Prolonged chimeric antigen receptor‐T apheresis to infusion time is associated with inferior outcomes in diffuse large B‐cell lymphoma

Summary

Chimeric antigen receptor (CAR) T‐cell therapy is effective for relapsed/refractory (r/r) diffuse large B‐cell lymphoma (DLBCL). However, delays between apheresis and infusion frequently occur due to limited facility capacity and prolonged bridging therapy, and the clinical impact of such delays remains uncertain. We retrospectively analysed R/R DLBCL patients who underwent CAR‐T‐cell therapy at Kyoto University Hospital. Among 90 patients, the median apheresis to infusion interval was 66 days (range, 28–203). Patients were categorized into delayed (≥66 days) and non‐delayed (<66 days) groups. Baseline characteristics at infusion were similar, whereas adverse features such as extranodal involvement and stable/progressive disease (SD/PD) were less common in the delayed group. Multivariate analysis identified delayed infusion as a significant negative prognostic factor for progression‐free survival (adjusted hazard ratio [aHR] 3.13; 95% confidence interval [CI] 1.63–6.00; p = 0.001), along with extranodal involvement (aHR 2.39; p = 0.004), SD/PD (aHR 2.66; p = 0.005), bulky disease (aHR 3.08; p = 0.008) and treatment with tisa‐cel (aHR 5.26; p = 0.001). Overall survival was also inferior in the delayed group (aHR 2.53, 95% CI 1.29–4.96; p = 0.007). Minimizing apheresis to infusion interval through improved workflows, institutional capacity and avoiding prolonged bridging in non‐responders may enhance outcomes.

INTRODUCTION

Diffuse large B‐cell lymphoma (DLBCL) can often be cured with front‐line immunochemotherapy. ¹ However, outcomes for patients with relapsed or refractory (R/R) disease remain poor, even after intensified chemotherapy and stem cell transplantation. ² , ³ , ⁴ In recent years, CD19‐directed chimeric antigen receptor (CAR) T‐cell therapy has emerged as a promising treatment, significantly improving prognosis in a subset of patients with R/R DLBCL who have exhausted conventional therapeutic options. ⁵ , ⁶ , ⁷ , ⁸

Despite its promise, CAR‐T‐cell therapy is not effective in all patients. Clinical trials and early real‐world data have identified several adverse prognostic factors associated with inferior responses, including poor performance status, elevated lactate dehydrogenase (LDH) levels and high tumour burden. ⁸ , ⁹ , ¹⁰ , ¹¹ Moreover, as CAR‐T‐cell therapy involves multiple complex steps before infusion, waiting periods between apheresis and infusion can be prolonged in real‐world practice for a variety of reasons. ¹² , ¹³ , ¹⁴ , ¹⁵ , ¹⁶ In particular, an increasing number of patients requiring CAR‐T‐cell therapy has exceeded available medical resources essential for post‐infusion patient management, resulting in prolonged waiting periods. Additionally, bridging therapy is often extended to achieve better disease control before infusion, reflecting previous reports indicating that improved pre‐infusion disease status is associated with favourable clinical outcomes. ¹¹ However, extended bridging therapy may in turn contribute to infusion delays. Such delays potentially lead to disease progression and reduce therapeutic efficacy. For many patients with R/R DLBCL, in order to deliver CAR‐T‐cell therapy effectively without compromising their outcomes, it is essential to clarify the clinical impact of infusion delays. However, evidence regarding the prognostic impact of delayed infusion in CAR‐T‐cell therapy is extremely limited. ¹¹ , ¹⁶

Therefore, in this study, we sought to evaluate the relationship between the time from apheresis to CAR T‐cell infusion and key clinical outcomes, such as overall survival and response rates, in patients with R/R DLBCL.

PATIENTS AND METHODS

Study cohort

We analysed all consecutive patients with R/R DLBCL who received axicabtagene ciloleucel (axi‐cel), lisocabtagene maraleucel (liso‐cel) or tisagenlecleucel (tisa‐cel) at Kyoto University Hospital between January 2019 and December 2023. The end of follow‐up was defined as September 2024. Diagnosis of DLBCL was based on the WHO classification of tumours of haematopoietic and lymphoid tissues (revised 4th edition). ¹⁷ This study was approved by the Ethics Committee of Kyoto University and was conducted in accordance with the principles of the Declaration of Helsinki. Written informed consent was obtained from all patients.

End‐points and variables

The schedule of CAR‐T‐cell infusion was determined according to the order of apheresis performed, and availability of predefined institutional resources, particularly inpatient units required for post‐infusion clinical management, and was not based on predefined clinical criteria. To evaluate the effects of prolonged time between apheresis to infusion, we divided our patients into two groups based on the median time between apheresis to infusion: a delayed infusion (≥median) group and a non‐delayed infusion (<median) group. The primary end‐point of this study was progression‐free survival (PFS). Overall survival (OS) was evaluated as a secondary end‐point. PFS was defined as the time from the date of CAR‐T infusion to the date of documented disease progression, relapse or death from any cause. OS was calculated from the date of CAR‐T infusion to death from any cause. Patients without an event were censored at the last date of follow‐up. Disease status at infusion was assessed using the Revised Response Criteria for Malignant Lymphoma. ¹⁸ Progression of relapse was defined based on morphological and clinical evidence of disease activity. All variables shown in tables and figures were retrospectively obtained from patient records. Disease status in the early bridging phase was assessed after two cycles of bridging therapy. Bulky disease was defined as a single nodal mass greater than 10 cm in the maximum dimension.

Statistical analysis

Continuous variables were summarized using medians and ranges, and categorical variables were summarized as counts and percentages. For comparisons between groups, patients and disease characteristics were compared using the Wilcoxon rank sum test for continuous variables and Fisher's exact test for categorical variables. Probabilities of PFS and OS were estimated using the Kaplan–Meier method, and comparisons between groups employed the Wald test in univariate analysis. Associations of the interval from apheresis to infusion with PFS and OS were analysed using restricted cubic spline (RCS) analysis with five knots at the 5th, 27.5th, 50th, 72.5th and 95th percentiles of the interval. ¹⁹ In multivariate analysis, Cox proportional hazard models were used to evaluate the effect of time from apheresis to CAR‐T‐cell infusion in combination with other clinically relevant variables at infusion, including disease status, extranodal involvement, bulky disease and types of infused CAR‐T‐cell product. Statistical significance was set at p < 0.05. All statistical analyses were performed using R (version 4.3.3; R Development Core Team).

RESULTS

Patient characteristics

In total, 90 patients with DLBCL were included in this study (Table 1). The study cohort comprised 43 females and 47 males, with a median age of 65 years (range, 20–76 years). At the time of infusion, disease status was complete response or partial response (CR/PR) in 50% of patients and stable disease or progressive disease (SD/PD) in 50% (Table 1). At the time of infusion, 52 patients (58%) had received more than three lines of therapy, and extranodal involvement was observed in 46 (51%). Bulky disease was present in 10 (11%). Regarding CAR T‐cell products, 65 (72%) received tisa‐cel, 19 (21%) received liso‐cel and 6 (7%) received axi‐cel. The median time from apheresis to infusion was 66 days (range, 28–203) (Figure S1). Patients were classified into two groups based on this interval: a delayed infusion (≥66 days) group and a non‐delayed infusion (<66 days) group.

TABLE 1.

Patient characteristics at infusion.

	Total (n = 90)	Delayed (n = 47)	Non‐delayed (n = 43)	p
Time from apheresis to infusion (days)				<0.001
Median, range	66 (28–203)	83 (66–203)	53 (28–64)
Sex				0.673
Female	43 (47.8)	21 (44.7)	22 (51.2)
Male	47 (52.2)	26 (55.3)	21 (48.8)
Age				0.085
Median, range	65 (20–76)	66 (20–76)	62 (28–75)
Hans				0.915
GCB	42 (46.7)	21 (44.7)	21 (48.8)
Non‐GCB	46 (51.1)	25 (53.2)	21 (48.8)
Missing	2 (2.2)	1 (2.1)	1 (2.3)
Transformed lymphoma				0.226
Transformed	23 (25.6)	15 (31.9)	8 (18.6)
De novo	67 (74.4)	32 (68.1)	35 (81.4)
Prior lines				0.398
≤3	38 (42.2)	22 (46.8)	16 (37.2)
>3	52 (57.8)	25 (53.2)	27 (62.8)
Number of bridging lines				0.113
<3	83 (92.2)	41 (87.2)	42 (97.7)
≥3	7 (7.8)	6 (12.8)	1 (2.3)
Number of cycles of bridging chemotherapy				<0.001
Median, range	2 (0–5)	3 (0–5)	2 (0–5)
Bridging with conventional chemotherapy a	63 (70)	34 (72.3)	29 (67.4)	0.651
Bridging with Pola‐based therapy	22 (24.4)	13 (27.7)	9 (20.9)	0.475
Bridging with radiotherapy	8 (8.9)	6 (12.8)	2 (4.7)	0.270
No bridging therapy	9 (10)	2 (4.3)	7 (16.3)	0.080
Disease status at infusion				0.034
CR/PR	45 (50.0)	29 (61.7)	16 (37.2)
SD/PD	45 (50.0)	18 (38.3)	27 (62.8)
Extranodal involvement at infusion				0.097
Yes	46 (51.1)	20 (42.6)	26 (60.5)
No	44 (48.9)	27 (57.4)	17 (39.5)
Bulky disease (>10 cm) at infusion				0.510
Yes	10 (11.1)	4 (8.5)	6 (14)
No	80 (88.9)	43 (91.5)	37 (86)
LDH b at infusion (IU/L)				0.465
Median, range	224 (125–2892)	223 (125–2892)	226 (167–1385)
Infused CAR‐T‐cell product				0.766
Tisa‐cel	65 (72.2)	34 (72.3)	31 (72.1)
Liso‐cel	19 (21.1)	9 (19.1)	10 (23.3)
Axi‐cel	6 (6.7)	4 (8.5)	2 (4.7)

Abbreviations: Axi‐cel, axicabtagene ciloleucel; CAR, chimeric antigen receptor; CR, complete response; GCB, germinal centre B‐cell‐like; liso‐cel, lisocabtagene maraleucel; PD, progressive disease; Pola, polatuzumab vedotin; PR, partial response; SD, stable disease; tisa‐cel, tisagenlecleucel.

^a Conventional chemotherapy included the following cytotoxic combination regimens (with or without rituximab): CHASE (cyclophosphamide, cytarabine, etoposide and dexamethasone), CHOP (cyclophosphamide, doxorubicin, vincristine and prednisolone), DeVIC (dexamethasone, etoposide, ifosfamide and carboplatin), EPOCH (etoposide, prednisolone, vincristine, cyclophosphamide and doxorubicin), ESHAP (etoposide, cisplatin, cytarabine and methylprednisolone), GCD (gemcitabine, carboplatin and dexamethasone), GDP (gemcitabine, dexamethasone and cisplatin), ICE (ifosfamide, carboplatin and etoposide).

^b Normal range: 124–222 IU/L.

There was no significant difference between the two groups in age (median, 66 vs. 62; p = 0.085). Proportions of patients with germinal centre B‐cell‐like (GCB) DLBCL (44.7% vs. 48.8%; p = 0.915) and transformed lymphoma (31.9% vs. 18.6%; p = 0.226) were comparable in the two groups. Other factors, including bulky disease (8.5% vs. 14%; p = 0.510), LDH levels (median, 223 IU/L vs. 226 IU/L; p = 0.465) and the proportion of patients treated with tisa‐cel (72.3% vs. 72.1%; p = 0.766), also showed no significant differences. Similarly, the proportion of patients with >3 prior therapy lines did not differ significantly (53.2% vs. 62.8%; p = 0.398). Regarding bridging therapy, patients in the delayed group received a significantly higher number of chemotherapy cycles than in the non‐delayed group (median, 3 vs. 2; p < 0.001), reflecting the longer waiting period. However, there was no significant difference in the number of distinct chemotherapy lines used between the two groups.

While baseline characteristics were generally comparable, several potentially adverse prognostic factors were less prevalent in the delayed group. For instance, the proportion of patients with extranodal involvement at infusion was lower in the delayed group (42.6%) than in the non‐delayed group (60.5%), although the difference was not statistically significant (p = 0.097). Furthermore, the proportion of patients with SD/PD at the time of infusion was significantly lower in the delayed group (38.3%) than in the non‐delayed group (62.8%; p = 0.034).

Univariable analysis of prognostic factors for PFS and OS

Table 2 summarizes the results of the univariable analysis of PFS and OS in patients treated with CAR‐T‐cell therapy. Patients with SD/PD at the time of infusion exhibited significantly worse PFS (hazard ratio [HR] 2.51; 95% confidence interval [CI] 1.41–4.45; p = 0.002) and OS (HR 2.66; 95% CI 1.35–5.23; p = 0.005) than those with CR/PR. Similarly, extranodal involvement at infusion was significantly associated with poorer PFS (HR 2.35; 95% CI 1.33–4.15; p = 0.003) and OS (HR 2.32; 95% CI 1.20–4.50; p = 0.012). The presence of bulky disease at infusion was also linked to significantly worse PFS (HR 2.70; 95% CI 1.26–5.77; p = 0.011) and OS (HR 2.62; 95% CI 1.15–5.98; p = 0.022). Furthermore, use of tisa‐cel as the infused CAR‐T‐cell product was associated with significantly inferior PFS (HR 4.47; 95% CI 1.78–11.24; p = 0.001) and OS (HR 3.31; 95% CI 1.18–9.30; p = 0.023) compared to other CAR‐T‐cell products. Of note, although the delayed group did not reach statistical significance for PFS (HR 1.45; p = 0.187) or OS (HR 1.76; p = 0.077), a trend towards worse outcomes was observed (Figure 1). The lack of statistical significance may have been influenced by differences in baseline patient characteristics between the two groups. Therefore, multivariable analysis was performed to adjust for potential confounding factors.

TABLE 2.

Prognostic factors for PFS and OS in univariate analysis.

Variable	PFS			OS
	HR	95% CI	p	HR	95% CI	p
Delayed versus non‐delayed infusion	1.45	0.84–2.51	0.187	1.76	0.94–3.30	0.077
Age ≥65 years versus <65 years	0.78	0.46–1.34	0.375	1.03	0.56–1.91	0.918
Male versus female	1.54	0.89–2.66	0.124	1.30	0.70–2.41	0.406
GCB versus non‐GCB	1.46	0.84–2.53	0.175	1.58	0.84–2.99	0.156
Transformed versus de novo lymphoma	0.99	0.53–1.82	0.965	1.27	0.65–2.47	0.483
Number of prior lines >3 versus ≤3	0.69	0.40–1.19	0.185	0.84	0.45–1.56	0.577
SD/PD versus CR/PR	2.51	1.41–4.45	0.002	2.66	1.35–5.23	0.005
Extranodal involvement (+) versus (−)	2.35	1.33–4.15	0.003	2.32	1.20–4.50	0.012
Bulky disease >10 cm (+) versus (−)	2.70	1.26–5.77	0.011	2.62	1.15–5.98	0.022
LDH > ULN versus ≤ULN	1.38	0.79–2.39	0.254	1.46	0.78–2.75	0.239
Tisa‐cel versus axi‐cel/liso‐cel	4.47	1.78–11.24	0.001	3.31	1.18–9.30	0.023

Abbreviations: Axi‐cel, axicabtagene ciloleucel; CI, confidence interval; CR, complete response; GCB, germinal centre B‐cell‐like; HR, hazard ratio; LDH, lactate dehydrogenase; liso‐cel, lisocabtagene maraleucel; OS, overall survival; PD, progressive disease; PFS, progression‐free survival; PR, partial response; R/R DLBCL, relapsed or refractory diffuse large B‐cell lymphoma; SD, stable disease; tisa‐cel, tisagenlecleucel; ULN, upper limit of normal (222 IU/L).

FIGURE 1.

Kaplan–Meier curves for (A) progression‐free survival (PFS) and (B) overall survival (OS), stratified by time from apheresis to infusion. Differences between groups were evaluated in univariate analyses.

Multivariate analyses of prognostic factors for PFS and OS

In the multivariable analysis for PFS, delayed infusion was identified as a significant poor prognostic factor (adjusted HR [aHR] 3.13; 95% CI 1.63–6.00; p = 0.001). Other significant factors included extranodal lesions (aHR 2.39; 95% CI 1.31–4.35; p = 0.004), SD/PD (aHR 2.66; 95% CI 1.33–5.29; p = 0.005), bulky disease at infusion (aHR 3.08; 95% CI 1.33–7.13; p = 0.008) and treatment with tisa‐cel (aHR 5.26; 95% CI 2.02–13.69; p = 0.001). OS was also inferior in the delayed infusion group (aHR 2.53; 95% CI 1.29–4.96; p = 0.007) (Figure 2). Similar results were observed when the time from apheresis to infusion (weeks) was analysed as a continuous variable (Table S1; Figure S2). These findings are further illustrated in the adjusted Kaplan–Meier survival curves, in which the delayed infusion group consistently demonstrated inferior PFS and OS after controlling for confounding factors (Figure 3).

FIGURE 2.

Multivariate analysis of factors for progression‐free survival (PFS) and overall survival (OS) following chimeric antigen receptor T‐cell therapy. Forest plots showing hazard ratios (HRs) with 95% confidence intervals (CIs) for variables included in the multivariate Cox regression model. (A) HRs for PFS. (B) HRs for OS. Each square represents the HR, and horizontal lines represent the 95% CI. p‐Values were calculated using the Wald test. Axi‐cel, axicabtagene ciloleucel; CR, complete response; liso‐cel, lisocabtagene maraleucel; PD, progressive disease; PR, partial response; SD, stable disease; tisa‐cel, tisagenlecleucel.

FIGURE 3.

Effects of delayed infusion on progression‐free survival (PFS) and overall survival (OS). Kaplan–Meier curves comparing patients with non‐delayed (blue) versus delayed (yellow) infusion for (A) PFS and (B) OS. CI, confidence intervals; HR, hazard ratio.

Subgroup analysis

We performed subgroup analyses to evaluate the impact of delayed therapy for various patient backgrounds (Figure 4). Overall, no subgroup demonstrated a benefit from delayed infusion. Notably, among patients with SD/PD, the detrimental effect of delay on PFS (HR 4.28; 95% CI 1.97–9.30; p < 0.001) and OS (HR 2.98, 95% CI 1.39–6.39; p = 0.005) was particularly pronounced (Figures 4 and 5). Kaplan–Meier analyses stratified by time from apheresis to infusion (delayed vs. non‐delayed) and disease status at infusion (CR/PR vs. SD/PD) confirmed this impact, demonstrating markedly inferior PFS and OS for patients in the delayed + SD/PD group compared to the other three subgroups (Figure 5). In addition, delayed infusion was associated with poorer PFS among patients with extranodal involvement (HR 2.50; 95% CI 1.25–4.99; p = 0.009) and poorer OS among those with elevated LDH (HR 3.07; 95% CI 1.27–7.44; p = 0.013) (Figure 4). In particular, among patients who initially presented with SD/PD early during bridging therapy, the vast majority remained with SD/PD at the time of infusion in both the delayed and non‐delayed groups (95% in the non‐delayed group and 88% in the delayed group), suggesting that extended bridging therapy in combination with delayed infusion may not contribute to a reduction in the proportion of patients with SD/PD (Figure S3). These results suggest the importance of non‐delayed infusion, particularly in patients with adverse prognostic factors such as SD/PD, extranodal involvement and elevated LDH.

FIGURE 4.

Forest plots of subgroup analyses for (A) progression‐free survival (PFS) and (B) overall survival (OS). Differences between groups were evaluated in univariable analyses. Axi‐cel, axicabtagene ciloleucel; CI, confidence intervals; CR, complete response; GCB, germinal centre B‐cell‐like; HR, hazard ratio; LDH, lactate dehydrogenase; liso‐cel, lisocabtagene maraleucel; PD, progressive disease; PR, partial response; SD, stable disease; tisa‐cel, tisagenlecleucel.

FIGURE 5.

Kaplan–Meier curves for (A) progression‐free survival (PFS) and (B) overall survival (OS), stratified by time from apheresis to infusion (delayed vs. non‐delayed) and disease status at infusion (complete or partial response [CR/PR] vs. stable or progressive disease [SD/PD]). Differences between groups were evaluated in univariable analyses. CI, confidence intervals; HR, hazard ratio.

DISCUSSION

In this study, we demonstrated that prolonged time from apheresis to CAR‐T‐cell infusion is an independent prognostic factor for worse PFS and OS in patients with R/R DLBCL after accounting for other significant prognostic factors using multivariate analysis. Adverse prognostic effects of disease status, extranodal involvement and bulky disease in CAR‐T therapy outcomes were reconfirmed in our cohort. Our findings provide novel insights that may inform future treatment strategies.

In our cohort, prolonged time from apheresis to infusion was clearly associated with inferior PFS and OS. While a large Center for International Blood and Marrow Transplant Research (CIBMTR) study of 1497 patients also reported that prolonged vein‐to‐vein interval adversely affected outcomes, ¹⁶ another real‐world analysis of 116 patients found no such association when comparing relapsed and non‐relapsed cases after CAR‐T therapy in univariable analysis. ¹¹ This discrepancy may stem from differences in study design, particularly the lack of multivariate adjustment for confounding factors such as extranodal involvement and SD/PD at the time of infusion in the previous report. Notably, our study included a wider range of infusion intervals, which were largely determined by the order of apheresis and inpatient unit availability rather than predefined clinical criteria, thereby introducing a degree of randomness that may have allowed clearer identification of prognostic impact. The median interval from apheresis to infusion in our cohort was 66 days (range, 28–203), notably longer than the 4–7 weeks reported in previous clinical trials and early real‐world data. ¹⁶ , ²⁰ , ²¹ , ²² This delay likely reflects real‐world barriers such as hospital capacity limitations and complex inter‐facility referral processes in Japan. Our findings highlight the importance of minimizing delays to prevent clinical deterioration and to ensure optimal treatment outcomes. Since the waiting period is highly influenced by hospital location, healthcare systems and evolving regulatory issues pertaining to CAR‐T‐cell products, continuous and real‐time monitoring of waiting time is crucial for addressing delays in clinical settings.

In this analysis, the identification of other major adverse prognostic factors in CAR‐T‐cell therapy, that is, disease status (SD/PD), extranodal involvement and bulky disease, is largely consistent with previous reports. Prior studies have suggested that these factors significantly impact treatment outcomes, ¹¹ and our univariate and multivariate analyses confirmed that they have a significant adverse effect on both PFS and OS. Additionally, patients treated with tisa‐cel showed poorer outcomes than those treated with liso‐cel or axi‐cel. While this trend is consistent with earlier reports, ²³ , ²⁴ it may be at least partly attributed to the fact that tisa‐cel was the first CAR‐T product approved in Japan. At our institution, tisa‐cel was more frequently used than other CAR‐T products administered during the initial phase of CAR‐T therapy implementation, when a higher proportion of patients presented with advanced disease. Therefore, differences in baseline characteristics, including disease severity, may have influenced outcomes observed in this group. In this study, no significant differences were observed in waiting periods or the number of bridging therapy lines between different CAR‐T products.

Subgroup analysis in this study indicated that the negative impact of delayed infusion was more pronounced among patients with SD/PD at infusion. Poor disease status at the time of infusion is associated with worse outcomes following CAR‐T‐cell therapy. ¹¹ Although intensifying bridging therapy could potentially improve disease status before infusion, it may also result in delayed infusion. Importantly, our findings indicate that continued bridging therapy rarely improved disease status for patients presenting with SD/PD early during bridging therapy (Figure S3). Consistent with this finding, the need for more intensive salvage was a key characteristic of the delayed SD/PD group; a significantly higher proportion of these patients received three or more lines of bridging therapy compared to their non‐delayed counterparts (27.8% [5 of 18] vs. 3.7% [1 of 27]; p = 0.031), strongly implying clinical deterioration during the prolonged wait, which is the likely driver of the poor outcomes observed in this subgroup. These findings support using the early response to bridging therapy as a key decision criterion. Specifically for patients with SD/PD who do not respond to the initial one to two cycles of bridging therapy, our results indicate that proceeding promptly with CAR‐T infusion is a more rational strategy than continuing potentially ineffective treatment. Importantly, this timeframe is clinically feasible, as it corresponds to the typical turnaround period from apheresis product shipment to CAR‐T‐cell product delivery in Japan. Further investigation is warranted, however, to prospectively confirm whether this early‐infusion strategy truly improves outcomes in this high‐risk subgroup. While the clinical impact of a delay was less pronounced in patients in remission, extended waiting periods may still lead to clinical deterioration. These include not only disease progression, which we observed in a subset of patients initially in CR at apheresis, but also complications such as infections and a decline in performance status that may compromise eligibility for infusion. Therefore, minimizing the wait time remains the ideal goal for all patients, even those in remission.

This study has several limitations. First, it is a retrospective, single‐centre analysis with a relatively small sample size, which may affect the generalizability of our findings. While our centre performs more CAR‐T‐cell therapies than most institutions in Japan and maintains fairly consistent eligibility criteria, the lack of multicentre data may introduce selection bias and limit the robustness of our conclusions. Second, our study reflects a period before the widespread use of novel salvage agents, such as bispecific antibodies. The increasing availability of these highly effective therapies may change the clinical decision‐making framework we have described and represents an important area for future study. Third, while our institutional protocol did not prioritize earlier apheresis based on disease aggressiveness, the potential influence of the pre‐referral process at outside institutions was not analysed in this study. Fourth, although an association between delayed CAR‐T‐cell infusion and worse PFS and OS was observed when analysed as a continuous variable, the optimal cut‐off for waiting time may vary depending on clinical and logistical factors. To validate our findings and provide a more comprehensive understanding of the prognostic impact of delayed CAR‐T‐cell infusion in patients with R/R DLBCL, further multicentre prospective studies with larger cohorts are warranted.

In conclusion, our study identified delayed CAR‐T‐cell infusion as an independent adverse prognostic factor for both PFS and OS in patients with R/R DLBCL. These findings underscore the critical importance of timely infusion and call for improved strategies to minimize delays in CAR‐T‐cell therapy.

AUTHOR CONTRIBUTIONS

SM, TJ and YA designed the study, reviewed and analysed data. SM, TJ and YA wrote the paper. TK, TS, CM, JK, M Nishikori, KY, M Nagao and AT‐K interpreted data and revised the manuscript. All authors critiqued the manuscript.

CONFLICT OF INTEREST STATEMENT

The authors declare no competing financial interests.

ETHICS STATEMENT

This study was approved by the Institutional Review Board and Ethic Committee of Kyoto University (G0697).

CONSENT TO PARTICIPATE

Written informed consent was obtained from all participating patients.

Supporting information

Data S1.

Morimoto S, Jo T, Kitawaki T, Sakamoto T, Mizumoto C, Kanda J, et al. Prolonged chimeric antigen receptor‐T apheresis to infusion time is associated with inferior outcomes in diffuse large B‐cell lymphoma. Br J Haematol. 2025;207(4):1484–1494. 10.1111/bjh.70090

DATA AVAILABILITY STATEMENT

Data employed in this study are not publicly available due to ethical restrictions. Releasing the data would exceed the scope of patient consent for research use.