What is the biggest mistake people make when using IMRT to treat cervical cancer?

I can't say for sure what the biggest mistake people make is, but some common issues I see when reviewing others' contours are:

1. Using insufficient margins around the vessels when contouring the nodal volume (CTV)
2. Using insufficient planning margins around the vaginal cuff (postop) or cervical mass/uterus (definitive).

Optimal methods for CTV delineation and for managing daily variation in target position are matters of considerable debate among experts, but consensus guideline papers (Small et al., Lim et al.) have generally advocated at least 7 mm margins around the vessels, sometimes larger to ensure nodal fat pads are fully encompassed, with modifications to cut out expansions into muscle or bone.

Failure to adequately encompass the obturator nodes is another common problem I see. Around the cuff/cervix, a 10-15 mm planning margin is generally recommended, to account for changes in bladder/rectal filling. My approach is to delineate a cuff or cervix/uterine ITV based on full and empty bladder scans. I add a 15 mm planning margin around the cuff or cervix/uterus on the scan used for planning, then take the union with the ITV (but ITV usually lies within this margin). I treat with an empty bladder and use daily kV/kV for setup and verify the cuff/cervix position using soft tissue imaging (e.g. CBCT).

With this approach a shift is rarely necessary but in some cases I have had to take patients off the treatment table or re-sim them due to a distended rectum. I think the important thing is not to be too conservative in the target delineation, because good normal tissue sparing is still possible even with generous volume delineation, and narrow delineation unnecessarily risks underdosing the target and/or will cause frequent interruptions for daily QA/verification.