Purpose Deep inspiration breath hold (DIBH) in photon therapy as well as proton therapy (in free breathing) has been suggested to reduce cardiopulmonary and integral dose without compromising target coverage. The goal is to investigate the dosimetric impact on proton treatment plans of set-up variation and daily anatomy visualized on CBCT prior to treatment in patient with left-sided breast cancer patients with an indication for locoregional irradiation. Method For 10 patients, proton plans (4.05 Gy in 15 fractions) were made in Erasmus iCycle with a robustness setting of 5 mm and 3% range uncertainties. The CTV included the left breast (n=5) or chestwall (n=5), axillary lymphnodes levels 1-4 and internal mammary chain (IMC). The daily estimated dose was calculated on five CBCT for proton plans of 10 patients treated for left-sided breast cancer with image guided radiotherapy and surface guidance to facilitate DIBH. A workflow in MIM was created to generate virtual CTs by deforming the pCT into the CBCT. Dosimetric parameters were evaluated for four separate CTVs; CTV Breast/Chestwall, axillary lymph node levels L1+L2, L3+L4 and left IMC and included the volume, D98, V95 and V107. For OAR this included the volume, Dmean, D2cc and V5Gy for the heart and V5gy and V20Gy for the left lung. To determine whether there was a significant difference between the DVH parameters and volume between the pCT and the mean estimated dose of five fractions a wilxocon signed rank test or two tailed t-test was used. In addition, contour deformation was validated with the Dice similarity score and mean distance to agreement, and it was tested whether there was a significant difference in DVH parameters for CTV chest wall/Breast, CTV L1-L2, and heart between contours derived by the method and contours manually corrected by an experienced breast radiation oncologist on the CBCT. Results The mean and standard deviation (std) values observed in this study for the DSC and MDA are 0.98 (0.1) and 0.30 (0.2), respectively, which is in agreement with the proposed tolerances for both evaluation tools. Moreover, no significant difference was found in volume and DVH parameters between the semi-automatic contours derived from the method and the manually corrected contours. A significant difference in volume with a small absolute volume for CTV breast, Chestwall and L1-L2 was observerd and no significant difference in any of the DVH parameters for CTV breast, chestwall, L1-L2, L3-L4 and IMC. The cardiac dose did not have a significant difference in DVH parameters, but the left lung had a significant increase in V5Gy[cc] and V20Gy [cc]. Conclusion Proton treatment plans created in Erasmus-iCycle on planning CT in DIBH for patients with left-sided breast cancer with indication for irradiation of the breast/chestwall, axillary lymph nodes and IMC achieve good coverage in the mean estimated delivered dose of five fractions treated with image guided radiotherapy and surface guidance for DIBH.