Background and purpose: Recently, algorithms have been developed to derive the patient dose from portal dose measurements using a liquid-filled electronic portal imaging device. These algorithms have already been validated for several phantom geometries irradiated under clinical conditions. It is the aim of the present study to investigate the applicability of a liquid-filled electronic portal imaging device in combination with these algorithms for two-dimensional midplane dose verification in clinical practice. Measurements and methods: Portal dose images were obtained during several patient treatments under routine clinical conditions. Measurements were performed to verify the midplane dose during radiotherapy of larynx cancer with 4 MV beams, breast and lung cancer with 8 MV beams and prostate cancer with both 8 and 18 MV beams. Midplane doses, determined from portal dose measurements and analyzed with our algorithms, were compared with midplane doses calculated with our three-dimensional (3D) treatment planning system (TPS). Results: For the larynx treatment the measured 2D midplane dose agreed within 2.0% with TPS calculations in most parts of the field. Larger differences were found in a small region below the skin due to the absence of electron equilibrium, which is not taken into account in our portal dose analysis. For breast irradiations the measured midplane dose showed a homogeneous distribution in the AP direction in the axial plane, while high dose regions were observed in the cranial and caudal part of the breast. Portal dose measurements and TPS calculations agreed within 2.5% for most of the prostate and lung irradiations. For a few of the prostate and lung treatments larger local differences were found due to differences between the actual patient anatomy and the planning CT data, e.g. as a result of variable gas filling in the rectum and anatomical changes in the lung. Conclusions: Portal dose measurements with a liquid-filled electronic portal imaging device can be used to determine the 2D midplane dose for various treatment sites in clinical practice. Portal in vivo dosimetry has proven to be important in detecting changes in the patient's anatomy and its influence on the dose delivery. It is concluded that portal dosimetry is an excellent tool for accurate and independent verification of the dose in the entire (2D) midplane during patient treatment. However, a limited number of patients were involved in this study and the results are therefore preliminary. More research is needed to fully assess the clinical value of portal dose measurements.