Performance of multi-modality molecular imaging of prostate cancer

Abstract

724

Objectives: Several non-invasive imaging techniques were assessed regarding their ability to detect primary prostate cancer (pPC) and whether they are associated with tumor aggressiveness.

Methods: Functional (diffusion-weighted, DW), precontrast 3T MRI with anatomical trans-pelvic (AP) coil and pre and post contrast anatomical 1.5 T MRI with endorectal coil (ERC) were performed together with C11-choline (CHOL) PET/CT in 13 patients with non-treated localized pPC on the same day. After radical prostatectomy and ex-vivo MRI of the prostatectomy specimen, imaging results were verified by whole mount histopathology (with local Gleason score) plus tissue microarray (TMA) analysis for the proliferation marker Ki67. Images were co-registered using mutual information software integrating all imaging modalities and pathology into a 3D voxel space for further analysis.

Results: From 30 tumor nodules identified by histology (volume range 0.01-8.8 ml), pelvic 3T MRI identified a 0.7 ml lesion as the smallest focus. From 14 foci >=0.7 ml, CHOL-PET/CT, DW-, AP- and ERC-MRI identified 6, 6, 7, and 8 foci, respectively. However, if all 3 MRI procedures were assessed combined, 13 of 14 foci were identified by at least one of the MRI modalities. CHOL-positive foci (but not DW-MR positive foci) were significantly correlated with the Gleason score (p<0.001). Ki67 scores available from 21 foci were varying between 0 and 25%. Increased Ki67 scores were preferentially seen in CHOL positive lesions, while the majority of foci were in fact Ki67 negative.

Conclusions: The combination of anatomic and functional MRI is superior to CHOL-PET/CT in identifying pPC. However, CHOL may serve as a marker of tumor aggressiveness offering the potential to select high-risk patients for intensified treatment.

Research Support: NIH P50 CA069568