|
Gene
Expression Alterations In Prostate Cancer (August 2002)
The article “Gene
Expression Alterations in Prostate Cancer Predicting Tumor Aggression
and Preceding Development of Malignancy” (July 15, 2004, JCO) by a group
of Pittsburgh researchers is a seminal documentation of the genetic
basis of prostate malignancy, the concept of precancerous “field
effect”, and the genetic pattern that distinguishes prostate cancer
aggressiveness. The basis of this study was the comparative analysis of
messenger RNA, the product of gene expression, among prostatectomy
specimens of prostate cancer tissue (63), histologically “normal” tissue
adjacent to cancer (60), and histologically normal prostate tissue (23)
from deceased organ donors. A 671 gene profile correctly distinguished
prostate cancer from the normal donor prostate tissue with a 99.9%
probability.
The usefulness of gene
expression profiling to predict the degree of cancer aggressiveness was
evaluated on the basis of a “70-gene” model that retrospectively
“correctly predicted 27 of the 29 (93%) aggressive tumors, and 32 of 37
(86.5%) no aggressive tumors, producing 86.5% specificity and 93%
sensitivity.” “An aggressive tumor [was] defined by any of the
following: cancer invasion into adjacent organs or seminal vesicles,
clinical relapse as evidenced by a rising PSA [ > 4 years
follow-up] following radical prostatectomy, or distant metastases.”
A comparison was made
between the accuracy of Gleason scoring in predicting aggressiveness. Of
62 cancers, 45 had Gleason scores of > 7 and 17 were < 6.
The “Gleason score has only limited accuracy in predicting true
aggressiveness with 33 of 62 (53%) correctly classified.” A second
analysis made by adding 23 additional cancer specimens yielded an
accuracy for prediction of aggressiveness for the 70 gene model of 78%
versus 52% for Gleason scoring. The specificity for the gene expression
technique was 82% v. 9% for the Gleason model.
The most thought
provocative portion of the study is the evidence that demonstrated the
very close similarity between the gene expression profile of apparently
histologically benign prostate tissue and the adjacent histologically
malignant cells. On the basis of this similarity “ninety percent (54 of
60) of AT [adjacent tissue] tissues were predicted to be tumors with a
high ( > 93% ) probability, whereas 10% (6 of 60) were predicted
to be tumors with low probability”. The “patterns of gene expression in
AT are much more similar to PC [prostate cancer] than to donor prostate,
supporting the ‘field effect’ hypothesis”. Further support for the
“field effect” derived from the observation that AT shared with PC the
same expression pattern for two well recognized PC signature genes, ie.
a downregulation of expression of glutathione-S-transferase pi (GSTpi),
and an upregulation of alpha-methylacryl-CoA racemase (AMACR)
expression.
Bottom Line:
“Collectively, these data suggest that genetic alterations in a gland
with prostate cancer are not limited to the malignant cells, and these
patterns of alteration may predict the population at risk for the
disease and for disease progression.”
«
Back to Article List
|
