Background: Monitoring ctDNA after definitive surgery has emerged as an accurate predictor of breast cancer (BC) recurrence. However, further data is needed to determine the proportion and dynamics of ctDNA positivity in different BC subtypes. Here we present ctDNA status and dynamics in patients from an ancillary study (A1) of the prospective randomized phase III JCOG1204 trial (UMIN000012429; Registration: November 2013 - January 2020, N = 1034) in which intensive post-operative follow-up is compared to standard follow-up in BC patients at high risk of recurrence. Methods: Eligible patients enrolled in JCOG1204 were required to have either (1) ≥4 lymph node metastases for estrogen receptor (ER)+ BC without neoadjuvant therapy, (2) ≥1 lymph node metastases for ER- without neoadjuvant therapy, or (3) ≥1 lymph node metastases after neoadjuvant therapy. Additional inclusion criteria for ancillary study JCOG1204A1 include no more than 4 years after surgery and relapse-free at the enrollment. A personalized, tumor-informed assay (Signatera, Natera Inc.) was used for ctDNA surveillance every 3 - 6 months for 2 years for ER- or 5 years for ER+ BC. ctDNA status, time of ctDNA detection and association with subtype were examined. The JCOG DSMC approved the publication of this interim analysis. Results: A total of 248 patients were enrolled (July 2020-October 2023) in JCOG1204A1. Personalized assays were successfully designed for 88% patients, the main reason for failure was insufficient tissue (83%; 25/30). A total of 1440 plasma ctDNA samples were collected from 211 patients. The median age at the time of surgery was 50 years. The most prevalent subtype was ER+ HER2- (Luminal; n = 130), followed by ER- HER2- (TN; n = 38), ER- HER2+ (HER2; n = 24), and ER+ HER2+ (Luminal-HER2; n = 18). 25 patients (11.8%) had at least one positive ctDNA result, with overall sample-level positivity proportion of 4.5% (65 / 1440); 76% (19/25) became detectable with longitudinal testing. Surveillance in TN [the first time point median 31.5 months (12.6-55.8) from surgery] did not result in ctDNA detection. Of the 130 patients with Luminal subtype [the first time point median 36.6 months (12-57) from surgery], 21 (16.2%) had ctDNA first detected at a median 43.5 months (18-72). Additionally, 5 patients (2.4%) had ctDNA detection for the first time after 5 years of surgery. The proportion of ctDNA positivity increased by ~2% annually in Luminal subtype; varied rates of ctDNA level increase have been observed. Conclusions: Monitoring ctDNA in high-risk BC reveals a detection proportion of 11.8% at the patient level (TNBC: 0%; Luminal: 16.2%) Longitudinal testing revealed ctDNA detection in patients with luminal subtype beyond 5 years from surgery, suggesting a correlation with late recurrence. Subsequent reports will focus on investigating the association of ctDNA status with relapse and overall survival.