Evaluation of Alternative In Vivo Drug Screening Methodology: A Single Mouse Analysis

Abstract

Traditional approaches to evaluating antitumor agents using human tumor xenograft models have generally used cohorts of 8 to 10 mice against a limited panel of tumor models. An alternative approach is to use fewer animals per tumor line, allowing a greater number of models that capture greater molecular/genetic heterogeneity of the cancer type. We retrospectively analyzed 67 agents evaluated by the Pediatric Preclinical Testing Program to determine whether a single mouse, chosen randomly from each group of a study, predicted the median response for groups of mice using 83 xenograft models. The individual tumor response from a randomly chosen mouse was compared with the group median response using established response criteria. A total of 2,134 comparisons were made. The single tumor response accurately predicted the group median response in 1,604 comparisons (75.16%). The mean tumor response correct prediction rate for 1,000 single mouse random samples was 78.09%. Models had a range for correct prediction (60%-87.5%). Allowing for misprediction of ± one response category, the overall mean correct single mouse prediction rate was 95.28%, and predicted overall objective response rates for group data in 66 of 67 drug studies. For molecularly targeted agents, occasional exceptional responder models were identified and the activity of that agent confirmed in additional models with the same genotype. Assuming that large treatment effects are targeted, this alternate experimental design has similar predictive value as traditional approaches, allowing for far greater numbers of models to be used that more fully encompass the heterogeneity of disease types. Cancer Res; 76(19); 5798-809. ©2016 AACR.

Figure 1

A. Distribution of deviation for 2134 observations. Single mouse prediction of response was compared to the median response for groups of tumor-bearing mice (solid tumors n=10; ALL models n=8). A score equaling zero indicates accurate prediction, whereas +1 and -1 refer to over- and under-prediction by one response classification. Panel B. Four ‘best’ tumor lines (F2, OS-2; F10, OS-33; F11, OS-31 osteosarcomas and B10, Rh65 rhabdomyosarcoma). Panel C. Four ‘worst case’ tumor lines (C5, BT-36 and C6, BT-41 ependymomas; G2, All-3 (primary B cell Precursor ALL) and G6, All-16 (primary T cell ALL)

Figure 2

Examples of over- and under-prediction by single mouse tumors. Left panels show growth of individual untreated (control) tumors. Right panels show growth of treated individual tumors. The tumor chosen at random (single mouse) is shown in the gray broken line (arrowed). For solid tumors, tumor volume (cm³) is plotted against time, and for ALL-16 the percent human CD45-positive cells in peripheral blood is plotted against time. Treatments start on day 0.

Figure 3

A. Objective response rates (ORR) were calculated for all tumor models tested for a particular drug (n=1 to n=55) for different studies, based upon the Group Median response; (red) responses predicted from a randomly chosen single mouse are plotted against group median response; (blue) the Single Mouse ORR mean ORR correlation based on 1,000 single mouse samples.

Figure 4

Objective response data presented as ‘waterfall’ plots for the ‘worst’ four studies using molecularly targeted agents. Each bar represents the response of a single tumor model. Left panels show distribution of responses based upon a single mouse analysis. Right panels show the median group response based on 8–10 mice per group. Colors indicate different tumor types.