Search

Introduction: The multileaf collimator (MLC) now commercially available as an integral or optional retrofit component of linear accelerator heads, permitting to shield automatically irregular fields by computerized movements of multiple tungsten leaves. In the present paper we discuss the main characteristics of different MLC versions, MLC dosimetric drawbacks and the clinical fields where the MLC could he widely used., Materials and Methods: Since February, 1995, we have studied the dosimetric characteristics and clinical implications of a multileaf collimator of a Clinac 2100 C/D linear accelerator used to replace conventional low melting alloy blocks. The scalloping effect of isodoses and the effective penumbra produced by the multileaf collimator in an irregular field were analyzed accurately. Secondly, radiation leakage through tungsten leaves was measured and compared with the values of low melting alloy blocks. Finally, the MLC was extensively used in clinical practice for the radiotherapy of different tumors., Results: Different dosimetric steps were followed to obtain the monitor units/dose ratio. Our single MLC-shielded irregular fields measurements also showed several physical and dosimetric disadvantages related to wider effective penumbra than with conventional shielding when the angle between field margin and the normal to the direction of leaf travel is 45 degrees. In clinical practice, the MLC can be widely used for the conformal radiotherapy of pelvic and thoracic tumors. Conventional low melting alloy blocks should be replaced with MLC for radiotherapy of selected brain and head and neck cancers., Discussion: The current use of a multileaf collimator improves both the accuracy and the effectiveness of radiation therapy and reduces the time for every treatment dose, which potentially increases the number of patients treated every day. The multileaf collimator is presently an important technical tool either to replace conventional shielding for static conformational radiotherapy or to administer 3D-planned dynamic radiotherapy.

Adequate irradiation of regional lymph-nodes in selected patients seems to improve disease-free and over-all survival rates. Technically, regional lymph-nodes are generally treated separately (supra-infra clavicular, apex of axilla and internal mammary chain-IMC). About the treatment of the IMC nodes, there are two main problems: the identification of the individual target volume, and its correct irradiation. For the latter problem, the most employed solutions are: the inclusion of IMC in the tangential fields, the use of the direct or oblique electron beam for IMC alone. None of them is optimal regarding to the tolerance of adjacent critical organs and from a geometric and dosimetric point of view. In this paper we propose and discuss a modified method. The IMC nodes together with the breast are irradiated by a three field technique (two tangential photon beams and one oblique electron beam). Their arrangement allows to treat every part of the target by two fields, reducing the risk of hot and cold spots. This is obtained by extending the internal tangential photon beams to cover the electron beam. Wedges and different weights must be used to achieve an uniform dose distribution.

The authors have addressed the problem of dose distribution calculation accuracy when irregularly shaped electron beams are used and structural dishomogeneities are present. The mathematical model used for the electrons from the "scattering foil" and from the colimator walls is the "gaussian pencil beam", and the total dose is obtained by adding to the above mentioned components the dose from the electrons diffused by the field defining frame and from bremsstrahlung photons. Depth dose curves, with and without frames, are compared with a satisfactory correspondence between the calculated and the measured curves. The calculation method is considered accurate within 2% of dose and 2 mm in position. When dishomogeneous structures are present and treated as semi-infinite layers, corrections are referred in 2-D to a single anatomical section and can be extended to 3-D if a sufficient number of sections is available for the whole volume definition. However current algorithms need to be improved when there is the possibility to use a much more powerful hardware.