10–25%) and similar across age groups. (Table 2) We observed no impact of Fludarabine and MMF dosing on NRM. Regression analysis confirmed the significance of donor type and HCT-CI on NRM. The HR was 3.9 (95% CI, 1.6–9.6) for mismatched siblings, was 3 (95% CI, 1.5–6.2) for URD matched PBSC, and 2.2 (95% CI, 1–4.7) for URD matched marrow p<0.01. High risk HCT-CI had a HR of 2.4 (95% CI, 1.5–3.9) when compared to low/intermediate risk HCT-CI, p<0.01. Table 3

Cause of death was most often disease relapse (48%), but complications of acute GVHD (18%), infection (8%), chronic GVHD (6%), organ failure (6%) with graft failure, acute respiratory distress syndrome, new malignancy, hemorrhage, accidental death, encephalitis representing the others (14%). Cause of death differed by donor type. For matched PBSC URD transplants the majority of deaths were acute GVHD related (40%), infectious (20%), and disease relapse (20%). Comparatively, with matched sibling donor transplants the majority of deaths were disease relapse (52%) and acute GVHD related (18%).

GVHD

At 100 days, the cumulative incidence of acute GVHD (aGVHD) grades II-IV was 43% (95% CI 37–49%) and 27% (95% CI, 22–32%) for grade III-IV. The incidence of day +100 aGVHD were primarily influenced by donor source with highest rates of grade II-IV and grades III-IV in PBSC URD at 67% (95% CI 42–92%) and 50% (95% CI, 26–74%) respectively versus 37% (95% CI, 32–46%) and 24% (95% CI, 19–30%) in matched sibling transplants. In regression analysis, the risk of severe aGVHD grades III-IV was highest in URD PBSC with a HR of 2.9 (95% CI, 1.3–6.4) p=0.01 in comparison to matched sibling donor transplants. We found no other patient, disease or treatment variables (including fludarabine and MMF dosing) that significantly altered overall or severe aGVHD rates.

The cumulative incidence of chronic GVHD (cGVHD) at 2 years was 36% (95% CI, 30–42%) for the entire cohort. Only a few URD PBSC recipients survived to 2 years and thus this subset was not recognized to have a higher incidence of cGVHD.

Discussion:

This prospective trial using a uniform RIC and GVHD prophylaxis platform for patients with hematologic malignancies highlights superior outcomes in patients with indolent lymphomas and CML, younger patients, those with sibling donors, and those with lower DRI and lower HCT-CI and highlights need for improvement for those with high risk DRI and those utilizing unrelated PBSC donors.

Relapse remains a substantial problem compromising long-term success of HCT for hematologic malignancies. Our cohort of patients revealed high rates of relapse in patients with high risk DRI, those with multiple myeloma and Hodgkin lymphoma and modest rates for those with AML and MDS. Requiring stringent remission status prior to transplant is a crucial initial step towards diminishing relapse risk post-transplant, especially those with AML/MDS^14–16 and may be equally important for those with high risk DRI. Augmenting conditioning intensity using myeloablative approaches when possible ² or considering an intermediate conditioning approach based on disease subtype/DRI are potential approaches for up-front mitigation of relapse risk. Implementing post-HCT anti-neoplastic treatments with maintenance therapy may further reduce risk of relapse after transplant.¹⁶ Numerous approaches for post-transplant maintenance strategies are in development: 1) FLT-3 ITD+ AML with post-transplant TKI maintenance,¹⁷ 2) MDS with post-transplant azacitidine or decitabine maintenance;^18–19 3) Myeloma utilizing lenalidomide (iMID), proteasome inhibition maintenance (BMT CTN 1302), or iMID + DLI approaches.²⁰ Lastly, immune modulation post-HCT could diminish relapse risk. Based on recently published data highlighting successful IL-15 superagonist (ALT-803) induced NK cell and CD8+ T cell expansion in patients relapsing post allogeneic transplant, ²¹ we are now utilizing ALT-803 prophylaxis post RIC allogeneic transplant for patients with AML and MDS in attempts to minimize relapse risk. Combining pre-transplant disease status specific requirements, personalized conditioning intensity approaches, and post-HCT maintenance pathways for high relapse risk diseases will improve long-term outcomes.

Our RIC platform was well-tolerated in an older patient population with overall low NRM that was not impacted by age as supported by prior publications. ^22,23 Unfortunately, despite prior reports from our institution correlating diminished NRM with lower F-ara-A levels, decreasing the fixed dose of fludarabine to 30 mg/m2 in our study did not translate to diminished NRM. These findings suggest that diminishing NRM with fludarabine dosing would require true pharmacokinetic dosing on an individual patient basis. Those with high risk HCT-CI and those using a PBSC unrelated donor experienced higher NRM and presents a population of patients where adjusted supportive care or adjusted GVHD prophylaxis modifications is needed.

We observed a statistically significant decreased risk of relapse in those recipients who were CMV seropositive. Of the 169 CMV+ recipients in our cohort, 75 (44%) experienced CMV reactivation. In our analysis, interestingly, CMV reactivation as a time-dependent covariate in regression analysis for the endpoint of relapse did not show an association. The data are mixed regarding impact of CMV seropositivity or CMV reactivation on post-transplant relapse and NRM. Our group has previously demonstrated a diminished risk of relapse in those undergoing RIC allogeneic transplant who experienced CMV reactivation citing expansion of a specific population of NK cells (CD56 ^dim CD57+NKG2C+) in response to the CMV reactivation. Those data revealed diminished relapse in the setting of CMV reactivation in comparison to CMV negative patients suggesting a protective effect of CMV reactivation on relapse post RIC allogeneic HCT.²⁴ Tiera et al reported a CIBMTR analysis showing no impact of CMV reactivation on relapse rates regardless of type of hematologic malignancy but did highlight increased NRM and diminished overall survival.²⁵ In an EBMT analysis, Schmidt-Hieber et all showed similar outcomes with diminished survival and leukemia free survival, no change in relapse, and increased NRM in CMV seropositive patients.²⁶ In contrast Green et al report diminished relapse in the setting of CMV reactivation for those with AML but not ALL, MDS, or lymphoma with offsetting increased NRM but no difference in survival.²⁷ Thus, our findings of diminished relapse but similar NRM in the setting of CMV seropositivity only are in partial contrast to the other published data. The differences may be explained by the smaller sample size in our study as comparison to the large CIBMTR/EBMT analyses.

Lastly, we observed rates of acute GVHD higher than expected in comparison to contemporary studies. Our overall rates of 43% grade II-IV and 27% grade III-IV were modestly higher with the largest contribution from unrelated PBSC and marrow. Clinical Trials Network (CTN) 0901 reported 31% grade II-IV and only 6.8% grade III-IV utilizing mini-methotrexate + CSA or tacrolimus, tacrolimus + sirolimus, or CSA + MMF. ¹ CALGB 100103/CTN 0502 reported even lower rates of 9.6% grade II-IV aGVHD with a mini-methotrexate + tacrolimus approach.² Our data suggests that alternative GVHD prophylaxis strategies may be preferred, especially for PBSC unrelated donor transplants or mismatched sibling transplant. Post-transplant cyclophosphamide has become an effective GVHD prophylaxis strategy for haploidentical bone marrow transplants yielding very low aGVHD rates ^28,29 and is now extending to related and unrelated donor transplants. ^30–33 Numerous other strategies for GVHD prevention and treatment including histone deacetylase inhibitors, interleukin 6 antibodies, proteasome inhibitors, etc are also in development.³⁴ Given the improvement in rates of severe aGVHD with post-transplant cyclophosphamide in addition to an MMF + tacrolimus backbone, future investigation of this GVHD prophylaxis with our RIC conditioning regimen may be valuable, especially for transplants utilizing and peripheral blood unrelated donor sources or mismatched sibling donors.

In summary, our fludarabine + cyclophosphamide +TBI RIC platform produced highly successful outcomes in indolent lymphoma and CML and those with low risk DRI. More stringent disease burden criteria at transplant, modifications to conditioning intensity approaches, utilization of alternative GVHD prophylaxis, and initiation of targeted post-transplant maintenance therapy approaches will strive to improve outcomes for the remaining hematologic malignancy patients treated with our RIC platform.

Highlights.

• "Our fludarabine + cyclophosphamide +TBI RIC platform produced highly successful outcomes in indolent lymphoma and CML and those with low risk DRI.
• "Flat rate fludarabine dosing decrease did not translate into diminished non-relapse mortality; thus, personalized pharmacokinetic fludarabine dosing may be needed to achieve that goal
• "Our rates of acute GVHD were higher than desired, especially in the URD receiving PBSC. Alternative GVHD prophylaxis strategies proposed for future studies.
• "More stringent disease burden criteria at transplant and initiation of targeted post-transplant maintenance therapy approaches are needed to improve outcomes for the higher risk hematologic malignancy patients treated with our RIC platform.