Background: There are limited data on predictive biomarkers for de-escalated ET or chemotherapy with dual anti-HER2 blockade in HR+/HER2+ early breast cancer (BC). In this translational pre-planned project of the WSG-TP-II phase II-trial (NCT03272477), we aimed to identify associations of gene expression signatures and sTILs with pCR. Methods: Patients with cT1c-cT4c, cN0-3 centrally confirmed HR+/HER2+ BC were randomized to 12 weeks of standard ET (n = 100) or paclitaxel (Pac; n = 107). All patients received trastuzumab + pertuzumab (T+P) q3w as neoadjuvant and adjuvant treatment. Gene expression signatures were analyzed using NanoString BC360 panel in baseline biopsies (T+P+ET: n = 72; T+P+Pac: n = 78). sTILs were analyzed in 93 (T+P+ET) and 97 patients (T+P+Pac) at baseline and in 65 (T+P+ET) and 57 patients (T+P+Pac) at week 3. Impacts of standardized BC360 gene expression signatures and sTILs on pCR (ypT0/is ypN0; primary endpoint) expressed in odds ratios (OR) were estimated by logistic regression analysis. Results: pCR rate in patients with BC360 analysis was 39.3% (T+P+ET: 23.6%; T+P+Pac: 53.9%). Overall, ERBB2 (OR 2.37; 95%CI 1.55, 3.62) and cytotoxic cells signature (OR 1.42; 95%CI 1.02, 1.99) were favorable for pCR, while apoptosis (OR 0.64; 95%CI 0.44, 0.92), estrogen receptor 1 (OR 0.60; 95%CI 0.42, 0.85), estrogen receptor signaling (OR 0.64; 95%CI 0.45, 0.91), and progesterone receptor (OR 0.67; 95%CI 0.47, 0.94) signatures were unfavorable. Analyzing by treatment arm, a similar pattern was observed in the T+P+Pac arm (ERBB2: OR 2.01; apoptosis: OR 0.59; estrogen- and progesterone-related signatures: OR 0.41-0.58), but not in the T+P+ET arm where only ERBB2 was prognostic for pCR (OR 7.24; 95%CI 2.12, 24.05). Baseline ≥30% sTIL levels (vs < 30% sTILs; n = 16 vs n = 174) were associated with pCR (OR 5.16; 95%CI 1.60, 16.66); significance was not achieved in either trial arm. Compared to < 30% sTILs (n = 90), low tumor cellularity at week 3 precluding sTILs analysis ( < 500 invasive tumor cells, n = 22), but not ≥30% sTILs (n = 10), was prognostic for pCR in all patients (OR 9.47; 95%CI 2.94, 30.50) and in individual trial arms (OR 6.00-11.79). Conclusions: This hypothesis-generating translational results suggest that gene expression signatures (particularly ERBB2), baseline sTILs, and low tumor cellularity at week 3 predict pCR after ET + double HER2 blockade. Future neoadjuvant trials are needed to prospectively test the use of baseline gene expression and sTILs analysis, and early on-treatment cellularity measurement to select patients with HR+/HER2+ tumors for de-escalated endocrine-therapy-based approaches. Clinical trial information: NCT03272477.