Is correction for lung density in radiotherapy treatment planning necessary?

From 1978-981 a series of 30 patients with cancer of the esophagus were treated at the National Cancer Institute. Each of these patients had a CT scan of the chest taken in the treatment position, but prior to any treatment being given. Using these scans a retrospective analysis of the effect of lung density on delivered dose was performed. This indicated that failure to correct for tissue inhomogeneity results in a much higher dose being delivered than is prescribed. This effect is dependent on the energy of the beam being used for treatment, it may exceed 30% for 60Co. It also showed that there is wide patient to patient variation in lung density and that this variation is nonrandomly distributed. The average lung density in his group of patients was 0.21 compared to the standard estimate of 0.35 but some had densities substantially lower than this, these being the patients with the largest lung volumes. This variability acts to further increase the discrepancy between prescribed and delivered dose even in a very homogeneous group of patients being treated under identical conditions for the same malignancy. The implications of this for future clinical trials in thoracic malignancies are discussed.