An accurate prostate cancer prognosticator using a seven-gene signature plus Gleason score and taking cell type heterogeneity into account, "Beyond the Abstract," by Dan Mercola, MD, PhD

BERKELEY, CA (UroToday.com) -

Overcoming cell-type heterogeneity provides improved prognostic test for prostate cancer

Efforts to identify biomarkers from expression data of prostate cancer tumor cells following prostatectomy, or from biopsy tissue that would identify cases with aggressive disease likely to recur following prostatectomy, or to identify cases with more indolent disease, have been sought since expression arrays became widely available in the late 1990s. Numerous studies have identified profiles of altered gene expression that correlate with the known outcome. Few biomarkers have proven to be reliable across study populations or have survived rigorous validation studies, and therefore few have found their way into clinical use.

Some of the reasons for this lack of robustness is the polyclonal nature of prostate cancer, the small sample numbers of most studies, and the heterogeneity of cell-type composition of the samples used in most studies. Cell-type heterogeneity is particularly vexing because even when precautions are taken to study “pure” tumor samples, stroma is interwoven among the tumor epithelial cells.

This study employed a simple mathematical approach to determine the expression changes of the epithelial tumor cells. The basis of the method is to consider that the measured RNA-expression value is the sum of RNA contributed from the various cell types present in each sample. For relatively pure tumor samples, there are four major cell types: the epithelial tumor cells, the stroma cells, small amounts of epithelial cells from normal glands and BPH, and epithelial cells of dilated cystic glands. To determine the RNA contribution from each cell type, knowledge of the relative proportion of each cell type is needed. The study used samples for which these proportions were determined by examination of histological sections, by four pathologists. The total RNA for a gene, or a value proportional to it ,can then be considered the arithmetic sum of four contributions, where each contribution is the percentage of a particular cell type times a number that is the ability of that cell type to make RNA for the particular gene. For a large number of samples, the average “ability” of the tumor epithelial cells to make the RNA – the desired quantity – may be determined by a standard linear regression analysis.

The study examined whether the ability of tumor cells to express RNA for a particular gene varied greatly for tumors from patients who exhibited relapse after prostatectomy for organ-confined asymptomatic cases compared to similar patients who did not relapse for at least four years following prostatectomy. Seven genes fit these criteria. The ability of the profile to identify relapse from non-relapse cases, for independent cases not used in the development of the profile, was tested with accuracy of 71%. When combined with Gleason score, the accuracy is improved and is 83%. The pre-surgery accuracy of other methods such as the Kattan nomogram for the probability of remaining disease-free for five years is in the range 75-80%. Moreover, of all the common predictive clinical variables tested, such tumor volume, pre-op PSA, pathological stage, surgical margins status, etc., only the Gleason score provided an improvement -- indicating these other variables do not provide additional information. Thus, correcting for cell-type heterogeneity provides a simple seven-gene profile with performance in a range that may be clinically useful. Large scale validation and application to formalin-fixed paraffin-embedded tissues will be required to translate the new profile to a useful clinical assay.

 

Written by:
Dan Mercola, MD, PhD as part of Beyond the Abstract on UroToday.com. This initiative offers a method of publishing for the professional urology community. Authors are given an opportunity to expand on the circumstances, limitations etc... of their research by referencing the published abstract.

Professor, Pathology and Laboratory Medicine
Director, Translational Cancer Biology
University of California at Irvine
Medical Sciences I, D-440
Irvine, CA 92697-4800

An accurate prostate cancer prognosticator using a seven-gene signature plus Gleason score and taking cell type heterogeneity into account - Abstract

More Information about Beyond the Abstract