Diffuse reflectance spectroscopy can differentiate high grade and low grade prostatic carcinoma

Abstract

Prostate tumors are graded by the revised Gleason Score (GS) which is the sum of the two predominant Gleason grades present ranging from 6-10. GS 6 cancer exclusively with Gleason grade 3 is designated as low grade (LG) and correlates with better clinical prognosis for patients. GS >7 cancer with at least one of the Gleason grades 4 and 5 is designated as HG indicate worse prognosis for patients. Current transrectal ultrasound guided prostate biopsies often fail to correctly diagnose HG prostate cancer due to sampling errors. Diffuse reflectance spectra (DRS) of biological tissue depend on tissue morphology and architecture. Thus, DRS could potentially differentiate between HG and LG prostatic carcinoma. A 15-gauge optical biopsy needle was prototyped to take prostate biopsies after measuring DRS with a laboratory fluorometer. This needle has an optical sensor that utilizes 8×100 μm optical fibers for tissue excitation and a single 200 μm central optical fiber to measure DRS. Tissue biopsy cores were obtained from 20 surgically excised prostates using this needle after measuring DRS at 5 nm intervals between 500-700 nm wavelengths. Tissue within a measurement window was histopathologically classified as either benign, LG, or HG and correlated with DRS. Partial least square analysis of DRS identified principal components (PC) as potential classifiers. Statistically significant PCs (p<;0.05) were tested for their ability to classify biopsy tissue using support vector machine and leave-one-out cross validation method. There were 29 HG and 49 LG cancers among 187 biopsy cores included in the study. Study results show 76% sensitivity, 80% specificity, 93% negative predictive value, and 50% positive predictive value for HG versus benign, and 76%, 73%, 84%, and 63%, for HG versus LG prostate tissue classification. DRS failed to diagnose 7/29 (24%) HG cancers. This study demonstrated that an optical biopsy needle guided by DRS has sufficient accuracy to differentiate HG from LG carcinoma and benign tissue. It may allow precise targeting of HG prostate cancer providing more accurate assessment of the disease and improvement in patient care.