Prognostic factors and in vitro cytotoxic sensitivity in colorectal cancer.

The poor response rates to chemotherapy for colorectal cancer justify attempts to rationalize selection of patients for treatment, and the development of systems to evaluate new cytotoxic agents. Refinement of prognostic indices may identify colorectal cancer patients at a higher risk of recurrence who merit more aggressive treatment. We report our experience with the stem cell assay and pulse thymidine labelling in 43 primary colorectal cancers. Thirty-six tumours were evaluable, and clonogenic growth was obtained in 30 (83 per cent). In 24 tumours (67 per cent) growth was adequate for meaningful interpretation of a cytotoxic drug assay. Frequency of growth and colony forming efficiency did not correlate with histopathological grade, Dukes' stage or tumour cell kinetic indices. Thymidine labelling indices correlated with Dukes' stage (A and B versus C and D, P less than 0.01, Mann-Whitney U test). Cytotoxic assays with 5-fluorouracil and 5'-deoxy-5-fluorouridine were undertaken in 18 cases (14 primary carcinomas, 4 malignant ascites), of which 14 were evaluable and 3/14 (21.5 per cent) were chemosensitive in vitro. Both drugs were equally effective in vitro at clinically attainable plasma concentrations. This is in accordance with the response rates observed clinically with 5-FU chemotherapy in colorectal cancer.