The objective was to investigate survival of breast cancer patients at Mulago Hospital. A retrospective study of the medical records of 297 breast cancer patients referred to the combined breast clinic housed in the radiotherapy department between 1996 and 2000 was done. The female/male ratio was 24 : 1. The age range was 22 -85 years, with a median of 45 years and peak age group of 30 -39 years. Twenty-three percent had early disease (stages 0 -IIb) and 26% had metastatic disease. Poorly differentiated was the most common pathological grade (58%) followed by moderately differentiated (33%) and well-differentiated (9%) tumours. The commonest pathological type encountered was 'not otherwise specified' (76%). Of all patients, 75% had surgery, 76% had radiotherapy, 60% had hormonotherapy and 29% had chemotherapy. Thirty-six (12%) patients received all the four treatment modalities. The 5-year survival probabilities (Kaplan -Meier) for early disease were 74 and 39% for advanced disease (P ¼ 0.001). The overall 5-year survival was 56%, which is lower than the rates in the South African blacks (64%) and North American whites (82 -88%).

Dosimetrical characteristics of 11 Varian a-Si-500 electronic portal imaging devices (EPIDs) in clinical use for periods ranging between 10 and 86 months were investigated for consistency of performance and portal dosimetry implications. Properties studied include short-term reproducibility, signal linearity with monitor units, response to reference beam, signal uniformity across the detector panel, signal dependence on field size, dose-rate influence, memory effects and image profiles as a function of monitor units. The EPID measurements were also compared with those of the ionization chambers' to ensure stability of the linear accelerators. Depending on their clinical installation date, the EPIDs were interfaced with one of the two different acquisition control software packages, IAS2/IDU-II or IAS3/IDU-20. Both the EPID age and image acquisition system influenced the dosimetric characteristics with the newer version (IAS3 with IDU-20) giving better data reproducibility and linearity fit than the older version (IAS2 with IDU-II). The relative signal response (uniformity) after 50 MU was better than 95% of the central value and independent of detector. Sensitivity for all EPIDs reduced continuously with increasing dose rates for the newer image acquisition software. In the dose-rate range 100-600 MU min(-1), the maximum variation in sensitivity ranged between 1 and 1.8% for different EPIDs. For memory effects, the increase in the measured signal at the centre of the irradiated field for successive images was within 1.8% and 1.0% for the older and newer acquisition systems, respectively. Image profiles acquired at a lower MU in the radial plane (gun-target) had gradients in measured pixel values of up to 25% for the older system. Detectors with software/hardware versions IAS3/IDU-20 have a high degree of accuracy and are more suitable for routine quantitative IMRT dosimetrical verification.