Kilovoltage Cone beam Computed Tomographic imaging (KV CBCT) is emerging as a new standard for on-treatment verification and Image Guided Radiotherapy. We have applied this technology to understand and quantify the interfractional and intrafractional changes, both in the bony anatomy and in the internal organs at different tumour sites. We further attempted to derive generic CTV to PTV (Clinical Target Volume to Planning Target Volume) margins. Prospective observational and interventional clinical studies using serial KV CBCTs were designed for prostate, distal oesophageal and rectal cancers. The effect of rectal volume changes on prostate position and dose were estimated. The important geometric uncertainties in distal oesophageal cancers were quantified and used for generating CTV to PTV margins. In rectal cancer, the relative efficacy and reproducibility of the use of the bellyboard device, Intensity Modulated Radiotherapy (IMRT) and their combination in sparing irradiated small bowel were evaluated. Rectal volumes decreased over the course of radiotherapy especially if the planning rectal volume is greater than 100cc, resulting in systematic shifts of the base of prostate and insufficient high dose coverage in 38% of fractions. Smaller planning rectal volumes (between 50 and 100cc) are more consistent. In distal oesophageal cancers, bony translational shifts in the cranio-caudal (CC) axis & rotations along the anetro-posterior axis and intra-fraction organ motion in the CC axis are dominant. Inter-fraction organ motion is also large. Our current CTV-PTV margins appear to be inadequate. Large inter-patient variations in the delivered doses are seen. In rectal cancer, use of both the bellyboard and IMRT result in significant small bowel dose reduction at most dose levels. Their combined use is better than either technique. CTV dose was not affected. There is significant variability in the bladder volumes in spite of bladder filling protocols over the treatment course.