Clinical utility of tumour-informed circulating tumour DNA in patients with testicular cancer.

Abstract

To investigate whether circulating tumour DNA (ctDNA) predicts oncological outcomes in the minimal residual disease (MRD) window after radical orchiectomy or retroperitoneal lymph node dissection (RPLND), and to evaluate the diagnostic performance of ctDNA in reference to imaging studies in clinical stage (CS) I disease. Longitudinal tumour-informed ctDNA assays (Signatera™) were collected prospectively from consecutive patients during 2022-2023. The ctDNA signature was informed from the orchiectomy or RPLND specimen. The MRD window was defined as the initial 90 days after orchiectomy or RPLND. Event-free survival (EFS) and recurrence-free survival (RFS) were assessed using the Kaplan-Meier method. Sixty patients with a median (interquartile range) follow-up of 18 (13-23) months underwent 287 ctDNA analyses; 45% had seminoma. At initial staging, 65% had CS (clinical stage) I, 20% CS II, 3.3% CS III, and 10% CS IS. Post-orchiectomy ctDNA clearance was achieved in 80% of patients, and patients with undetectable MRD ctDNA had superior 3-month (100% vs 25% [95% confidence interval [CI] 7.5-83]) and 12-month EFS (91% [95% CI 79-100] vs (13% [95% CI 2-78]) compared to detectable ctDNA, respectively. Undetectable MRD window ctDNA in patients with CS I undergoing surveillance was associated with an 18-month RFS of 88%. The diagnostic performance of ctDNA compared to imaging studies for CS I per patient was: sensitivity 33.3%, specificity 93.5%, positive predictive value 33.3%, and negative predictive value (NPV) 93.5%. Detectable pre-RPLND ctDNA fully correlated with tumour presence, while undetectable ctDNA corresponded to benign histology or teratoma on pathology. Undetectable RPLND MRD ctDNA was associated with superior 9-month RFS vs detectable ctDNA (100% vs 25% [95% CI 5-100], respectively). ctDNA status in the MRD window predicted survival outcomes after orchiectomy and RPLND. Undetectable MRD window ctDNA demonstrated high NPV relative to imaging studies in CS I disease. Further investigation is warranted.