Koji Kawamura, Koji Kato, Takeshi Sugio, Shinichi Kobayashi, Fumihiko Kimura, Noriko Doki, Naoyuki Uchida, Masatsugu Tanaka, Ayumu Ito, Yasuo Mori, Tetsuya Eto, Shuichi Ota, Yuta Hasegawa, Hirohisa Nakamae, Yuta Katayama, Masashi Sawa, Shinichi Kako, Takeshi Maeda, Takahiro Fukuda, Ken Tabuchi, Yoshiko Atsuta, Junya Kanda, Hideki Nakasone, Yoshinobu Kanda
These findings emphasize the clinical relevance of donor selection after relapse following allo-HCT from an MRD and highlight the need for prospective studies to refine donor selection strategies in this setting.
BACKGROUND: Relapse after allogeneic hematopoietic stem cell transplantation (allo-HCT) from human leukocyte antigen (HLA)-matched related donors (MRDs) remains a major clinical challenge. While second allo-HCT may be curative, the optimal donor selection strategy for such patients remains unclear.
OBJECTIVES: This study aimed to evaluate the impact of donor change in patients who relapse following allo-HCT from an MRD, and to clarify its effect on outcomes after second allo-HCT.
STUDY DESIGN: Using data from the Japanese national transplant registry, we retrospectively analyzed 922 individuals who relapsed after first allo-HCT from an MRD and underwent a second allo-HCT between 2001 and 2021. Donor sources comprised the same MRD (SM-RD), a different MRD, HLA-mismatched related donors (1MM-RD, 23MM-RD), matched or mismatched unrelated donors (M-UD, 1MM-UD), and unrelated cord blood (U-CB).
RESULTS: In multivariate analysis, transplantation from 23MM-RDs was significantly associated with inferior overall survival (OS) compared with SM-RDs (HR, 1.49; P = 0.0062), while other donor sources were not. Additional adverse prognostic factors included non-remission at second transplant, poor performance status, and early relapse (<6 months after first transplant). Same donor selection did not correlate with increased non-relapse mortality, and donor change did not confer a survival advantage, suggesting that SM-RD remains a valid and reasonable option. In an additional sensitivity analysis, the association between 23MM-RD and inferior OS was attenuated and was not statistically significant in the contemporary era (2015-2021), although the donor-by-era interaction was not significant.
CONCLUSIONS: These findings emphasize the clinical relevance of donor selection after relapse following allo-HCT from an MRD and highlight the need for prospective studies to refine donor selection strategies in this setting.