Mika Okazawa-Sakai, Takao Fujisawa, Tatsuyuki Chiyoda, Hiroshi Yagi, Ichinosuke Hyodo, Shogen Boku, Yoh Hayasaki, Masanori Isobe, Daisuke Shintani, Kosei Hasegawa, Tomomi Egawa-Takata, Kimihiko Ito, Kei Ihira, Hidemichi Watari, Kazuhiro Takehara, Kiyoko Kato, Hiromichi Nakajima, Kenichi Harano, Yoshiaki Nakamura, Hideaki Bando, Takayuki Yoshino, Daisuke Aoki
This paired tissue-plasma longitudinal analysis highlights an interpretive challenge in cfDNA testing during first-line platinum-to-PARPi maintenance therapy: putative CH candidates may become detectable as tumor-derived cfDNA declines. Given cfDNA-only inference without matched WBC sequencing and the small serial cohort, these observations are supportive of prior reports of CH dynamics but require validation in larger WBC-integrated cohorts.
OBJECTIVE: To describe cell-free DNA (cfDNA)-inferred putative clonal hematopoiesis (CH) candidates during first-line platinum-based chemotherapy followed by poly(ADP-ribose) polymerase inhibitor (PARPi) maintenance in ovarian cancer.
METHODS: In SCRUM-Japan MONSTAR-SCREEN-1, we analyzed clinically reported paired tumor tissue and plasma cfDNA profiling (324-gene assays). Baseline (B1) data were assessed in 35 treatment-naïve patients; longitudinal cfDNA was available at B1, after platinum without progression (B2), and during/after PARPi without progression (B3) in 8 patients. Putative CH candidates were defined as pathogenic variants (variant allele frequency <40%) in prespecified CH-related genes detected in plasma but not detected in matched tumor tissue at clinical reporting thresholds.
RESULTS: At B1, putative CH candidates were reported in 19/35 patients (54.3%), most commonly DNMT3A; positivity was associated with age ≥60 years. In the longitudinal cohort, tumor-derived TP53 variants were below the assay reporting threshold at B3 in all patients, whereas putative CH TP53 variants were reported in 0/8 patients at B1 and 6/8 patients at B3. DNA damage response gene candidates (TP53/ATM/CHEK2) more frequently became detectable above the reporting threshold during B2-B3 than during B1-B2, while epigenetic-gene candidates showed relatively stable detectability. No therapy-related myeloid neoplasm events were observed during follow-up.
CONCLUSION: This paired tissue-plasma longitudinal analysis highlights an interpretive challenge in cfDNA testing during first-line platinum-to-PARPi maintenance therapy: putative CH candidates may become detectable as tumor-derived cfDNA declines. Given cfDNA-only inference without matched WBC sequencing and the small serial cohort, these observations are supportive of prior reports of CH dynamics but require validation in larger WBC-integrated cohorts.
TRIAL REGISTRATION: UMIN-CTR Clinical Trial Identifier: UMIN000036749.