Background: Dual trastuzumab plus pertuzumab (HP) has shown promising pathologic complete response (pCR) rates in HER2[+] EBC although lower to CT regimens. Identification of new markers of sensitivity to HP could help to de-escalate CT. PHERGain assessed early metabolic response by F-PET to neoadjuvant HP and the opportunity of CT de-escalation with a response-adapted strategy in patients (pts) with HER2[+] EBC. Methods: PHERGain randomized (1:4 ratio) centrally-confirmed HER2[+] stage I-III EBC pts to receive either docetaxel (T), carboplatin (C), and HP (cohort A) or HP ± endocrine therapy (ET) (cohort B). Randomization was stratified by hormone receptor (HR) status. Pts with subclinical metastases by F-PET were included in a different cohort (cohort C). Centrally-reviewed F-PET was performed prior to randomization and after 2 cycles of TX (cohorts A/B). Cohort A pts completed a total of 6 cycles regardless of F-PET results. Cohort B/PET-responder (RX) pts continued with HP ± ET for 6 cycles, while PET-non-RX pts were switched to receive 6 cycles of TCHP. After surgery, cohort B/PET-RX pts who did not achieve a pCR received 6 cycles of TCHP and all pts from cohorts A/B completed 18 cycles of HP. Cohort C pts received 6 cycles of TCHP. Co-primary endpoints were breast/axilla pCR rate (ypT0/isN0) among cohort B/PET-RX pts and 3-year invasive disease-free survival (iDFS) in pts allocated to cohort B. Results: A total of 376 pts were included (71 pts in cohort A, 285 pts in cohort B, and 20 pts in cohort C). In cohort B, median age was 50 years, 49.2% had node-positive disease, and 67.4% had HR+ tumors. pCR in cohort A was achieved in 41 pts (57.7%, 95% CI 47.4-69.4%) and it was observed in 101 pts included in cohort B (35.4%, 95% CI 29.9-41.3%). Among cohort B pts, 227 (79.6%) were PET-RX; 86 of them (37.9%, 95% CI 31.6-44.5%) obtained a pCR. Among PET-non-RX pts, 15 (25.9%, 95% CI 15.3-39%) achieved a pCR after adding CT (TCHP). PET-RX pCR by HR status was 44.3% for HR[-] and 35% for HR[+] (p = 0.184). The incidence of commonly reported adverse events (AEs) was higher in pts allocated to cohort A (grade≥3 AEs 58.8 vs 12%; serious AEs 29.4 vs 4.6%). The rate of pts with a ≥10% global health status decline in cohorts A and B were 40.8 and 23.5%, respectively. Conclusions: F-PET identify pts with HER2[+] EBC who are more likely to benefit from CT-free dual HER2-blockade with HP. Follow-up is ongoing for iDFS endpoint. Depending on the results of this second co-primary endpoint, this strategy could select a group of HER2[+] EBC pts who would not need CT. Clinical trial information: NCT03161353.