Many observations on potential inadequate coverage of tumour volume at risk in advanced cervical cancer (CC) when conventional radiation fields are used, have further substantiated by investigators using MRI, CT or lymphangiographic imaging. This work tries to obtain three dimensional margins by observing enlarged nodes in CT scans in order to improve pelvic nodal chains clinical target volumes (CTVs) drawing, and by looking for corroborative evidence in the literature for a better delineation of tumour CTV. Eleven consecutive patients (seven males, four females, mean age 62 years, range 43-78) with CT diagnosis of nodal involvement caused by pathologically proved carcinoma of the cervix (n = 2), carcinoma of the rectum (n = 2), carcinoma of the prostate (n = 2), non-Hodgkin lymphoma (n = 2), Hodgkin lymphoma (n = 1), carcinoma of the penis (n = 1) and carcinoma of the corpus uteri (n = 1) were retrospectively reviewed. Sixty CT scans with 67 enlarged pelvic nodes were reviewed in order to record the more proximal structures (muscle, bone, vessels, cutis or subcutis and other organs) to each enlarged node or group of nodes according to the four surfaces (anterior, lateral, posterior and medial) in a clockwise direction. A summary of the observations of each nodal chain and the number of occurrences of every marginal structure on axial CT slices is presented. Finally, simple guidelines are proposed. Tumour CTV should be based on individual tumour anatomy-mainly for lateral beams as it results from sagittal T2 weighted MRI images. Boundaries of pelvic nodes CTVs can be derived from observations of enlarged lymph nodes in CT scans.