Assessment of mouse tumour xenograft volumes in vivo by ultrasound imaging, MRI and caliper measurement

The purpose of the study was to compare the accuracy and precision of rodent tumour volumes estimated from caliper measurements (CM), from sonography (US) and from MRI. Studies were carried out on 27 female athymic mice bearing xenografts from a human gastric tumour line, MKN45. Twenty mice were examined by US and MRI, and tumour weight determined. Seven mice were imaged weekly to measure tumour growth. Tumour volumes were estimated from CM by assuming the rumours to be prolate spheroids and from US by assuming them to be ellipsoids. In MRI contiguous multislice data were acquired and volumes were estimated by pixel counting. MRI, US and CM each reported volumes which correlated very well with ex vivo mass. However, MRI was best for large rumours (>1 ml), probably because these deviate most from the ellipsoidal model, while CM gave poor results for small tumours (<0.05 ml), presumably because no correction was made for skin thickness. All three techniques reported exponential growth, but MRI data showed least deviation from the exponential.