|
PSADT: Biology And PSA Dynamics
(August 2005)
“We dance around in a
ring and suppose, But the Secret sits in the middle and knows.”
- “The Secret Sits” by Robert Frost.
It’s easy enough to plug in a series of PSA values and
their respective dates into the program available at Nomograms.org and
get a straightforward answer. But accurately modeling the biological
dynamics of PSA is an elusive goal because of a variety of subtle
reasons. PSADT is very much a derived calculation based on assumptions
and compromises.
It has become a cannon of prostatology that the numerical
value of PSA is proportional to the volume of prostate cells - benign
and malignant, and that by converting a sequence of PSA values to their
natural logarithms a log-linear function emerges. This suggests an
exponential pattern of growth. This relationship was first described by
Stamey, McNeal, and Schmidt in CANCER, 1993 in which they proposed that
in untreated patients “One gram of cancer on average produces 3.5 ng/mL
of PSA” ... and that one gram of BPH tissue “elevates the PSA levels at
an average of 0.3 ng/mL”. When PSA was measured sequentially “there was
an exponential (log-linear) increase in PSA with time in 86%
[emphasis mine] of 43 patients we followed up, a linearity that allowed
us to calculate a doubling time for PSA”. This generalization has been
criticized by some as not fully reflecting the complexity of the
dynamics. The critics suggest that not all examples follow the first
order kinetics growth pattern that is the assumed basis for calculation
of PSADT and could lead to a loss of accuracy in some instances.
That PSA is proportional to prostate cancer volume is
generally accepted, but a specific PSA value does not indicate a
specific volume. In the untreated patient the contribution of PSA
arising from BPH introduces a confounding variation. This issue may
explain why PSADT in the pretreatment period has been much less
informative than in post treatment measurements. Malignant cells produce
PSA at 10 or more times their benign counterparts, and cells in the
transition zone produce less PSA than those in the peripheral zones
(presumably because of a lesser density of androgen receptors). Hence, a
70 year old cancer free man with a PSA value of 3 ng/mL will have a an
estimated gland of 50cc3, whereas a 50 year old with the same PSA has an
estimated gland size of 40 cc3 because of the greater amount of BPH
tissue in the older man. Interpretation of the PSADT is complicated by
this issue. In some models there is an effort to negate the effect of
BPH derived PSA on PSADT by subtracting the estimated contribution of
the BPH derived PSA using a formula: gland volume multiplied by 0.066
equals the amount of benign PSA from BPH.
Inoue, a U of W biostatistitian, using metaanalysis data
identified a transition point in the pattern of rising total PSA at
which the faster rising PSA produced by the malignant cells becomes
dominant and emerges from the shadow of the PSA derived from underlying
BPH. This inflection was termed the “change point” and his analysis
placed this acceleration of PSA increase at a point prior to the
identification of biochemically suspected disease - somewhere earlier
than a PSA value of 4 ng/mL. This biologic insight introduces an aspect
of non-linearity in this range of PSA that has implications for the
interpretation both of PSA screening results and calculations of
pretreatment PSADT.
Scardino, Wheeler et al. (J Urol. June 1994 - “Prostate
specific antigen and Gleason grade: an immunological study of prostate
cancer”), investigated the relation of Gleason grade to PSA. By
carefully mapping the grade(s) of cancer in 86 prostatectomy specimens
and using an immunological detection of the density of cellular PSA
staining they identified a “strong inverse correlation between Gleason
grade and the PSA content of prostate cancer”. “Serum PSA levels
correlated with total tumor volume, but the PSA levels per cm3 decreased
with increasing grade.” “While many grade 4 (found in 79% of specimens
they studied) and grade 5 (49%) cells were positive, the intensity of
staining was weak”.
It follows that each tumor, with its own heterogeneous
mix of cellular differentiation, produces a total amount of PSA that is
a composite reflecting the diversity of grades of cellular
differentiation uniquely represented in that particular tumor.
Despite these shortcomings the PSADT
is proving to be clinically useful, although our incomplete
understanding of prostate cancer biology limits the PSADT to be only the
“best guide currently available”.
«
Back to Article List |
