Fig. 1

(A) Cumulative disease-free survival (DFS) stratified by mLN-RVT 0%, mLN-RVT 1–10%, and mLN-RVT 11–100% using a univariate cox proportional hazards model. (B) Kaplan-Meier survival curves of DFS stratified by (A) mLN-MPR(-) and mLN-MPR(+) subgroups (p value = 0.011); (C) stratified by mLN-MPR(+)/PT-MPR(+), mLN-MPR(+)/PT-MPR(-), mLN-MPR(-)/PT-MPR(+), and mLN-MPR(-)/PT-MPR(-) subgroups (p value = 0.030); (D) stratified by ypN0/PT-MPR(+), ypN0/PT-MPR(-), ypN1-N2/PT-MPR(+), and ypN1-N2/PT-MPR(-) subgroups (p value = 0.117) (E) Representative 2 cases of inconsistent pathologic responses of residual viable tumor in primary lesion and paired regional LNs, and across different LN stations (Case 1, Case 2). (F) Histogram showing the percent RVT in resected involved LNs of patients with PT-MPR(-) (left, green bar chart) and PT-MPR(+) (right, blue bar chart). Eight patients had PT-MPR(+) while RVT in mLN > 10%; and twelve patients had PT-MPR(-) but mLN reached mLN-MPR(+). (G) Typical examples of patients with partial regression of mLN exhibiting distinct immune-related phenotypes: immune-inflamed subtype (Case 3, LUSQ, with 0% RVT in PT, and 25% in N2 LN), and immune-evacuation subtype (Case 4, LUSQ, with 95% RVT in PT, and 30% in N2 LN). (H) Imaging mass cytometry image of invasion margin (stroma-tumor border area) from 2 partial regression mLN (Case 5, immune-inflamed subtype; Case 6, immune-evacuation subtype). The two solid line boxes represented the region of interests (ROIs). Blue color: CD3; White color: CD8; Green color: PD-1. (I) The histograms showing the CD3, CD8, and PD-1 densities, respectively, in ROIs between immune-inflamed subtype and immune-evacuation subtypes