A simple and robust method for in vivo midline dose map estimations using diodes and portal detectors

Introduction: This work investigates the possibility of using a pair of dio des on the beam axis in conjunction with a portal imaging detector to estim ate in vivo midline dose distributions, without any additional patient info rmation, related to the external body contour. Materials and methods: In the proposed method, the patient is considered eq uivalent to a parallelepiped phantom with a thickness z equal to the patien t's physical thickness on the field axis with a variable electronic density rho, depending on the water-equivalent thickness. Based on this assumption , if the air gap between portal detector and patient is kept small (within 10-15 cm), the relative exit dose map may be assumed to be equal to the cor responding map measured at the portal detector level by geometrical back pr ojection to the corresponding exit points. The relative exit dose map is th en normalized at the on-axis value measured by the exit diode. The entrance dose map is derived by correcting the absolute dose value measured with th e diode at the entrance surface by the off-axis ratios. For each pair of en trance and exit doses, the midline dose may be estimated by applying algori thms reported in literature. The method was tested in 6 MV beams using port al film as detector and the Huyskens and Rizzotti algorithms for midline do se estimation. Tests on homogeneous cubic phantoms, homogeneous phantoms wi th varying thickness symmetrically (simulating head and neck regions) and a symmetrically (simulating abdomen/pelvis region), and a half-sphere phantom with simulating the breast, were performed. Midline doses estimated with t he proposed method have been compared with corresponding ones measured by i onisation chamber. Results and discussion: Results confirm that the proposed method can be used to estimate midplane d ose maps within 2-3% for most clinically suitable situations. For homogeneo us symmetrical phantoms the agreement between estimated and measured midlin e doses decreases with the phantom-portal film distance, the field sizes an d the thickness. For homogeneous asymmetrical phantoms the percentage devia tions are generally within 3%. Discrepancies larger than 3% (up to 5-6%) ar e found only for 'stressed' irradiation geometries which are not linked wit h any clinical condition. Conclusions: The obtained results not only show the accuracy of the propose d method but, due to its simplicity, suggest a rapid clinical implementatio n of this method in relevant clinical situations such as head-neck, breast and abdomen/pelvis irradiation. Previous investigations which confirmed the possibility of using portal detectors for transit dosimetry in inhomogeneo us regions suggest the further exploration of the accuracy and the limits o f the proposed method in such cases. (C) 2001 Elsevier Science Ireland Ltd. All rights reserved.