PurposeWe compared the setup errors determined by an optical imaging system (OSIS) in women who received breast‐conserving surgery (BCS) followed by whole‐breast radiotherapy (WBRT) with those from cone‐beam computed tomography (CBCT) carried out routinely.MethodsWe compared 130 setup errors in 10 patients undergoing WBRT following BCS by analyzing the translational and rotational couch shifts via CBCT and OSIS. Patients were treated with intensity‐modulated radiotherapy (IMRT). The patient outline extracted from the planning reference Computed tomography (CT) was used as the reference for OSIS and CBCT alignment during treatment. We detected the setup uncertainty using CBCT and OSIS at the first five fractionations of RT and then twice a week.ResultsThe absolute translational setup error (mean ± Standard deviation (SD)) in x (lateral), y (longitudinal), and z (vertical) axes detected by the OSIS was 0.14 ± 0.18, 0.15 ± 0.14, and 0.13 ± 0.13 cm, respectively. The rotational setup error (mean ± SD) in Rx (pitch), Ry (roll), and Rz (yaw) axes was 0.77 ± 0.54, 0.76 ± 0.61, and 1.23 ± 0.95, respectively. Significant difference is observed only in one direction (Rx, p = 0.03) in the paired setup errors obtaining from OSIS and CBCT, without significant differences in five directions.ConclusionOSIS is a repeatable and reliable system that can be used to detect misalignments with accuracy, which is capable of supplementing CBCT for WBRT after BCS. We believe that an OSIS may be easier to use, quicker, and reduce overall dose as this method of patient alignment does not require ionizing radiation.