NICHE-2：短程新辅助免疫后的长期生存与ctDNA

Figure 1: Kaplan–Meier estimates of disease-free survival Time to event was calculated from date of surgery to the date of recurrence or death from any cause. Tick marks indicate data censored at the last moment the patient was known to be disease-free and alive. 3-year disease-free survival was 100% ([97·5% CI 100–100]; one sample log-rank test p<0·0001).

Figure 2: Kaplan–Meier estimates of overall survival Time to event was measured from date of enrolment to date of death from any cause. Tick marks indicate data censored at the last moment the patient was known to be alive. 5-year overall survival was 100% (95% CI 100–100).

Of the 111 patients in the efficacy analysis, ctDNA could be evaluated in 108 patients (data were not available for two patients with pathological complete response and one patient with near-pathological complete response). All patients for whom ctDNA could be evaluated are included at the baseline timepoint. Subsequent timepoints (week 2 and presurgery) include only patients with detectable ctDNA at baseline. For some patients, plasma samples were not available for all timepoints, as indicated in the column ctDNA missing. Subgroups of pathological response include pathological complete response, defined as absence of residual viable tumour in both the primary tumour and lymph nodes; near-pathological complete response, defined as ≤10% residual viable tumour in the resection specimen; partial response, defined as ≤50% residual viable tumour in the resection specimen; and non-response, defined as ≥50% residual viable tumour in the resection specimen. ctDNA=circulating tumour DNA. *Contains all 108 patients for whom ctDNA could be evaluated. †ctDNA status is stratified by pathological response. The total number of patients stratified by pathological response equals 107, as pathological response was unavailable for one patient. Table: ctDNA status across treatment timepoints in all patients and by pathological response category

Figure 3: Alluvial plot of ctDNA dynamics during neoadjuvant treatment Alluvial plot showing individual patient ctDNA dynamics across treatment time points for all patients for whom plasma samples were available at baseline. Each horizontal bar (flow) represents a single patient and depicts ctDNA status at baseline, after cycle 1 (week 2), and after cycle 2 (before surgery). Flows can be followed from left to right to track changes in ctDNA status for each patient over time. Flow colours indicate ctDNA positivity or negativity at the timepoint towards which the patient is transitioning. The vertical bars (strata) at each timepoint show ctDNA positivity or negativity at baseline and at week 2. At the presurgery timepoint, ctDNA status is linked to the patient’s pathological response at surgery. ctDNA=circulating tumour DNA.

Figure 4: ctDNA dynamics during neoadjuvant treatment (A) Longitudinal ctDNA concentration in all patients measured at baseline, after cycle 1 (week 2), and after cycle 2 at the presurgery timepoint. Overall differences across timepoints were assessed using the Friedman test, followed by pairwise Wilcoxon signed-rank tests. (B) Baseline ctDNA concentration stratified by clinical T stage based on radiographic staging at baseline. The cT2, cT3, and cT3–4a group includes patients classified as cT2, cT3, and cT3–4a (n=31). The cT4 group includes patients classified as cT4a and cT4b (n=66). A tumour was deemed cT3–4a when the radiologist could not differentiate between cT3 and cT4 disease. Consequently, these patients were assigned to the lower stage group (cT2, cT3 and cT3–4a). The cT4 group was reserved strictly for tumours that the radiologist unequivocally staged as cT4a or cT4b. Comparisons between groups were performed using the Wilcoxon rank-sum (Mann–Whitney U) test. (C) ctDNA dynamics according to pathological response category: pathological complete response (n=66), near-pathological complete response (n=27), and pathological partial response (n=4). Comparisons between response groups at each timepoint were first assessed using the Kruskal–Wallis test, followed by pairwise Wilcoxon rank-sum tests when appropriate. (D) ctDNA dynamics in cT4 tumours stratified by pathological response: pathological complete response (n=42) and near- pathological complete response (n=22). Comparisons between pathological complete response and near-pathological complete response at each timepoint were made using the Wilcoxon rank-sum test. ctDNA concentrations are shown on a logarithmic (log10) scale. Dots represent individual patients. Boxplots indicate the median and interquartile range, with whiskers representing 1·5× the IQR. p values were adjusted for multiple testing using the Benjamini-Hochberg false discovery rate method for A, C, and D. Only (near) significant p values are shown. Patients with undetectable ctDNA at baseline (n=9) and one patient for whom pathological response was unavailable were excluded from this analysis. Due to the small sample size, comparisons involving the patient who had a pathological non-response (n=1) were not performed. ctDNA=circulating tumour DNA. MTM=mean tumour molecules. *Adjusted p values.