Background: Human epidermal growth factor receptor 2 (HER2) status is an important predictive biomarker in breast cancer (BC). Tumor heterogeneity has been described, with changes in HER2 expression levels between lesions and over the disease course. HER2 expression is assessed on tissue biopsies, at primary diagnosis and in metastatic lesions. A whole-body imaging technique such as PET/CT could help understand expression levels in different lesions. A 68Ga-labeled single domain antibody (sdAb) targeting the HER2 receptor has been developed and proven safe (Keyaerts et al., 2016). Imaging is performed at 90 min post-injection (pi). We report results of a phase II trial to assess the repeatability of the technique in 20 patients and the correlation of tracer uptake with HER2 tissue expression of the lesions present at the time of imaging. Methods: Twenty patients (pts) with a locally advanced or metastatic BC with at least one lesion of minimum 12 mm were included. Pts were injected intravenously with a typical protein mass of 100 µg and a radioactive dose ranging from 98-168 MBq 68GaNOTA-anti-HER2 sdAb. PET/CT images were obtained at 90 min pi. A second tracer injection followed by PET/CT was done with a maximal interval of 8 days. To assess repeatability, up to 5 lesions per pt were selected, with no more than 2 in a single organ. Peak Standard Uptake Values (SUVpeak) of the lesions were measured on both scans and compared with a t-test and Bland-Altman Plots. Images were compared to other available medical or imaging data and interpreted considering the subject’s disease course. Serum and plasma samples were collected before injection and between 60 and 365 days pi and stored for future detection of anti-drug antibodies (ADA) and liquid biopsies analysis for the presence of HER2 amplification. Tissue samples were assessed by central labs using mass spectrometry, immunohistochemistry and in fluorescence situ hybridization. Results: Twenty women with BC (6 HER2+, 14 HER2-) with a mean age of 58.6 y (37-81) were included. Three pts were scanned only once (2 due to withdrawal of consent, 1 due to covid pandemic). Repeatability of the technique was visually scored as excellent. For quantification, 50 lesions were compared on both scans in 17 pts without significant differences between the two measurements (p=0.40). The repeatability coefficient (RC) was 38.2%. The mean absolute percentage difference (MAPD) was 13.6%, comparable to repeat values reported for 18F-FDG. In 3 out of 6 HER2-positive (HER2+) patients, lesions showed high uptake, even better visible than using 18F-FDG in 2 of them. In 2 HER2+ subjects with a negative scan, lesions were confirmed to be true negatives: one patient did not relapse from BC but had tuberculosis; the other was confirmed to have a radiopneumonitis after radiotherapy and no relapse. In 1 HER2+ patient, the uptake was unexpectedly low. However, the HER2 status was also not reconfirmed in the metastatic setting for this subject. In 1 HER2-negative patient, the tumor HER2 status was changed from negative to positive based on a subsequent image-guided biopsy performed in this study. High tracer uptake was also seen in many of the patients presenting with HER2-low BC (IHC 1+ or 2+), indicating the potential of the tracer to detect low-level HER2 expression. Additional correlation to centrally performed tissue and blood analysis is ongoing. Conclusion: 68GaNOTA-Anti-HER2 PET/CT shows high uptake in HER2-expressing BC lesions but also in HER2-low lesions. The technique shows good repeatability and, in some cases, even better sensitivity than 18F-FDG PET/CT. Specificity was confirmed in relapse-free lesions such as tuberculosis and radiopneumonitis. Its sensitivity makes it a promising technique to assess HER2+ and HER2-low lesions in BC patients.