Impact of Bone Marrow Involvement on the Success of Hematopoietic Stem Cell Mobilization in Patients With Lymphoma

ABSTRACT

The influence of bone marrow (BM) involvement at diagnosis on hematopoietic stem cell (HSC) mobilization outcomes remains to be fully elucidated, particularly in the contemporary plerixafor era. To address this issue, this multicenter retrospective cohort study including four lymphoma treatment centers evaluated patients who underwent their first autologous HSC mobilization procedure between January 2017 and June 2023. Mobilization success was defined as achieving a minimum CD34+ cell yield ≥ 2.0 × 10⁶ cells/kg. Propensity score matching (PSM) was performed to rigorously adjust for baseline confounders and ensure robust group comparisons. The study cohort comprised 669 patients, of whom 138 (20.6%) presented with BM involvement at the time of diagnosis. Analyzes revealed that mobilization success rates were highly comparable between patients with and without BM involvement, both before (79.0% versus [vs.] 81.2%; p = 0.563) and after (79.1% vs. 83.5%; p = 0.500) PSM. Furthermore, the median CD34+ cell yield did not differ significantly between the 2 cohorts either before (3.2 vs. 3.7 × 10⁶/kg; p = 0.192) or after (3.2 vs. 3.5 × 10⁶/kg; p = 0.639) PSM. These findings indicate that initial BM involvement does not significantly impair the success of HSC mobilization in patients with lymphoma, a conclusion that remained highly consistent even after controlling for baseline variables.

1. Introduction

High‐dose chemotherapy followed by autologous stem cell transplantation (ASCT) is an important therapeutic strategy for both Hodgkin's lymphoma and non‐Hodgkin's lymphoma (HL and NHL, respectively), and has been associated with significant improvements in progression‐free and overall survival (PFS and OS, respectively) [1, 2, 3, 4]. The success of ASCT critically relies on the efficient mobilization and adequate collection of hematopoietic stem cells (HSCs), with a minimum target ≥ 2.0 × 10⁶ CD34+ cells/kg required for hematopoietic engraftment [5, 6, 7, 8]. However, previous studies have reported HSC mobilization failure rates ranging from 5% to 30% in patients with lymphoma, with rates as high as 40% in individuals with high‐risk features for poor mobilization [6]. As such, accurately identifying clinical‐ and treatment‐related factors that influence mobilization outcomes is of paramount importance for optimizing stem cell mobilization strategies and maximizing the success rate of ASCT.

The bone marrow (BM) is the primary niche for HSC production and is the direct site of action for many mobilizing agents [9, 10, 11]. Whether lymphoma infiltration of the BM affects mobilization efficiency is a clinically relevant question because it influences physicians' decisions about proceeding to HSC collection and ASCT, and the choice of mobilization strategies and agents. To date, however, a definitive consensus regarding the influence of BM involvement on mobilization outcomes has not been established. Early reports identified BM involvement at diagnosis as a predictor of poor mobilization [12], whereas a more recent study did not consistently find such an association [13]. Although expert guidelines list BM involvement as a risk factor for poor mobilization [5], the distinction between infiltration at diagnosis and mobilization remains poorly defined. In the era of widespread plerixafor use [14] and more refined mobilization approaches, it remains unclear whether baseline BM involvement exerts a clinically meaningful effect on contemporary mobilization outcomes. To address this knowledge gap, we conducted a multicenter study with a large cohort and adjusted for key confounders to evaluate the impact of baseline BM involvement on mobilization success.

Furthermore, beyond BM involvement, various other clinical and treatment‐related factors can also affect mobilization efficacy, including patient age, choice of mobilization strategy, number of previous treatment lines, specific drug exposures, and history of radiotherapy [7, 15]. Notably, the introduction of plerixafor, a CXCR4 inhibitor [16], has revolutionized HSC mobilization and numerous studies have consistently demonstrated that it dramatically increases the success rate of stem cell collection [17, 18]. To further enable risk stratification, some researchers have attempted to integrate clinical indicators and treatment characteristics into predictive models aimed at estimating the likelihood of mobilization failure or forecasting CD34+ cell yield [19, 20, 21].

Accordingly, we conducted a large‐scale, multicenter retrospective cohort study to clarify the impact of BM involvement at diagnosis on HSC mobilization success. In addition, other potential factors associated with mobilization success were also explored.

2. Materials and Methods

2.1. Study Design and Patients

This retrospective cohort study was conducted across 4 lymphoma treatment centers in China and included patients who underwent their first autologous HSC mobilization between January 2017 and June 2023. Clinical baseline characteristics, treatment history, and laboratory data of all patients were extracted from electronic medical records housed at the participating institutions. This study was reviewed and approved by the institutional review boards of all participating centers.

2.2. HSC Mobilization and Collection

Patients received granulocyte colony‐stimulating factor (G‐CSF) at a dose of 10 μg/kg via a single daily subcutaneous injection every morning for up to 8 consecutive days. The decision to use either steady‐state mobilization (G‐CSF alone) or chemo‐mobilization was made at the discretion of the treating physician. For patients undergoing chemo‐mobilization, the mobilization regimens were predominantly based on disease‐specific chemotherapy. The chemo‐mobilization regimens administered in the cohort included (R)‐CHOP (rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone), gemcitabine combined with a platinum agent (cisplatin or oxaliplatin), a regimen comprising ifosfamide, cisplatin, etoposide, dexamethasone, and cytarabine‐based therapies, and other regimens. If the patients had completed all planned chemotherapy cycles, steady‐state mobilization with G‐CSF alone was used. Plerixafor was administered at a dose of 0.24 mg/kg each evening before apheresis for a maximum of 4 consecutive days. The decision to initiate plerixafor, as well as the specific administration strategy (upfront or pre‐emptive), was at the discretion of the treating physician. Peripheral blood CD34+ cell counts were monitored every morning from day 4 of mobilization to the final day of apheresis. Peripheral blood stem cell apheresis was initiated on the morning of day 5 and performed once daily, continuing until a cumulative collection of ≥ 2.0 × 10⁶ CD34+ cells/kg was achieved or for a maximum of 4 days. After each apheresis session, the absolute CD34+ cell yield within the harvested products was quantified.

All of the collected stem cell products were cryopreserved. Immediately after apheresis, the stem cells were suspended in a cryoprotectant solution containing 5% dimethyl sulfoxide (DMSO), human serum albumin (HSA), and 6% hydroxyethyl starch (HES). A non‐controlled‐rate freezing protocol was used, with the grafts stored in an ultra‐low temperature freezer at −80°C. On the day of infusion, the stem cell bags were rapidly thawed in a 37°C water bath, and graft viability was verified using a standard trypan blue exclusion assay.

2.3. Data Collection

The patients' clinical and laboratory data were collected, including sex, age, BM involvement status at diagnosis, lymphoma classification, diagnosis stage, remission status before mobilization, mobilization strategies, lines of previous chemotherapy, previous lenalidomide‐containing therapy, previous platinum‐containing therapy, previous radiotherapy, platelet count, and the number of CD34+ cells collected. The baseline diagnosis of BM involvement was established through a comprehensive assessment of BM aspiration and biopsy specimens, including cytomorphological examination, flow cytometry, and immunohistochemistry (IHC) to confirm the presence or absence of lymphoma infiltration. Before mobilization, all patients underwent repeated BM aspiration and biopsy. The absence of residual lymphoma cell infiltration was confirmed using the same combination of diagnostic modalities (cytomorphology, flow cytometry, and IHC).

2.4. Definitions

The key variables in this study are described as follows. The mobilization approaches were categorized as follows: steady‐state mobilization, which refers to the use of G‐CSF alone to mobilize HSCs; and chemo‐mobilization, which involves the combination of cytokines and chemotherapy [6, 7]. Remission status before mobilization was assessed using the Lugano response criteria for lymphoma [22]. A CD34+ cell yield of ≥ 2.0 × 10⁶ cells/kg was considered to be the minimum threshold for mobilization success [23, 24].

2.5. Statistical Analysis

Continuous variables are expressed as median with interquartile range (IQR), and categorical variables are expressed as frequency and percentage (n [%]). For comparisons between independent groups, Pearson's chi‐squared (χ2) test was used for categorical variables, while the Mann–Whitney U test was applied to continuous variables. For paired samples, continuous variables were compared using the Wilcoxon signed‐rank test and categorical variables were analyzed using the McNemar test. Potential factors associated with mobilization success were screened using univariate logistic regression analysis. Variables with p < 0.10 in the univariate analysis were subsequently included in a multivariate logistic regression model to identify independent predictors of mobilization success.

To control for confounding factors that could affect the comparison between groups with and without BM involvement, propensity score matching (PSM) was performed using 1:1 nearest‐neighbor matching without replacement, with a caliper width equal to 0.2 of the standard deviation of the logit of the propensity score. Standardized mean differences (SMDs) were used to assess the covariate balance before and after matching. An absolute SMD of < 0.1 was considered to be negligible imbalance, and values between 0.1 and 0.2 were considered to be acceptable balance. Covariates for the PSM model were selected based on baseline characteristics that exhibited significant differences (p < 0.05) between the “BM involvement” and “Without BM involvement” groups, and factors known to biologically or clinically influence stem cell mobilization kinetics according to the existing literature and clinical expertise. All statistical tests were two‐sided, and differences with p < 0.05 were considered to be statistically significant. Statistical analyzes were performed using SPSS (IBM Corp., Armonk, NY, USA) and R software (R Core Team; R Foundation for Statistical Computing, Vienna, Austria).

3. Results

3.1. Baseline Characteristics

Data from 669 patients diagnosed with lymphoma were included in this study. The clinical characteristics of the cohort are summarized in Table 1. The median age at mobilization was 45 years (IQR 32.0–55.0), with a male/female ratio of 1.57:1. BM involvement was observed in 138 (20.6%) patients at diagnosis. The distribution of lymphoma subtypes was as follows: diffuse large B‐cell lymphoma was the most prevalent subtype (n = 337 [50.4%]), followed by peripheral T/NK‐cell lymphoma (n = 128 [19.1%]), HL (n = 103 [15.4%]), and lymphoblastic lymphoma (n = 38 [5.7%]). Other identified subtypes included mantle cell lymphoma (n = 39 [5.8%]), follicular lymphoma (n = 5 [0.7%]), and other B‐cell lymphomas (n = 19 [2.8%]). Chemo‐mobilization was performed in 282 (42.2%) patients, whereas the remaining 387 (57.8%) underwent steady‐state mobilization with G‐CSF alone. Within the chemo‐mobilization subgroup (n = 282), (R)‐CHOP was the most frequently administered regimen and was administered to 144 (51.1%) patients. This was followed by a combination of ifosfamide, cisplatin, etoposide, and dexamethasone in 92 (32.6%) patients. Twenty (7.1%) patients received gemcitabine plus platinum, 16 (5.7%) were mobilized using cytarabine‐based regimens, and the remaining 10 (3.5%) underwent other chemotherapy regimens. A subset of 52 (7.8%) patients had a history of radiotherapy. The radiotherapy sites included nodal regions (n = 16), mediastinum (n = 14), pelvis (n = 5), spine (n = 1), central nervous system (n = 5), and other extranodal sites (n = 11).

TABLE 1.

Baseline characteristics of patients stratified by BM involvement.

Characteristics	Overall (N = 669)	BM involvement (n = 138)	Without BM involvement (n = 531)	p
Sex, n (%)				0.363
Male	409 (61.1%)	89 (64.5%)	320 (60.3%)
Female	260 (38.9%)	49 (35.5%)	211 (39.7%)
Age (years), median (IQR)	45.0 (32.0, 55.0)	48.5 (36.3, 56.0)	43.0 (31.0, 55.0)	0.020
Lymphoma classification, n (%)				0.089
NHL	565 (84.5%)	123 (89.1%)	442 (83.2%)
HL	104 (15.5%)	15 (10.9%)	89 (16.8%)
Stage, n (%)				< 0.001
I	50 (7.5%)	0 (0.0%)	50 (9.4%)
II	123 (18.4%)	0 (0.0%)	123 (23.2%)
III	91 (13.6%)	0 (0.0%)	91 (17.1%)
IV	405 (60.5%)	138 (100.0%)	267 (50.3%)
Pre‐mobilization remission status, n (%)				0.328
CR	368 (55.0%)	81 (58.7%)	287 (54.0%)
PR	301 (45.0%)	57 (41.3%)	244 (46.0%)
Mobilization strategies, n (%)				0.723
G‐CSF alone	387 (57.8%)	78 (56.5%)	309 (58.2%)
G‐CSF + chemotherapy	282 (42.2%)	60 (43.5%)	222 (41.8%)
Plerixafor use, n (%)				0.339
Yes	68 (10.2%)	11 (8.0%)	57 (10.7%)
No	601 (89.8%)	127 (92.0%)	474 (89.3%)
Number of previous chemotherapy lines, n (%)				0.120
1 line	368 (55.0%)	84 (60.9%)	284 (53.5%)
≥ 2 lines	301 (45.0%)	54 (39.1%)	247 (46.5%)
Platinum‐containing therapy, n (%)				0.320
Yes	272 (40.7%)	51 (37.0%)	221 (41.6%)
No	397 (59.3%)	87 (63.0%)	310 (58.4%)
Lenalidomide‐containing therapy, n (%)				0.984
Yes	22 (3.3%)	4 (2.9%)	18 (3.4%)
No	647 (96.7%)	134 (97.1%)	513 (96.6%)
Radiotherapy, n (%)				0.016
Yes	52 (7.8%)	4 (2.9%)	48 (9.0%)
No	617 (92.2%)	134 (97.1%)	483 (91.0%)
Pre‐mobilization PLT (×109/L), median (IQR)	169.0 (126.0, 224.0)	167.0 (121.5, 228.5)	171.0 (127.0, 224.0)	0.424
Pre‐apheresis PLT (×109/L), median (IQR)	147.0 (102.0, 189.0)	149.5 (102.8, 192.5)	145.0 (102.5, 188.5)	0.663
MNC (×109/L), median (IQR)	3.3 (2.2, 4.8)	3.6 (2.4, 5.1)	3.2 (2.1, 4.7)	0.070
PB CD34+ cell count (/μL), median (IQR)	12.2 (4.0, 28.8)	12.3 (5.0, 36.9)	12.1 (3.6, 27.3)	0.417
Number of apheresis days (days), median (IQR)	3 (2, 3)	3 (2, 3)	3 (2, 3)	0.491

Abbreviations: BM, bone marrow; CR, complete remission; G‐CSF, granulocyte colony‐stimulating factor; HL, Hodgkin lymphoma; IQR, interquartile range; MNC, mononuclear cell count; NHL, non‐Hodgkin lymphoma; PB CD34+ cell count, peripheral blood CD34+ cell count; PLT, platelet count; PR, partial remission.

3.2. Impact of BM Involvement on Mobilization Outcomes

Overall, 540 (80.7%) patients achieved mobilization success (≥ 2.0 × 10⁶ CD34+ cells/kg), with a median CD34+ cell yield of 3.5 × 10⁶/kg (IQR 2.2–6.2 × 10⁶/kg). The mobilization outcomes did not differ significantly between patients with and without BM involvement (Figures 1 and 2). PSM was subsequently used to rigorously control for baseline confounding factors and to ensure the robustness and statistical reliability of the findings. SMDs were used to assess the balance of baseline characteristics between the 2 groups before and after PSM; detailed balance diagnostic results are provided in Table S1.

FIGURE 1.

Comparison of mobilization outcomes between patients with and without bone marrow (BM) involvement. (A, B) Stacked bar charts illustrating the mobilization success rates (defined as a CD34+ cell yield of ≥ 2.0 × 10⁶/kg) before (A) and after (B) propensity score matching (PSM). (C, D) Stacked bar charts displaying the proportion of patients achieving a CD34+ cell yield of ≥ 5.0 × 10⁶/kg before (C) and after (D) PSM. p values indicate the statistical significance of differences between the two groups. BM, bone marrow.

FIGURE 2.

Comparison of CD34+ cell yield between patients with and without bone marrow (BM) involvement. (A, B) Raincloud plots illustrating the distribution of CD34+ cell yield before (A) and after (B) propensity score matching (PSM). p values indicate the statistical significance of differences between the groups. BM, bone marrow.

Specifically, the success rates were comparable between patients with and without BM involvement, both before (79.0% versus [vs.] 81.2%; p = 0.563) and after (79.1% vs. 83.5%; p = 0.500) PSM (Figure 1A,B, Table S2). Furthermore, the proportions of patients achieving an optimal collection (≥ 5.0 × 10⁶ CD34+ cells/kg) were also similar between the two groups before and after PSM (Figure 1C,D, Table S2). Similarly, the median CD34+ cell yield exhibited no significant differences between the 2 groups before (3.2 vs. 3.7 × 10⁶/kg; p = 0.192) and after PSM (3.2 vs. 3.5 × 10⁶/kg; p = 0.639) (Figure 2A,B, Table S2).

In the present study, plerixafor was administered to 68 patients. Those who received plerixafor achieved a significantly higher mobilization success rate than those who did not (95.6% vs. 79.0%; p = 0.001). Subgroup analyzes revealed that plerixafor consistently improved mobilization success rates, which increased from 77.9% to 90.9% in patients with BM involvement, and from 79.3% to 96.4% in those without. The effect of plerixafor integration was further evaluated across different mobilization strategies. In the steady‐state mobilization group (n = 387), the median CD34+ cell yield was 3.12 × 10⁶/kg and the success rate was 76.7%, while in the chemo‐mobilization group, the yield was 4.31 × 10⁶/kg with a success rate of 86.0%. Stratified analysis (Figure S1A) revealed that the lowest success rate occurred in patients undergoing steady‐state mobilization without plerixafor (n = 329; success rate, 73.5%), which was lower than that in patients undergoing chemo‐mobilization without plerixafor group (n = 272; success rate, 85.6%). Notably, the integration of plerixafor bolstered the success rate of steady‐state mobilization to 94.8% (n = 58), whereas the chemo‐mobilization plus plerixafor group (n = 10) reached 100%. Similar trends were observed in CD34+ cell yield (Figure S1B).

The distribution of CD34+ cell yield across different histological categories is reported in Figure S2A. The distribution of the CD34+ cell yield across different irradiation sites is shown in Figure S2B. There was a trend toward lower CD34+ cell yield in patients who received radiotherapy to the axial skeleton and major hematopoietically active regions, specifically the pelvis, spine, and mediastinum.

3.3. Multivariate Analyzes of Factors Influencing Mobilization Success

Following univariate screening, candidate variables were further evaluated to identify independent factors of mobilization success. Multivariate analysis revealed that mobilization strategies using G‐CSF combined with chemotherapy or the addition of plerixafor, absence of previous radiotherapy, and higher pre‐apheresis platelet counts were significantly associated with an increased likelihood of successful mobilization. Previous exposure to platinum‐containing therapy was identified as a significant independent risk factor that negatively affected mobilization outcomes (Table 2, Figure 3).

TABLE 2.

Univariate and multivariate logistic regression analyzes of factors associated with mobilization success.

Characteristics	Univariate analysis		Multivariate analysis
	OR (95% CI)	p	OR (95% CI)	p
Sex
Male	Reference
Female	0.891 (0.603–1.318)	0.565
Age, years	0.995 (0.982–1.009)	0.517
BM involvement
Yes	Reference
No	1.147 (0.721–1.823)	0.563
Stage
I	Reference
II	0.463 (0.179–1.198)	0.112
III	0.594 (0.218–1.618)	0.308
IV	0.572 (0.235–1.389)	0.217
Lymphoma classification
NHL	Reference
HL	0.761 (0.460–1.258)	0.287
Mobilization strategies
G‐CSF alone	Reference		Reference
G‐CSF + chemotherapy	1.888 (1.250–2.851)	0.003	3.844 (2.364–6.252)	< 0.001
Plerixafor use
No	Reference		Reference
Yes	5.747 (1.777–18.588)	0.004	8.253 (2.497–27.283)	< 0.001
Number of previous chemotherapy lines
1 line	Reference		Reference
≥ 2 lines	0.561 (0.380–0.826)	0.003	1.233 (0.617–2.462)	0.553
Platinum‐containing therapy
No	Reference		Reference
Yes	0.484 (0.328–0.714)	< 0.001	0.439 (0.222–0.866)	0.018
Lenalidomide‐containing therapy
No	Reference
Yes	0.806 (0.292–2.227)	0.678
Radiotherapy
Yes	Reference		Reference
No	2.913 (1.605–5.285)	< 0.001	2.883 (1.465–5.673)	0.002
Pre‐mobilization PLT, ×109/L	1.002 (1.000–1.005)	0.071	1.001 (0.998–1.004)	0.585
Pre‐apheresis PLT, ×109/L	1.005 (1.002–1.008)	0.003	1.007 (1.003–1.011)	< 0.001
MNC, ×109/L	1.041 (0.954–1.136)	0.363

Note: Variables with p < 0.10 in univariate analysis were included in the multivariate model. Mobilization success was defined as achieving a CD34+ cell yield ≥ 2.0 × 10⁶ cells/kg.

Abbreviations: BM, bone marrow; CI, confidence interval; CR, complete remission; G‐CSF, granulocyte colony‐stimulating factor; HL, Hodgkin lymphoma; MNC, mononuclear cell count; NHL, non‐Hodgkin lymphoma; OR, odds ratio; PLT, platelet count; PR, partial remission.

FIGURE 3.

Forest plot of multivariate analysis for factors associated with mobilization outcomes. The forest plot illustrates the independent predictors for mobilization outcomes. Group 1 represents the probability of achieving the minimum threshold for mobilization success (≥ 2.0 × 10⁶ CD34+ cells/kg) and group 2 represents the probability of achieving the optimal target (≥ 5.0 × 10⁶ CD34+ cells/kg). Solid dots indicate the estimated odds ratios (ORs), and horizontal lines represent the corresponding 95% confidence intervals (CIs). The vertical dashed line denotes an OR of 1.0 (reference). For continuous variables, the ORs for Age and Pre‐apheresis PLT are presented per 1‐unit increase (i.e., per year of age and per 1 × 10⁹/L increment in platelet count, respectively). The use of plerixafor was identified as a significant independent predictor in the multivariate model but was omitted from this plot to avoid scale distortion and maintain visual clarity for the other variables. The actual values for plerixafor use are as follows: For group 1, OR = 8.253 (95% CI: 2.497–27.283, p < 0.001); for group 2, OR = 0.934 (95% CI: 0.494–1.765, p = 0.833).

3.4. Post‐ASCT Outcomes

Among the 669 patients who underwent stem cell mobilization, 596 (89.1%) subsequently underwent ASCT. The median time to neutrophil engraftment was 11 days (IQR 10–12 days), and the median time to platelet engraftment was 12 days (IQR 10–15 days). When stratified according to baseline BM involvement, there were no significant differences in engraftment kinetics between patients with and without BM involvement, with a median time to neutrophil engraftment of 11 vs. 10 days (p = 0.417), and platelet engraftment of 13 vs. 12 days (p = 0.090).

The estimated 24‐month PFS rate in the ASCT cohort was 84.9% (95% confidence interval [CI], 81.4%–88.6%), and the 24‐month OS rate was 96.3% (95% CI, 94.4%–98.3%) (Figure S3A,B). Baseline BM involvement was not associated with significantly inferior post‐transplant survival: the 24‐month PFS rates were 81.6% and 85.6% in patients with and without BM involvement, respectively (p = 0.077; Figure S3C), and the corresponding 24‐month OS rates were 93.6% and 97.2%, respectively (p = 0.055; Figure S3D).

4. Discussion

Effective HSC mobilization is a critical prerequisite for successful ASCT. Although previous studies have explored the predictive factors for mobilization success [15, 21, 25], the specific impact of BM involvement at diagnosis remains controversial, particularly in the current plerixafor era. In this multicenter retrospective cohort of 669 patients with lymphoma, including 138 (20.6%) with BM involvement at diagnosis, there were no statistically significant differences in mobilization outcomes between patients with and without BM involvement. Crucially, this comparability was robustly confirmed after PSM, in which the rates of mobilization success—collecting ≥ 2.0 × 10⁶ CD34+ cells/kg—remained similar between the two groups (79.1% vs. 83.5%; p = 0.500). Our large‐scale, real‐world data provide an evidence base for optimizing mobilization management by demonstrating that baseline BM involvement does not inherently preclude successful mobilization.

The clinical significance of whether BM involvement affects HSC mobilization merits attention because it informs the potential need for intensified strategies, as well as the risk of prolonged hospital stays and increased healthcare costs. Although early studies were conducted before the widespread adoption of plerixafor [12, 26], contemporary research has been constrained by limited sample sizes [13]. By leveraging a large, multicenter lymphoma cohort, our study addresses these gaps and provides an evidence base to guide clinical decision making and optimize patient care. Furthermore, it is essential to distinguish BM involvement at the initial diagnosis from persistent infiltration pre‐mobilization. In our cohort, all patients underwent repeat BM aspiration and biopsy before mobilization, and the absence of residual lymphoma infiltration was confirmed by cytomorphology, flow cytometry, and IHC.

With premobilization status controlled for, our study demonstrated that baseline BM involvement does not preclude mobilization success. This aligns with evidence from a study by Olivieri et al. [19], indicating that poor mobilization is primarily associated with heavy disease infiltration (≥ 30%) at the time of mobilization rather than disease status at diagnosis. Although our results suggested that patients with baseline BM involvement yielded numerically lower median CD34+ counts (3.2 vs. 3.7 × 10⁶/kg), this difference was statistically insignificant and did not translate into a significant decrease in mobilization success rates. BM involvement at the time of diagnosis is a reversible insult to the hematopoietic niche. Unlike the irreversible injury to the hematopoietic microenvironment and HSC pool caused by advanced age [27], radiotherapy, or extensive chemotherapy [28], the mobilizing capacity of the BM can be substantially restored once tumor clearance is achieved via induction therapy. Consequently, baseline BM involvement should not be regarded to be a barrier to ASCT, provided that marrow remission is confirmed before mobilization.

Regarding the effect of mobilization strategies on achieving the minimum collection threshold, univariate and multivariate analyzes identified chemo‐mobilization (G‐CSF plus chemotherapy) as an independent protective factor. Mobilization strategies are generally categorized into chemo‐mobilization and steady‐state mobilization (G‐CSF alone) [6, 7]. Substantial evidence has indicated that chemo‐mobilization may reduce tumor burden [6] and, more importantly, increase the yield of HSCs [29, 30, 31, 32]. The choice of chemotherapy regimen was individualized based on the pathological profile and institutional therapeutic protocol. For instance, (R)‐CHOP‐based chemo‐mobilization was utilized for a high‐risk subset of DLBCL patients (age‐adjusted International Prognostic Index score = 3) undergoing ASCT consolidation in the first remission in accordance with the contemporary Chinese Society of Clinical Oncology guidelines [33]. Gokarn et al. [29] reported that GDP‐based and cytarabine‐containing regimens were associated with higher mobilization success rates in patients with HL. Nevertheless, the limitations of chemo‐mobilization should be acknowledged, including higher treatment‐related toxicity, the need for inpatient hospitalization, and higher overall costs [7]. Therefore, the selection of a mobilization strategy requires a meticulous balance between the clinical benefits and potential risks. Furthermore, regarding the impact of previous chemotherapeutic exposure on mobilization efficacy, the cumulative marrow‐suppressive effects of platinum‐based agents, such as cisplatin, carboplatin, and oxaliplatin, may impair hematopoietic function and, consequently, compromise mobilization outcomes [5]. Additionally, the pre‐apheresis platelet count was identified as a significant predictor of mobilization success, which is consistent with previous findings suggesting that peripheral platelet levels serve as reliable indicators of residual reserve capacity [13, 23, 34]. Moreover, the baseline platelet count has also been shown to be a reliable predictor of mobilization response to plerixafor [35].

Plerixafor, a novel CXCR4 antagonist, improves mobilization outcomes by promoting the release of hematopoietic progenitor cells into peripheral circulation through inhibition of the CXCR4/SDF‐1 axis [36, 37]. Currently, clinical strategies for plerixafor use are primarily categorized into 3 approaches: pre‐emptive mobilization; delayed remobilization; and first‐line use [14]. Studies have confirmed that the addition of plerixafor to conventional mobilization strategies significantly enhances CD34+ cell collection efficiency [38]. A phase III randomized clinical trial conducted in China demonstrated that, among patients with NHL, the proportion of those who achieved the target collection threshold (≥ 5.0 × 10⁶ CD34+ cells/kg) was significantly higher in the plerixafor group compared with the placebo group (62% vs. 20%; p < 0.0001) [9]. These findings are consistent with those reported in pivotal studies conducted in the European Union and the United States, which reported similar results (59.3% vs. 19.6%) [39]. Furthermore, the percentage of patients achieving ≥ 2 × 10⁶ CD34+ cells/kg was also significantly greater in the plerixafor group than that in the control group (88% vs. 66%) [9]. Our real‐world analysis further delineated the specific magnitude of the benefits of plerixafor in diverse clinical scenarios. We demonstrated that patients who received plerixafor achieved a significantly higher mobilization success rate than those who did not (95.6% vs. 79.0%; p = 0.001). Notably, this substantial benefit was observed regardless of baseline BM status; the addition of plerixafor improved success rates from 77.9% to 90.9% in patients with BM involvement and from 79.3% to 96.4% in those without BM involvement. Furthermore, our data demonstrated that while steady‐state mobilization without plerixafor yielded the lowest success rate (73.5%), the integration of this CXCR4 inhibitor dramatically bolstered its success rate to 94.8%.

After successful mobilization and collection, the optimal preservation of stem cells is pivotal for ASCT. The standard cryopreservation protocol relies on controlled‐rate freezing and storage in liquid nitrogen (−196°C) or the vapor phase (−156°C) with 10% DMSO [40]. However, given the potential infusion‐related toxicity associated with DMSO, there is increasing interest in reducing its concentration without compromising stem cell viability and engraftment potential [41, 42]. In alignment with this clinical need, our center used a modified cryopreservation strategy involving a non‐controlled‐rate freezing protocol at −80°C, with a reduced DMSO concentration of 5% combined with HSA and 6% HES. Similar alternative cryopreservation conditions have been reported in other studies. Notably, Gokarn et al. [43] highlighted that a protocol using 4.35% DMSO combined with methylcellulose and uncontrolled rate freezing at −80°C could sustain stem cell viability for up to 8 years. Similarly, Pinar et al. [44] observed that, although viability may gradually decrease over time, HSCs stored at −80°C remain capable of supporting durable engraftment. Collectively, these efforts provide a feasible option for resource‐limited healthcare settings.

Our findings provide a different perspective on the common assumption that BM involvement is a major barrier to mobilization. By clarifying the limited impact of BM involvement, our results support the development of more inclusive and individualized mobilization strategies that may ultimately facilitate broader patient access to autologous transplantation. The limitations of this study were as follows. First, there is a small proportion of patients receiving plerixafor (10.2%). Future studies with expanded cohorts and the collection of direct drug expenditure data are required to enable pharmacoeconomic evaluations and to inform optimal administration strategies. Another limitation arises from the study's retrospective design. Consequently, detailed quantitative pathological parameters, such as exact BM cellularity percentages and specific infiltration patterns, could not be completely standardized or retrieved across different institutions. Further efforts are required to develop predictive models for mobilization success and validate them in large‐scale cohorts, with the aim of guiding the selection of mobilization strategies and optimizing the use of plerixafor.

5. Conclusion

Results of this study demonstrated that BM involvement at diagnosis does not significantly compromise the success of HSC mobilization in patients with lymphoma, even after rigorously controlling for baseline confounding factors.

Author Contributions

Jie Lv: conceptualization, data curation, formal analysis, investigation, writing – original draft, visualization. Zailin Yang: conceptualization, data curation, formal analysis, investigation, writing – original draft. Weiping Liu: conceptualization, funding acquisition, methodology, project administration, resources, supervision, writing – review and editing. Yuqin Song: conceptualization, funding acquisition, methodology, project administration, resources, writing – review and editing. Chen Zhang: conceptualization, investigation, project administration, supervision, writing – review and editing. Jun Zhu: supervision, writing – review and editing. Yao Liu: data curation, investigation, resources. Ling Wang: conceptualization, data curation, investigation, resources, writing – review and editing. Xiaoqian Liu: conceptualization, data curation, formal analysis, investigation, writing – original draft.

Funding

This work was supported by the National Natural Science Foundation of China (Grant 82270195), Capital's Funds for Health Improvement and Research (Grants 2022‐1‐2152 and 2024‐1‐2151), Beijing Natural Science Foundation (Grants L244063 and L244025), Noncommunicable Chronic Diseases‐National Science and Technology Major Project (Grant 2023ZD0500800), Beijing Research Ward Excellence Program (Grants BRWEP2024W032150206, BRWEP2024W032150205, BRWEP2024W032150207, BRWEP2024W032150208 and BRWEP2024W032150213) and Beijing Xisike Clinical Oncology Research Foundation (Grant Y‐Young 2023‐0286).

Ethics Statement

This study was conducted in accordance with the Declaration of Helsinki and approved by the institutional review boards of all participating centers, including Peking University Cancer Hospital & Institute (Approval No. 2022YJZ78), Chongqing University Cancer Hospital (Approval No. CZLS2023085‐A‐100), Yantai Yuhuangding Hospital (Approval No. 322669), and Qingdao Central Hospital (Approval No. JS202313401). A waiver of written informed consent was granted by the respective ethics committees, as the study involved a retrospective analysis of de‐identified data, and all data were handled in accordance with institutional confidentiality and data protection policies.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Figure S1: Impact of plerixafor on mobilization outcomes stratified by mobilization strategies. (A) Proportion of patients achieving mobilization success (defined as a CD34+ cell yield of ≥ 2.0 × 10⁶/kg) across different mobilization strategies. (B) Comparison of total CD34+ cell yield among the four subgroups. Statistical significance: *p < 0.05; ***p < 0.001; ns, not significant.

Figure S2: Distribution of CD34+ cell yield across different histological subtypes and radiotherapy sites. (A) CD34+ cell yield stratified by histological subtypes. (B) CD34+ cell yield stratified by radiotherapy sites.

Figure S3: Post‐ASCT survival outcomes. (A) Progression‐free survival (PFS) and (B) overall survival (OS) for the entire ASCT cohort. Comparison of (C) PFS and (D) OS stratified by baseline bone marrow (BM) involvement.

Table S1: Standardized mean differences of covariates before and after propensity score matching.

Table S2: Mobilization outcomes before and after PSM stratified by BM involvement.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.