Abstract
Background Bone marrow (BM) aspiration is essential for monitoring post-transplant relapse in AML/MDS and is typically performed unilaterally under the assumption of even blast distribution. However, our clinical observations revealed cases of asymmetric blast distribution, leading to inconsistent cytologic results and potential diagnostic errors. To address this, we conducted a prospective study using bilateral BM aspiration of posterior superior iliac spine (PSIS) to investigate the spatial distribution pattern of blasts.
Method A total of 57 patients undergoing HSCT were enrolled, with 98 bilateral BM aspirations performed. Inclusion criteria were as follows: 1. Patients with AML or MDS who had post-transplant relapse or suspected relapse; 2. Provision of written informed consent; 3. Following bilateral BM aspiration, cases were included if the proportion of blasts exceeded 5% in any side of PSIS, or proportion was between 0–5% with MRD+ on any side. Exclusion criteria were as follows: 1. Patients with diseases other than AML or MDS; 2. Patients were unwilling or unable to tolerate bilateral aspirations; 3. Patients without evidence of relapse after bilateral aspiration or with factors (e.g., BM necrosis) that hindered blast quantification were excluded. The study is registered at ChiCTR.org.cn (ChiCTR2100045918).
Result A total of 22 patients undergoing 36 bilateral BM aspirations were included in the analysis. Asymmetric distribution (Asy-distribution) was defined as a blast percentage <5% on one side and ≥5% on the other, or a ≥1.5-fold difference between sides (n=17); others were classified as symmetric (Sy-distribution, n=19). Patients exhibiting ≥1 Asy-distribution were assigned to the Asy-group (n=11), others to the Sy-group (n=11). Among Asy-distribution, 6 cases showed <5% blasts on one side with ≥5% on the contralateral side (6/17; 6/11 patients). In each cytology pair, the side with higher blast proportion was defined as “high-side,” the other as “low-side.” In Asy-distribution, lymphocyte proportion was significantly lower in the high-side (P=0.0081, Wilcoxon matched-pairs test), while no such difference was observed in Sy-distribution. Single-cell RNA-seq was performed on 3 Asy- and 2 Sy-distribution pairs. In Asy-samples, compared with low-side, leukemia cells from the high-side showed upregulation of histone/chromatin genes (e.g., H3-3A, H3-3B, H1-10) and mitochondrial genes (MT-ND2, MT-ND4, MT-ND5), while inflammatory genes (S100A8, S100A9, NLRP3) were downregulated. GSEA revealed activation of E2F targets, G2/M checkpoint, and MYC targets, with suppression of apoptosis, IL-2–STAT5, and IL-6–JAK–STAT3 signaling pathways. Similarly, compared with Sy-distribution, blast cells from Asy-distribution samples displayed similar DEG and pathway patterns to those from the high-side. CD8⁺ T cells in both high-side and Asy-distribution exhibited reduced expression of cytotoxic genes (GZMB, PRF1, NKG7, CXCR4), as did NK cells (CXCR4). GSEA indicated suppression of immune effector pathways, including IFN-γ signaling, suggesting an immune-evasive phenotype in regions with higher leukemia burden. No significant differences were found between the Asy- and Sy-groups in demographics, overall survival, leukemia-free survival, or aGVHD incidence. Notably, the Asy-group showed a significantly higher 1-year cumulative incidence of cGVHD (45% vs 0%, P=0.013), but the correlation between asymmetric distribution and occurrence of cGVHD before relapse did not reach statistical significance (P=0.059).
Conclusion This study revealed asymmetric distribution of blasts in the bone marrow, accompanied by reduced lymphocyte proportions and immune transcriptional differences. In both high-side vs. low-side and Asy- vs. Sy-distribution comparisons, blasts from the high-side or Asy-distribution consistently exhibited enriched proliferative signatures, suppressed apoptotic and inflammatory pathways, and impaired immune cytotoxicity. Despite a limited cohort, nearly half of the patients exhibited asymmetry. While uneven HSC distribution post-transplant has been suggested, its mechanisms remain unclear. Whether asymmetry is driven by leukemia, immune-mediated, or influenced by cGVHD remains unknown. These findings underscore a clinically overlooked phenomenon and suggest that unilateral marrow sampling may be inadequate for relapse assessment in post-transplant AML/MDS. Larger studies are warranted to elucidate underlying mechanisms.
