The response of a human tumour xenograft to chemotherapy: intrinsic variation between tumours and its significance in planning experiments.

The HT29R human colon adenocarcinoma cell line has been used to investigate the problem of inter-tumour variability in the xenograft system and how this may affect the planning of chemotherapy experiments. Subcutaneous inoculation of 10(6) in vitro cells into immuno-suppressed mice yielded an 80% take rate of moderately poorly differentiated, mucin-secreting tumours with a mean doubling time of 6 days. Statistical estimates from experiments with this system demonstrated that whereas a delay of one doubling time in treated compared with control groups could be detected with relatively small animal numbers, 3 to 4 times as many animals were needed to detect a delay of half a doubling time. The use of 2 tumours per animal yielded the same results as one tumour per animal but reduced the number of animals needed to achieve the same degree of statistical significance.