Sofroni_Ervis.pdf (2.36 MB)

Tissue characterization of prostate cancer using quantitative analysis of low frequency ultrasound

Download (2.36 MB)
posted on 22.05.2021, 14:41 by Ervis Sofroni
Current accepted methodologies used for detection of the prostate rumor involve measurements of prostate specific antigen (PSA) levels, patient age followed by ultrasound guided biopsies leaving a lot to desire in the ability to correctly identify lesions. Also PSA level test has been shown to produce a high number of false positives leading to unnecessary invasive biopsies. The goal of this thesis is to investigate the use of trans-rectal conventional low frequency (1-10MHz) ultrasound as a non-invasive imaging modality for the detection of prostate tumors. Currently we are investigating the use of multiparameter spectroscopic analysis of the ultrasound radio frequency signal in combination with ultrasound elastrographic imaging of the prostate and correlating the results with whole-mount histopathology from radical prostatectomy. Ten patients with prostate cancer prior to surgery were subjected to trans-rectal conventional low frequency ultrasound scans. Parametric maps are generated for each individual spectral parameter. Ratios of disease area versus normal prostatic tissue are identified using low frequency ultrasound and compared with the equivalent ratios obtained from whole-mount histopathology. Preliminary results show that areas of suspected disease identified by spectral parameters correlate with areas of disease presence in the corresponding whole-mount sections. An initial software platform performing visualization of areas of disease based on parametric maps generated from spectral analysis methods was developed.





Master of Science


Computer Science

Granting Institution

Ryerson University

LAC Thesis Type


Thesis Advisor

Alireza Sadeghian

Usage metrics

Computer Science (Theses)