This work reports a practical method for the determination of the in vivo breast middle dose value, D(m) on the beam central axis, using a signal S(t), obtained by a small thimble ion chamber positioned at the center of the electronic portal imaging device, and irradiated by the x-ray beam transmitted through the patient. The use of a stable ion chamber reduces many of the disadvantages associated with the use of diodes as their periodic recalibration and positioning is time consuming. The method makes use of a set of correlation functions obtained by the ratios S(t)/D(m), determined by irradiating cylindrical water phantoms with different diameters. The method proposed here is based on the determination of the water-equivalent thickness of the patient, along the beam central axis, by the treatment planning system that makes use of the electron densities obtained by a computed tomography scanner. The method has been applied for the breast in vivo dosimetry of ten patients treated with a manual intensity modulation with four asymmetric beams. In particular, two tangential rectangular fields were first delivered, thereafter a fraction of the dose (typically less than 10%) was delivered with two multi leaf-shaped beams which included only the mammarian tissue. Only the two rectangular fields were tested and for every checked field five measurements were carried out. Applying a continuous quality assurance program based on the tests of patient setup, machine settings and dose planning, the proposed method is able to verify agreements between the computed dose D(m,TPS) and the in vivo dose value D(m), within 4%.