AUTHORS
Zhe Zhou, Yangwei Li2, Desheng Gong3, Shanshan Fan4, Yinyin Chang, MSc5, Lina Du6, Xiaodong Lyu7
AFFILIATIONS
1 The Affiliated Cancer Hospital of Zhengzhou University & Henan Cancer Hospital, Zhengzhou, China
2 The Affiliated Cancer Hospital of Zhengzhou University & Henan Cancer Hospital, Zhengzhou, China
3 Clinical Laboratories, Shenyou Bio, Zhengzhou, China
4 Clinical Laboratories, Shenyou Bio, Zhengzhou, China
5 Clinical Laboratories, Shenyou Bio, Zhengzhou, China
6 Clinical Laboratories, Shenyou Bio, Zhengzhou, China
7 The Affiliated Cancer Hospital of Zhengzhou University & Henan Cancer Hospital, Zhengzhou, China
ABSTRACT
Background: Chimerism analysis is a critical tool for assessing engraftment and predicting relapse after allogeneic hematopoietic stem cell transplantation (allo-HSCT). Conventional STR-based methods, with a sensitivity of only 1%–5%, are insufficient to capture early signals before the clinical onset of relapse. This study utilized a high-sensitivity chimerism monitoring platform to systematically evaluate the prognostic value of refined chimerism-based stratification and to develop a prognostic prediction model integrating chimerism levels, dynamic evolutionary patterns, and clinical characteristics.
Methods: This retrospective study enrolled 549 patients with hematologic disorders who underwent allo-HSCT, including 278 with acute myeloid leukemia, 153 with acute lymphoblastic leukemia, 77 with myelodysplastic syndrome, and 41 with other diagnoses. Chimerism levels were retrospectively collected at three time points post transplantation: day 30 (T1, ±15 days), day 60 (T2, ±18 days), and day 90 (T3, ±15 days). Chimerism was assessed using MSNPseq, an NGS assay that amplifies 100 fragments, each harboring 4–8 SNPs. This assay achieved >2000× sequencing depth for 96.8% of the targeted loci.
Results: A chimerism threshold of 99% was applied to stratify patients into a mixed chimerism group (defined as <99%) and a complete chimerism group (defined as ≥99%). At T1, there were no significant between group differences in PFS or OS (P-value>0.05, log-rank test). In contrast, at both T2 and T3, the mixed chimerism group exhibited significantly inferior PFS and OS relative to the complete chimerism group (P<0.001). Within 3 months following transplantation, a sustained decline in chimerism was associated with significantly inferior PFS (P=0.001) and OS (P<0.001). Patients who exhibited a decrease in chimerism rate of more than 1% at either T2 or T3 had significantly worse PFS (P=0.018) and OS (P=0.003). Based on chimerism dynamics within 3 months post transplant, patients meeting any of the following criteria were classified into the poor prognosis group: (1) chimerism level <99% at T2 or T3 (n=194); (2) continuous decline across all three time points (n=66); or (3) a decrease of >1% between any two consecutive time points (n=127). A prognostic scoring model integrating chimerism risk grouping with clinical features (age, sex, transplant type, aGVHD, CMV status, and EBV status) was constructed using LASSO regression. This model effectively stratified patients into high risk and low risk groups with significantly different PFS and OS outcomes. The model achieved a relapse detection rate of 52.3%, representing a marked improvement over the 11% rate obtained with the conventional 95% chimerism threshold strategy.
Conclusion: This study establishes a dual dimensional chimerism based risk stratification framework that integrates a static 99% threshold with dynamic kinetic patterns—defined as sustained decline or a >1% decrease between consecutive time points—which were both significantly associated with inferior PFS and OS. Integrating this chimerism framework with clinical features via LASSO regression yields a model that substantially outperforms the conventional 95% only threshold in risk stratification and relapse prediction.
DISCLOSURES
Yinyin Chang, MSc: No relevant conflicts of interest to declare.