Figure 4

Bias assessment of each method for tumour volume calculation. HF-US volumes show less bias than in vivo calliper volumes when compared to ex vivo calliper volumes. The ratios of HF-US to ex vivo calliper volumes (y-axis) were compared to the average value of the measurements (x-axis). Bland-Altman plots were generated comparing the bias between HF-US and ex vivo calliper volumes (a) and in vivo calliper volumes compared to ex vivo calliper volumes (b). The solid line denotes the bias (the average of the differences between the two measurement methods) and the dashed lines define the 95% confidence limits. The dotted line defines zero. HF-US could detect much smaller tumour volumes than callipers therefore n = 11 in (a) and n = 10 in (b).