Comparison of the significance of three histopathological thresholds of malignancy in experimental colorectal tumours.

The purpose of this experiment was to study sequential histogenesis of colonic epithelial tumours in the dimethylhydrazine (DMH) induced rat colon cancer model. Seventy outbred female Wistar rats treated with DMH (40 mg/kg, subcutaneously weekly for five weeks) were killed and autopsied in batches of 10 every five weeks from the 10th to 40th weeks from first treatment. The resulting 378 colonic lesions were assigned to benign or malignant categories using each of three standard histopathological thresholds of malignancy: alpha, the transition from dysplasia to intraepithelial carcinoma; beta, invasion through the crypt basement membrane; and gamma, invasion through the muscularis mucosae. These comprised 79 'benign' and 299 'malignant' or 273 'benign' and 141 'malignant' lesions depending on the threshold (alpha or gamma) assigned (p less than 0.001). Decreasing ratios of pre-threshold to post-threshold lesions between 15 and 40 weeks (alpha, 2.0 to 0.051; beta, 3.5 to 0.57; gamma, 8.0 to 0.87) provide some support for an 'adenoma-carcinoma' progression for each. Comparison of time-dependent prevalence curves confirms that the alpha threshold (cyto-architecture) is qualitatively different from the beta and gamma thresholds (invasion), showing that the adenoma-carcinoma and de novo hypotheses need not be mutually exclusive. The time-dependent prevalence data support de novo origin of some carcinomas, as well as at least two other modes of accrual for neoplastic lesions.