Ultrasound and magnetic resonance imaging for group stratification and treatment monitoring in the transgenic adenocarcinoma of the mouse prostate model.

Background The transgenic adenocarcinoma of the mouse prostate (TRAMP) is a widely used genetically engineered spontaneous prostate cancer model. However, both the degree of malignancy and time of cancer onset vary. While most mice display slowly progressing cancer, a subgroup develops fast-growing poorly differentiated (PD) tumors, making the model challenging to use. We investigated the feasibility of using ultrasound (US) imaging to screen for PD tumors and compared the performances of US and magnetic resonance imaging (MRI) in providing reliable measurements of disease burden. Methods TRAMP mice (n = 74) were screened for PD tumors with US imaging and findings verified with MRI, or in two cases with gross pathology. PD tumor volume was estimated with US and MR imaging and the methods compared (n = 11). For non-PD mice, prostate volume was used as a marker for disease burden and estimated with US imaging, MRI, and histology (n = 11). The agreement between the measurements obtained by the various methods and the intraobserver variability (IOV) was assessed using Bland-Altman analysis. Results US screening showed 81% sensitivity, 91% specificity, 72% positive predictive value, and 91% negative predictive value. The smallest tumor detected by US screening was 14 mm(3) and had a maximum diameter of 2.6 mm. MRI had the lowest IOV for both PD tumor and prostate volume estimation. US IOV was almost as low as MRI for PD tumor volumes but was considerably higher for prostate volumes. Conclusions US imaging was found to be a good screening method for detecting PD tumors and estimating tumor volume in the TRAMP model. MRI had better repeatability than US, especially when estimating prostate volumes.