Prostate Cancer Diagnosis by Multiparametric Ultrasound

Prostate cancer (PCa) is the most frequently diagnosed cancer among American males, accounting for 20% of new cancer diagnoses, but PCa is directly responsible for only 9% of cancer related deaths. PSA screening facilitates the detection of PCa, but many of the cancers detected by PSA screening are low to intermediate grade, resulting in overtreatment of non-aggressive PCa. Recent studies suggest an urgent need to selectively identify “clinically significant” PCa that will result in cost-effective treatment of PCa. In the past few years, multiparametric MRI (mp-MRI) has become the standard of care for diagnosis of aggressive PCa. Although mpMRI guided biopsy can selectively detect higher grade PCa lesions, several recent studies acknowledge that a considerable number of clinically important lesions are missed by mp-MRI. Furthermore, given the high cost of MRI, the geographic variability in the availability of MRI systems, the inexact methods used to register and fuse MRI with ultrasound for biopsy, and the dangers of Gadolinium in patients with renal insufficiency, there is a need for a more cost-effective, alternative to detect high-grade PCa. Recent studies have demonstrated that contrast-enhanced ultrasound (CEUS) can selectively detect patients with “clinically significant” PCa. New methods have been proposed to enhance conventional ultrasound detection of PCa, including subharmonic imaging (SHI), contrast-enhanced ultrasound dispersion imaging (CUDI) and viscoelastography. The objective of this project is to develop a system capable of a multiparametric combination of ultrasound techniques (mpUS) for detection of “clinically significant” PCa. We will compare 3D mp-US and mp-MRI to identify significant PCa (defined as: Gleason score ≥ 7, a single core with > 50% involvement, or > 25% of biopsy cores positive for PCa or as Gleason grade group ≥ 2, PSA > 10 or clinical stage of cT2b or worse). The first year of the study will be dedicated to implementation of our mp-US technique on a commercially available 3D transrectal probe, and to adapting a MRI-based fusion/registration system to mp-US. During the second year, we will optimize mp-US with a machine learning approach, based upon pathologic correlation with 50 radical prostatectomy patients. The final 3 years of the study will be dedicated to a clinical trial with 300 participants suspected of having PCa. Each participant will receive an experimental intervention in the form of biopsy of up to 3 suspicious areas identified by transrectal ultrasound evaluation of the prostate with mp-US. For comparison, a maximum of 3 targeted biopsy cores will also be obtained from each participant, based on mp-MRI (the clinical intervention). Following the targeted biopsy, each participant will also receive a systematic biopsy (standard of care) consisting of 6 laterally directed biopsy cores and 6 medially directed biopsy cores. The study is powered to demonstrate that targeted biopsy based upon mp-US is not inferior to targeted biopsy based upon mp-MRI.