Prior status of graft‐versus‐host disease affects donor lymphocyte infusion outcomes in patients with relapsed haematological malignancies after allogeneic stem cell transplantation

Summary

Optimizing the effectiveness of donor lymphocyte infusion (DLI) for relapse after allogeneic stem cell transplantation (alloHSCT) has been challenging. We investigated whether the benefits of achieving full donor chimerism (FDC) and developing graft‐versus‐host disease (GVHD) after DLI are affected by a history of GVHD before DLI. We retrospectively analysed 56 patients who received DLI for relapse after alloHSCT at our institute from 2015 to 2022. Among 29 patients without GVHD before DLI, those who achieved FDC after DLI had a significantly higher 1‐year overall survival (OS) compared to those who did not (73.7% vs. 20.0%; p < 0.001). Furthermore, in the same cohort, patients who developed GVHD after DLI showed a tendency towards higher OS compared to those who did not (60.0% vs. 52.6%; p = 0.067). In contrast, in patients with GVHD before DLI, there was no significant difference in OS between patients with or without achieving FDC (64.3% vs. 33.3%, p = 0.226) or developing GVHD (60.0% vs. 53.3%, p = 0.866). Our study showed that achieving FDC or developing GVHD after DLI improved OS, but only in those without a prior history of GVHD. The graft‐versus‐leukaemia effect may be better exploited in patients without a history of GVHD.

INTRODUCTION

Relapse in haematological malignancies after allogeneic stem cell transplantation (alloHSCT) is a significant challenge. It is the leading cause of treatment failure, with 15%–58% of patients with acute myeloid leukaemia (AML) experiencing relapse depending on the disease risk, the intensity of conditioning and the donor type. ¹ , ² When compared to other causes of death, such as infection and graft‐versus‐host disease (GVHD), studies from the Center for International Blood and Marrow Transplant Research cite relapse as the main contributor to post‐transplant mortality among patients with haematological malignancies. ³ Additional salvage therapies are often pursued for those who relapse after alloHSCT, which involves a cellular therapy with either a second alloHSCT or a donor lymphocyte infusion (DLI).

The mechanism of DLI lies in its potential to enhance graft‐versus‐leukaemia (GVL) effects and suppress disease relapse. During its introduction in the 1990s, it successfully achieved durable remissions in patients with relapsed chronic myeloid leukaemia. ⁴ , ⁵ However, it was less effective for patients with relapsed acute leukaemias. ⁴ , ⁶ , ⁷ Overtime, pre‐emptive DLI was also shown to suppress relapse in patients with mixed chimerism (MC) by promoting full donor chimerism (FDC), as well as in patients with minimal residual disease or with high‐risk diseases. A retrospective, single‐institution study reported higher rates of 5‐year overall survival (OS) in patients who received pre‐emptive or prophylactic DLI compared to patients who received DLI for relapse. ⁸ This suggests that the use of chimerism data to trigger DLI could be beneficial to prevent disease relapse, in addition to using DLI therapeutically.

Despite extensive research related to the utility of DLI, it is still unclear how to predict prior to the treatment who would benefit the most from DLI, or whether a DLI or second alloHSCT is the better option. A large retrospective study found no significant differences in OS in patients with AML who relapsed after a first alloHSCT between the two approaches, a DLI or a second alloHSCT. ⁹ Although preliminary findings from a single‐centre study suggested that a second alloHSCT may benefit patients who achieved complete remission from chemotherapy induction after relapse, ¹⁰ evidences are limited.

While GVHD is recognized as a potentially fatal toxicity of DLI, it is a double‐edged sword with the potentially favourable effect of GVL. In a study highlighting the complex relationship between GVHD and GVL, the development of chronic GVHD (cGVHD) after DLI was associated with a higher rate of remission and improved survival. ¹¹ It is unknown if a patient's prior history of GVHD affects the outcomes after DLI. The aim of this study was to evaluate if there is a relationship between a patient's prior GVHD history before DLI and DLI response.

METHODS

Patients and study design

The Penn State Cancer Institute institutional review board approved this retrospective study (STUDY00020839). We selected patients with haematological malignancies who received DLI as a treatment for post‐alloHSCT relapse and had available chimerism data at our institution from 2015 to 2022. Patients who received DLI pre‐emptively (i.e. patients who exhibited MC without relapse) were excluded. During the study period, a total of 65 patients received DLI for haematological relapse. Of these, nine patients had no available chimerism data after DLI and were excluded. A total of 56 patients were studied retrospectively. Loss of heterozygosity or mismatched human leukocyte antigen (HLA) loss was not studied in this historical cohort, although it is recognized as a major mechanism by which malignant cells escape the donor's anti‐leukaemic T cells, leading to DLI refractoriness. ¹²

Chimerism analysis

Chimerism was tested by short tandem repeat (STR) method or quantitative polymerase chain reaction (qPCR) on total unfractionated cells from peripheral blood (PB) or bone marrow (BM) as previously described. ¹³ Briefly, when recipient chimerism is ≥5%, the STR method with GenePrint 24 (Promega, Fitchburg, WI, CAT# B1870) was used. When recipient chimerism is <5%, the qPCR method with KMRTrack (GenDx, Utrecht, the Netherlands, CAT# 8355520) was used. Chimerism was tested monthly in the first 3 months after transplant, followed by every 3 months in the first year and every 6–12 months after the first year, or at the physician's discretion. We defined FDC as >95% and MC as 5%–95% of donor chimerism in total cells from PB or BM samples, following the guidelines provided by the American Society for Transplantation and Cellular Therapy. ¹⁴

Term definitions and DLI procedure

Haematological relapse was defined as either circulating blasts or ≥5% blasts in the BM in patients with acute leukaemia. For lymphoma and multiple myeloma patients, relapse was determined based on imaging and/or increased levels of disease markers. The history of GVHD before the first DLI includes any grades of acute GVHD (aGVHD) or cGVHD. DLI was generally given to patients with 10 × 10⁶/kg, 30 × 10⁶/kg and all the rest (30–100 × 10⁶/kg) of CD3⁺ T cells for the first, second and third doses respectively. For those who received haploidentical donors, the patients received 1 × 10⁶/kg of T cells for DLI without dose escalation.

Statistical methods

The medians and ranges are provided for continuous variables, and percentages are shown for categorical variables. All survival analyses as well as the analysis of the incidences of treatment‐related mortality (TRM) and GVHD were performed using the first DLI date as day 0. The probability of OS and the comparison between the groups with p‐values were calculated by the Kaplan–Meier and the log‐rank test respectively. Landmark plots were used to visualize the effect of achieving FDC and developing GVHD after DLI on OS. Landmark day was set at 60 days after DLI because approximately 75% of patients had their chimerism evaluated within this time frame. In the landmark analysis, patients were grouped based on whether they had achieved FDC or developed GVHD within 60 days after DLI, and any events occurring after 60 days were considered as not having occurred in the analysis. Patients who died within 60 days after DLI were excluded from the analysis. The Cox proportional hazards regression model was used to analyse overall mortality; achieving FDC and developing GVHD after DLI were treated as time‐dependent covariates. The incidences of GVHD and TRM were calculated using a competing risk model and cumulative incidence estimates, and univariate comparisons of incidence curves were made using the Gray test. Relapse after alloHSCT and death without relapse, as well as GVHD and death without GVHD, were considered competing events. All tests were two‐sided, and p‐values < 0.05 were considered statistically significant. All statistical analyses were performed with EZR Version 1.65. ¹⁵

RESULTS

Patient characteristics

The patient characteristics are shown in Table 1. Among the 56 patients in the study, the median age at the time of DLI was 61 years old (range: 2–77 years old), and 55.4% were male. 71.4% had a diagnosis of AML, 5.4% acute lymphoblastic leukaemia (ALL), 8.9% myelodysplastic syndrome (MDS)/myeloproliferative neoplasm and 14.3% lymphoma/multiple myeloma. At the time of alloHSCT, 67.9% of patients had achieved complete remission, and 46.4% relapsed more than 6 months after alloHSCT. Regarding donor types, 75.0% of patients received a transplant from an unrelated donor, 17.9% had an HLA‐matched related donor and 7.1% had a haploidentical donor. Among those patients with an unrelated donor, 76.2% had matched unrelated donors (8/8 match), and 23.8% had mismatched unrelated donors (7/8 match). Before DLI, 16 patients had a history of aGVHD (grade I = 4, II = 8, and III = 4), and nine patients had a history of cGVHD (mild = 7, moderate = 2). There was no significant difference in OS after DLI between patients with and without a history of GVHD before DLI (Figure S1). 60.7% of patients received DLI at least 6 months after alloHSCT. To assess the potential impact of treatment between relapse and DLI on clinical outcomes, we summarized the diverse treatment approaches in Table 2, which includes FDC achievement and outcomes after DLI. Overall, 83.9% of patients received two or fewer DLIs after relapse and 51.8% received a total DLI dose of CD3⁺ T cells between 1 × 10⁷ and 5 × 10⁷ cells/kg.

TABLE 1.

Patient's characteristics.

Number of patients	56
Age at DLI (%)
Median [range], years	61 [2, 77]
<60 years	25 (44.6)
≥60 years	31 (55.4)
Sex (%)
Female	25 (44.6)
Male	31 (55.4)
Disease (%)
AML	40 (71.4)
ALL	3 (5.4)
MDS/MPN	5 (8.9)
Lymphoma/MM	8 (14.3)
Disease status at HCT (%)
CR	38 (67.9)
Non‐CR	18 (32.1)
Donor type (%)
HLA‐matched related	10 (17.9)
Unrelated	42 (75.0)
Haploidentical	4 (7.1)
GVHD prophylaxis (%)
Ciclosporin‐based	8 (14.5)
Tacrolimus‐based	32 (58.2)
PTCy	15 (27.3)
Duration from HCT to relapse (%)
Median [range], days	142 [29, 2549]
≤6 months	30 (53.6)
>6 months	26 (46.4)
Duration from HCT to DLI (%)
Median [range], days	264 [56, 3063]
≤6 months	22 (39.3)
>6 months	34 (60.7)
% Blast at relapse (%)
<20%	24 (42.9)
≥20%	18 (32.1)
Lymphoma/MM or unknown	14 (25.0)
History of aGVHD before DLI (%)
No	40 (71.4)
Max grade I	4 (7.1)
Max grade II	8 (14.3)
Max grade III	4 (7.1)
History of cGVHD before DLI (%)
No	47 (83.9)
Mild	7 (12.5)
Moderate	2 (3.6)

Abbreviations: aGVHD, acute GVHD; ALL, acute lymphoblastic leukaemia; AML, acute myeloid leukaemia; cGVHD, chronic GVHD; CR, complete response; DLI, donor lymphocyte infusion; FDC, full donor chimerism; GVHD, graft‐versus‐host disease; HCT, allogeneic stem cell transplantation; HLA, human leukocyte antigen; MDS/MPN, myelodysplastic syndromes/myeloproliferative neoplasms; MM, multiple myeloma; PTCy, post‐transplant cyclophosphamide.

TABLE 2.

Summary of pre‐DLI treatment and post‐DLI outcomes.

Disease	Treatment before DLI; N (%)	Median time from HCT to relapse, days [range]	Median time from HCT to DLI, days [range]	Alive; N (%)	Achieved FDC post‐DLI		Maintained FDC from pre‐DLI in achieved FDC post‐DLI patients
					N (%)	Alive; N (%)	N (%)	Alive; N (%)
AML N = 40	No treatment N = 8 (20.0%)	102 [81, 209]	171 [97, 536]	N = 4 (50.0%)	N = 5 (62.5%)	N = 4 (50.0%)	N = 4 (50.0%)	N = 3 (37.5%)
	Low‐intensity therapy N = 16 (40.0%)	176 [30, 2549]	238 [56, 2772]	N = 3 (18.8%)	N = 9 (56.3%)	N = 2 (12.5%)	N = 4 (25.0%)	N = 1 (6.3%)
	Standard chemotherapy N = 16 (40.0%)	142 [29, 1674]	277 [56, 3063]	N = 4 (25.0%)	N = 12 (75.0%)	N = 4 (25.0%)	N = 8 (50.0%)	N = 2 (12.5%)
ALL N = 3	Standard chemotherapy N = 3 (100%)	207 [75, 237]	322 [272, 392]	N = 1 (33.3%)	N = 3 (100%)	N = 1 (33.3%)	N = 3 (100%)	N = 1 (33.3%)
MDS/MPN N = 5	Low‐intensity therapy N = 3 (60.0%)	256 [209, 586]	361 [272, 690]	N = 2 (66.7%)	N = 2 (66.7%)	N = 2 (66.7%)	N = 1 (33.3%)	N = 1 (33.3%)
	Standard chemotherapy N = 2 (40.0%)	286 [221, 351]	339 [268, 410]	N = 0 (0%)	N = 2 (100%)	N = 0 (0%)	N = 1 (50%)	N = 0 (0%)
Lymphoma/MM (N = 8)	No treatment N = 1 (12.5%)	501	592	N = 1 (100%)	N = 1 (100%)	N = 1 (100%)	N = 0 (0%)	–
	Low‐intensity therapy N = 2 (25.0%)	150 [118, 182]	219 [182, 257]	N = 0 (0%)	N = 2 (100%)	N = 0 (0%)	N = 1 (50.0%)	N = 0 (0%)
	Standard chemotherapy N = 5 (62.5%)	92 [56, 315]	428 [80, 937]	N = 0 (0%)	N = 3 (37.5%)	N = 0 (0%)	N = 3 (37.5%)	N = 0 (0%)

Note: Low‐intensity therapy included such as hypomethylating agents (azacitidine or decitabine) with or without venetoclax, targeted therapies (e.g. sorafenib, gilteritinib, jakafi), proteasome inhibitors (ixazomib), immunotherapy (elotuzumab‐based regimen) and localized radiation.

Abbreviations: ALL, acute lymphoblastic leukaemia; AML, acute myeloid leukaemia; DLI, donor lymphocyte infusion; FDC, full donor chimerism; HCT, allogeneic stem cell transplantation; MDS/MPN, myelodysplastic syndromes/myeloproliferative neoplasms; MM, multiple myeloma; OS, overall survival.

Effect of achieving FDC and developing GVHD after DLI

Of the 56 patients, 39 (69.6%) achieved FDC within 60 days after DLI, while 17 (30.4%) did not (MC). Of those who achieved FDC post‐DLI, 24 started with FDC before DLI and maintained FDC, while 15 converted from MC to FDC. Among patients who achieved FDC post‐DLI, 33.3% (8/24) of those with maintained FDC and 40.0% (6/15) of those who converted from MC were alive at last follow‐up. In the landmark analysis, six patients in the FDC group and one patient in the MC group were excluded because they died within 60 days after DLI. Patients who achieved FDC after DLI had a higher OS than those who failed to achieve FDC (1‐year OS were 69.6% in FDC vs. 25.0% in MC, p < 0.001) (Figure 1A). Among patients who achieved FDC within 60 days, the 1‐year OS was 60.6% in the maintained FDC group and 90.0% in the converted FDC group. Although not statistically significant (p = 0.138), the converted group showed a favourable trend (Figure S2). Among patients with a history of GVHD before DLI, there was no statistically significant difference between those who achieved FDC and those who did not (1‐year OS were 64.3% in FDC vs. 33.3% in MC, p = 0.226) (Figure 1B). In contrast, those who achieved FDC after DLI had a significantly superior OS than those who did not among patients without a history of GVHD before DLI (1‐year OS 73.7% in FDC vs. 20.0% in MC; p < 0.001) (Figure 1C). These results were consistent even when grade I aGVHD was excluded from the history of GVHD before DLI (Figure S3A,B). In addition, similar trends were observed when the analysis was restricted to patients with AML (Figure S4A–C).

FIGURE 1.

The probability of overall survival in landmark analyses (landmark day = 60 days after the first donor lymphocyte infusion [DLI]), comparing patients who achieved full donor chimerism (FDC) after DLI and those who did not (mixed chimerism: MC). (A) All patients, (B) patients with a history of graft‐versus‐host disease (GVHD) before DLI and (C) patients without a history of GVHD before DLI. The solid line represents patients with FDC after DLI, while the dashed line represents those with MC after DLI.

Eighteen (32.1%) patients developed GVHD within 60 days after DLI, including one patient with grade I aGVHD, nine with grade II aGVHD, three with grade III aGVHD, two with grade IV aGVHD, two with moderate cGVHD and one with grade III aGVHD plus severe cGVHD. In the landmark analysis, three patients with GVHD post‐DLI and four patients without GVHD post‐DLI were excluded because they died within 60 days after DLI. Among the overall cohort, developing GVHD after DLI did not significantly affect the probability of OS (60.0% in GVHD vs. 52.9% in Non‐GVHD; p = 0.181) (Figure 2A). There was no significant difference in OS between developing GVHD or not after DLI in patients with a history of GVHD (60.0% in GVHD vs. 53.3% in Non‐GVHD; p = 0.866) (Figure 2B). In contrast, developing GVHD after DLI showed a tendency towards higher OS compared to those who did not among patients without a history of GVHD (60.0% in GVHD vs. 52.6% in Non‐GVHD; p = 0.067) (Figure 2C). These results were consistent even when grade I aGVHD was excluded from the history of GVHD before DLI (Figure S5A,B). Similar results were also confirmed in a subgroup analysis limited to patients with AML (Figure S6A–C).

FIGURE 2.

The probability of overall survival in landmark analyses (landmark day = 60 days after the first donor lymphocyte infusion [DLI]), comparing patients who developed graft‐versus‐host disease (GVHD) after DLI and those who did not (non‐GVHD). (A) All patients, (B) patients with a history of GVHD before DLI and (C) patients without a history of GVHD before DLI. The solid line represents patients who developed GVHD after DLI, while the dashed line represents those who did not develop GVHD after DLI.

Figure 3 shows the subgroup analysis for the impact of achieving FDC and developing GVHD after DLI on overall mortality, stratified by the history of GVHD before DLI (achieving FDC and developing GVHD were treated as time‐dependent covariates). The effect of achieving FDC after DLI on overall mortality was less pronounced in patients with a history of GVHD before DLI compared to those without a history of GVHD (hazard ratio [HR]: 0.55, 95% confidence interval [CI]: 0.19–1.64 with a history of GVHD vs. HR: 0.31, 95% CI: 0.12–0.76 without a history of GVHD). A similar trend was observed for the effect of developing GVHD after DLI on OS (HR: 1.01, 95% CI: 0.36–2.85 with a history of GVHD vs. HR: 0.60, 95% CI: 0.25–1.48 without a history of GVHD).

FIGURE 3.

Subgroup analysis of the effect of achieving full donor chimerism (FDC) and developing graft‐versus‐host disease (GVHD) after donor lymphocyte infusion (DLI) on overall mortality, stratified by the history of GVHD before DLI. CI, confidence interval.

In a multivariate analysis, achieving FDC and developing GVHD after DLI were treated as time‐dependent covariates. Achieving FDC after DLI was associated with a lower risk of overall mortality (HR: 0.36; 95% CI: 0.18–0.71; p < 0.01), while developing GVHD after DLI was not (HR: 0.62; 95% CI: 0.28–1.36; p = 0.23) (Table 3).

TABLE 3.

Univariate and multivariate analyses of factors predicting overall mortality after DLI.

Factors		Univariate analysis		Multivariate analysis
		Hazard ratio (95% CI)	p value	Hazard ratio (95% CI)	p value
Age at DLI	≥60 (vs. <60) years	0.88 (0.48–1.63)	0.68
Sex	Male (vs. female)	0.79 (0.43–1.46)	0.45
Disease	Others (vs. AML)	1.15 (0.56–2.35)	0.70
Disease status at HCT	Non‐CR (vs. CR)	1.55 (0.82–2.93)	0.18
Donor type	Others (vs. matched related)	1.42 (0.62–3.25)	0.41
GVHD prophylaxis	PTCy (vs. CNI‐based)	1.22 (0.59–2.52)	0.58
Duration from HCT to relapse	>6 (vs. ≤6) months	0.66 (0.35–1.22)	0.18
Duration from HCT to DLI	>6 (vs. ≤6) months	0.55 (0.29–1.03)	0.06	0.53 (0.25–1.12)	0.10
% Blast at relapse	≥20 (vs. <20) %	1.02 (0.51–2.04)	0.96
	Lym/MM or unknown (vs. <20%)	0.68 (0.29–1.57)	0.37
History of GVHD before DLI	Yes (vs. no)	1.12 (0.60–2.10)	0.73
Achieving FDC post‐DLI*	Yes (vs. no)	0.36 (0.18–0.71)	<0.01	0.45 (0.22–0.91)	0.03
Developing GVHD post‐DLI*	Yes (vs. no)	0.76 (0.39–1.47)	0.41	0.62 (0.28–1.36)	0.23

Note: ‘Achieving FDC post‐DLI’ and ‘Developing GVHD post‐DLI’ were treated as time‐dependent covariate (*).

Abbreviations: AML, acute myeloid leukaemia; CI, confidence interval; CNI, calcineurin inhibitor; CR, complete remission; DLI, donor lymphocyte infusion; FDC, full donor chimerism; GVHD, graft‐versus‐host disease; HCT, allogeneic stem cell transplantation; Lym, Lymphoma; MM, multiple myeloma; PTCy, post‐transplant cyclophosphamide.

GVHD occurrence and TRM after DLI

The cumulative incidence of grade II–IV aGVHD after DLI was 14.5% at 30 days after DLI and 32.7% at 100 days after DLI (Figure S7A). The cumulative incidence of grade III–IV aGVHD after DLI was 5.4% at 30 days after DLI and 14.3% at 100 days after DLI (Figure S7B). Finally, the cumulative incidence of all severities of cGVHD was 8.9% at 1 year after DLI and 16.4% at 2 years after DLI (Figure S7C).

Among the 40 patients who had died at the time of data collection, 33 had deaths related to relapsed disease. The remaining seven patients had non‐relapse causes of death, including GVHD, sepsis, cardiac arrest, respiratory failure, liver dysfunction, failure to thrive and gastrointestinal haemorrhage. The cumulative incidence of TRM was 5.4% at 1 year after DLI and 9.3% at 2 years after DLI (Figure S7D).

DISCUSSION

Our retrospective study confirmed the association between achieving FDC and lower overall mortality after DLI. In addition, we found that achieving FDC or developing GVHD after DLI was associated with significantly better OS, but only in those without a prior history of GVHD.

In previous studies, chimerism data have been highlighted as a reliable means to monitor disease in the post‐transplant period to identify patients at risk for relapse. ¹⁴ , ¹⁵ A study of paediatric AML patients who have undergone alloHSCT found a positive association between increasing MC and a higher risk for relapse; these findings suggested early immunological intervention can improve event‐free survival rates in children with increasing MC. ¹⁶ Similar findings were discovered in adult patients with AML in a study from 2007 that concluded patients who developed FDC or decreasing MC within 6 months from alloHSCT had a significantly lower risk of relapse compared to those with MC or increasing MC. ¹⁷ In an abovementioned study that investigated pre‐emptive DLI in MC patient populations, achieving FDC after DLI in both the relapsed and non‐relapsed settings showed improved survival. ⁸ Taken together with our findings that only included patients with relapsed disease, chimerism analysis is not only valuable for the post‐alloHSCT period to monitor the risk of relapse at fixed time points but also as a prognostic marker after treatment for relapse with DLI. ¹⁸ , ¹⁹ Furthermore, although previous studies have proposed that DLI in the pre‐emptive setting is more promising than DLI in relapse, our study shows that DLI in the relapsed setting may still lead to reasonable outcomes. ⁸ , ²⁰

Several studies have shown an association between developing GVHD after initial alloHSCT and lower relapse rates and/or improved survival. ²¹ , ²² , ²³ , ²⁴ A multicentre retrospective study found that aGVHD before relapse adversely impacted 2‐year OS from relapse in patients with MDS. ²⁵ However, little is known about how having a history of GVHD affects outcomes in patients who eventually receive a second cellular therapy for relapse after alloHSCT. Additionally, clinical decisions on whether to select DLI or second alloHSCT for relapse after alloHSCT have been difficult for clinicians, given the lack of standardized therapeutic algorithms. Assuming that GVHD is very closely related to the effects of GVL, as discussed in numerous reports, our findings suggest that patients without a history of GVHD may benefit the most from DLI, and this patient‐specific factor should be taken into account when considering additional therapies for relapse. ²⁶ , ²⁷ A large study conducted by the European Society of Blood and Marrow Transplantation (EBMT) reported that MDS and secondary AML patients without a history of cGVHD after initial alloHSCT had better outcomes after a second alloHSCT. However, among their group of patients who received DLI, the effectiveness of DLI was not dependent on a history of cGVHD. ²⁸ In another EMBT study on patients with ALL, cGVHD before relapse did not significantly impact the distribution of different post‐relapse therapies, which included GVL‐based therapies such as DLI. ²⁹ These reports do not correspond with our findings; however, these EBMT studies did not evaluate the effect of chimerism status or the development of GVHD after DLI. To enhance the effectiveness of DLI, it may be beneficial to consider the pre‐DLI status of GVHD along with the chimerism and GVHD status after DLI.

Despite these uncertainties surrounding GVHD and DLI, other factors have been identified that affect DLI's efficacy. In multiple studies we have cited thus far, remission duration is one of the most relevant and commonly discussed prognostic measures after treatment of relapse, and earlier relapse within 6 months after alloHSCT is associated with poorer OS. ⁸ , ⁹ , ²⁸ In our cohort, although not statistically significant, a longer interval of greater than 6 months between alloHSCT and DLI tends to lower mortality. In light of these findings, there is a need for more innovative therapies for patients who experience a shorter remission period as their prognosis after cellular therapies remains dismal. In addition, earlier studies show that cGVHD appears more strongly linked to GVL than aGVHD. ³⁰ We could not stratify patients based on the type of GVHD due to the small sample size and events of GVHD.

In conclusion, although our study is a retrospective analysis with a small sample size from a single institution, our findings confirmed that relapsed patients who achieved FDC after DLI are associated with lower mortality. However, this trend was not observed when analysing the subset of patients with a prior history of GVHD before DLI separately. Furthermore, those who developed GVHD after DLI demonstrated a better OS compared to those who did not, but only in patients without a prior history of GVHD. Therefore, for relapsed patients who have a history of GVHD, treatment strategies other than DLI, such as a second alloHSCT, could be considered as an alternative option. Further studies are needed to validate with more subjects.

AUTHOR CONTRIBUTIONS

C.M.V., Y.I., S.M. and K.M. participated in the research design, data analysis and writing of this article. J.C., K.R., N.S., M.N., H.Z., S.N., C.E., J.M., B.S., C.V., R.G., V.B. and H.S. participated in the patients and donors clinical management and writing of this article.

FUNDING INFORMATION

This work was supported by the Arlene Witmer Memorial Fund for Bone Marrow Cancer Research, the John and Denise Gilliland Fund for the Cancer Genetics Program, G. R. Sponaugle Employee Cancer Research Fund, Earl ‘Bumps’ Clouser Memorial Lymphoma Research Endowment, Megan and Andrew Enders Stewardship, and Richard E. and Stephanie A. Ziegler Charitable Foundation Endowment in Hematology. In addition, the authors would like to thank the patients who participated in this study and the technicians who participated in the experimental procedures.

CONFLICT OF INTEREST STATEMENT

All authors declare no conflicts of interest associated with this manuscript.

ETHICS STATEMENT

This study was approved by the institutional review board of Penn State Cancer Institute (STUDY00020839).

PATIENT CONSENT

Informed consent was waived by the institutional review board as part of the exemption determination for this retrospective study.

Supporting information

Data S1.

Velasco CM, Inoue Y, Cioccio J, Rakszawski K, Songdej N, Nickolich M, et al. Prior status of graft‐versus‐host disease affects donor lymphocyte infusion outcomes in patients with relapsed haematological malignancies after allogeneic stem cell transplantation. Br J Haematol. 2025;207(2):515–524. 10.1111/bjh.20215

DATA AVAILABILITY STATEMENT

The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.