Endoscopic versus radiology-based location of rectal cancer

Rigid proctosigmoidoscopy is recommended for measuring the height of rectal neoplasms but appears to be performed in only a minority of patients. Our aim was to compare endoscopic and radiological measurement of rectal tumour location with a focus on differentiation between mid and high rectal cancer.

Medical records of 66 rectal cancer patients were reviewed. Tumour location defined at colonoscopy (66 patients), rigid proctosigmoidoscopy (20 patients) and endorectal ultrasound (35 patients) was recorded. Rectilinear and curvilinear methods were used to estimate the distance between the lower tumour level and the anal verge on sagittal CT or MR images (66 patients). Agreement, intra- and inter-observer variation of radiology-based measurements were -assessed using intra-class correlation (ICC) and within-subject coefficient of variation (WSCV).

Tumour location was performed at rigid proctosigmoidoscopy in 30% of patients. Intra- and inter-observer agreement for radiology-based measurements were high. Tumour location using the rectilinear method or proctosigmoidoscopy was similar on average, for a difference of only 0.34 cm (SD 2.0 cm, p = 0.330), although agreement was -moderate (ICC = 0.54, WSCV = 16.7%). Measurements based on colonoscopy and the curvilinear radiological method were -characterized by a systematic overestimation of the location, increasing with tumour height.

Radiology-based measurement of the lower tumour level is a reproducible alternative for tumour location at rigid or flexible endoscopy. Its validity should be further assessed.