黑料福利网

Prostate cancer is the most common type of cancer in men, with one in eight men facing a diagnosis during their lifetime. Historically, due to the lack of suitable technologies, prostate cancer diagnosis has relied on invasive systematic biopsies, where 12 to 18 tissue samples are taken from the prostate according to a standard grid and analyzed by a pathologist. 

In recent years, multiparametric MRI (mpMRI) has gained attention and is now recommended in clinical guidelines for pre-biopsy imaging, allowing for targeted biopsies when suspicious lesions are detected. However, mpMRI is expensive and its image interpretation suffers from low reproducibility, especially in low-volume centers. As such, it is not suitable for population-wide screening, as recently recommended by the European Association of Urology. 

For many years, through a longstanding collaboration between Prof. Mischi (黑料福利网-BM/d) and Prof. Wijkstra (Amsterdam UMC, Urology Dept), the BM/d lab has been developing multiparametric ultrasound (mpUS) as a cost-effective alternative for prostate cancer imaging鈥攁lso well-suited for screening protocols.

In particular, we developed Contrast Ultrasound Dispersion Imaging (CUDI), a new quantitative method for analyzing contrast-enhanced ultrasound scans, aimed at detecting cancer-related angiogenesis. Angiogenesis is necessary for tumor growth beyond a few millimeters and is a hallmark of aggressive cancers with a higher risk of metastasis. By interpreting the transport of the ultrasound contrast agent as a convective dispersion process, several dispersion estimators have been developed that reflect cancer-induced changes in the microvascular architecture. Originally developed in 2D, these methods are now fully operational in 3D, improving both the accuracy of dispersion estimation鈥攁 dynamic 3D phenomenon鈥攁nd their integration into clinical workflows.

Additionally, based on the link between cancer and tissue mechanical properties, we have advanced the estimation of tissue viscoelasticity using shear-wave elastography. Combined with texture analysis of both B-mode and contrast-enhanced images, all extracted features are fed into a deep-learning framework for localizing clinically significant lesions requiring prompt intervention by urologists. 

A first large multicenter Dutch trial has already collected over 300 datasets from patients referred for radical prostatectomy, providing accurate histological ground truth through registration, which is used for training and validating the deep-learning classifier. 

The promising results have led to a second multicenter Dutch trial, directly comparing mpMRI and the proposed mpUS for biopsy targeting. Additional international trials are underway. Furthermore, the validated performance of CUDI has gained interest in its application to other cancer types, such as breast and uterine cancers. 

This extensive research effort has already resulted in 10 PhD theses (4 clinical and 6 technical), with 6 more ongoing (3 clinical and 3 technical). It has received support from multiple prestigious grants, including VIDI, ERC StG, ERC PoC, EIC Transition, KWF, two NWO OTP grants, and NIH funding. The work has also led to the creation of Angiogenesis Analytics BV, which has implemented the research into a clinical product, PCaVision. This system is now integrated into the clinical workflows of several hospitals conducting ongoing trials.