In recent years intensity modulated radiation therapy with photons (xIMRT) has gained attention due to its ability to reduce the dose in the tissues close to the tumour volume. However, this technique also results in a large low dose volume. Electron IMRT (eIMRT) has the potential to reduce the integral dose to the patient due to the dose fall off in the electron depth dose curves. This dose fall off makes it possible to modulate the dose distribution in the direction of the beam by selecting appropriate electron energies. The use of a computer based energy selection method was examined in combination with the IMRT technique to optimise the electron dose distribution. It is clearly illustrated that the energy optimisation procedure reduces the dose to lung and heart in a breast cancer treatment. To shape the multiple electron subfields (beamlets) that are used in eIMRT, an electron multi leaf collimator (eMLC) is needed. However, photons produced in a conventional electron treatment head could penetrate such an added eMLC, thus producing an undesirable dose contribution. The leakage levels normally achieved are acceptable for standard single electron field treatments but could become unacceptably high in eIMRT treatments where a lot of small subfields are combined. To limit this photon contribution, the photon MLC (xMLC) was used to shield off large parts of the photon leakage. The effect of this xMLC shielding on the reduction of photon leakage, the electron beam penumbras, and electron output (dose level), was studied using Monte Carlo methods for different electron treatment head designs. The use of helium as a mean to reduce the electron scatter in the treatment head, and thus the perturbating effect of the xMLC on electron beam penumbra and output, was also investigated. This thesis shows that the effect of the xMLC shielding on the electron beam penumbra and output can be made negligible while still obtaining a significantly reduced x-ray leakage dose contribution. The result is a large gain in radiation protection of the patient and a better dynamic range for the eIMRT dose optimisation. For this optimisation a computer based electron energy selection method was developed and tested on two clinical cases.