Why linear
extent, not percent, of cancer should be used to measure cancer in needle
biopsies
True LD, Wallner K
University of Washington, Seattle, WA
Mod Pathol. 2002; 15:184A
Objectives:
The
extent of cancer (CA) in needle biopsies is a predictor of extent of
CA in a prostatectomy specimen. Clinical decisions are based on minimal
cancer in needle biopsies, i.e., tumors <3mm are considered latent.
Many practices report extent of CA as percent, not as linear extent.
We hypothesized that linear extent provides a more accurate and reproducible
measurement than percent.
Methods:
The
linear extent of cancer in 100 biopsies (with a wide range of amount
of CA) that had been originally characterized by estimated percent involvement
by CA was determined. Correlation coefficients (CC) and standard deviations
of estimated percent (SD percent) and of measured extent (SD - ME),
based upon repeat blinded measurements, were calculated.
Results:
The correlation between percentage and measured involvement decreased
with decreasing amount of cancer. Conversely, the SD increased. The
correlation coefficients between estimated percent CA and actual CA
is < 0.8 for small cancers.
Extent
of CA 1mm 2mm 3mm 5mm All
biopsies
CC 0.23
0.70 0.64 0.88 0.88
SD percent 9 11
18 15
18
SD (ME) 0.5 0.4
0.5 0.6 1.2
Conclusions:
Due to poor correlation between estimated percentage of minimal cancer
in biopsies and measured extent of CA, and due to a high variance in
estimated percent compared with measured extent, we recommend that actual
length of cancer be reported instead of percent of cancer.
Editorial
Comment
The characterization of “insignificant” tumors in needle biopsies
is a timely topic. There is an increasing frequency of diagnosed cancers
in stage T1c. In a recent survey in our Institution we found 11.11%,
17.39%, 15.38%, 24.24%, 29.41%, and 52% of stage T1c cancers diagnosed
in the years of 1997, 1998, 1999, 2000, 2001 and 2002, respectively.
It is noteworthy the frequency of 52% in 2002. Patients with “insignificant”
cancer diagnosed in a needle biopsy may be candidates for watchful waiting.
It must be emphasized, however, that “insignificant” does
not mean “latent” cancer. So far, there is no marker to predict
clinical behavior for a particular tumor. “Insignificant”
means that there is a high probability for a tumor to have <0,5cc
in a possible radical prostatectomy, being therefore incipient. This
paper reflects the controversy among pathologists related to how we
should estimate the extension of a tumor in a needle biopsy. Epstein
JI et al. (J Urol. 1998; 160:2407-11) estimate according to the percent
of each core involved by cancer. Noguchi et al. (J Urol. 2001; 166:104-9),
from Stanford University, found that the combination of 1 positive core
with cancer length less than 3mm with no Gleason grade 4/5 involvement
is probably the best predictor of less than 0.5cc prostate cancer on
radical prostatectomy. For these authors the use of PSA, or PSA density,
in combination with needle biopsy findings did not enhance prediction
of tumor significance. Noguchi’s paper aroused an extensive editorial
comment by Epstein with an also comprehensive reply by the authors (J
Urol. 2001; 166:109-10).
Dr.
Athanase Billis
Chair, Department of Pathology
State University of Campinas, Unicamp
Campinas, São Paulo, Brazil