I think they are simply saying that the longer YOU go without
recurrence, the less likely it is that YOU will recur. (As time goes
by, obviously, more people recur cumulatively; but the longer each
individual goes without recurrence, the less likely it is that that
individual will recur)

So based on the data from your post-op biopsy, the nomogram predicts
likelihood of recurrence (by reading off the horizontal axis at the
bottom of the chart) immediately post-op.

The longer you go post-operatively with no recurrence, the more likely
it becomes that you will never recur; so you take your score from the
horizontal axis at the bottom of the chart, then trace up vertically
until you reach the number of months (vertical axis)  post-op that you
have been undetectable (defined as under 0.4); that's why the
probability line slants from left to right. So if you have had no
recurrence after 4 years, for example, your probability of remaining
recurrence-free is now greater than it was on the day after your op.

Hope that helps.

I have no idea what your number 2 para means. Leonard Evans, who is a
mathematician and a frequenter contributor here will probably be able
to explain it.

Fred

This is the nomogram that is currently used by the calculator at the
Sloan Kettering web site.  They give you the choice of using the older
"Historical Model" or the "Current Model".   The original model had
some defects and the newer model is an attempt to resolve them.

Although in principle I have the background to understand the
statistical models and techniques that are used,  I have to admit that
I've never tried to understand them in depth.  That would require an
investment in time that I don't feel it worth my while to make.   But
I think I understand enough of it to get the general flavor of what
they are saying.  So, with that caveat, let me try to address your
questions.

1.    I think that previous results have not made it clear whether
prospects improved with time.  It was definitely true for high risk
patients:  the longer they went without a recurrence the less likely
they were  to recur.  But the data I saw was not very clear about what
happened for moderate to low risk patients.  What  was clear in those
cases was that the risk stayed low and didn't increase
significantly.   What the current authors report does suggest,
however, is that even for moderate to low risk patients, the longer
they go without recurrence, the less likely the are to recur, at least
up to a point. That is encouraging..

I' m not sure which nomograms you are referring to which show an
increased risk over time.   I haven't seen any that I believe.

The results they report here are roughly consistent with those from
Walsh and his coresearchers, but perhaps a bit more optimistic.   It
is my impression that Scardino, Kattan, as well as Walsh, are leaders
in this field, so one should take seriously what they say.  The only
other predictive model I've seen people refer to is one supposedly
based on neural network methods, which was done at a VA hospital,
which gives much more pessimistic predictions.  In fact they are so
far off from what Scardino-Kattan and Walsh predict that either their
method is wrong or they are describing and entirely different
population.   When  predictions diverge so much,  it is usually the
outlier that is questionable.

Could you give a reference to a nomogram which predicts increased risk
over time?

Be that as it may, there is one point that is confusing here, and I'm
not sure what to make of it.  If all the nomogram is predicting is the
likelihood of your remaining recurrence free at 10 years from RP, then
of course, if you have gone 7 years without recurrence,  your
likelihood of still being recurrence free at 10 years is very high.
But what you are interested in is how likely you are to remain
recurrence free in after another 10 years.   It is not clear to me
just what they are saying about that.   They do say in the article
that they estimate the overall recurrence rate after 10 years, for men
who have gotten that far , as being about 3 percent.   But they don't
say how this varies with the post surgical pathology.

Let me use my case an an example.  I was last tested 89 months
following RP, and I was recurrence free.  According to the nomogram,
with my PSA and post surgial results,  the likelihood of my remaining
recurrence free at 10 years ( a bit less than three years from now) is
over 99 percent.  And according to their data, the likelihood of my
remaining recurrence free after that is at least 97 percent,  but it
might be even higher given the good post-surgical pathology report.
Their study doesn't really address that point.  I am willing to accept
such odds.

So, if I  understand the article properly, the nomogram is mostly
useful for men who are only a few years beyond RP.   It certainly
doesn't help to know the likelihood of your being recurrence free at
10 years if you have already got to 11 years, for example.

2.  Here they are discussing the validation.   The original study
applied to a specific study population.  The validation then tried to
see how well it worked on a different population.   In testing that
they asked first if it at least ranked patients correctly with respect
to greater and lower risk, and they found it did.  The secon dpart, if
I understand them correctly, refers to whether the actual predictions
were at the same level.  They discovered that their predictions were a
bit optimistic when applied to the validations population, but
apparently not greatly so.

There are of course several caveats that need to be considered.   It
is fairly clear that if Scardino is your surgeon, then your chances of
recurrence are going to be lower than if you pick a random surgeon in
your community.    The question is how much less.   The validation
gave a partial answer to this: not very much for the other surgeons
whose patients were in the validation study.  But otherwise it doesn't
really give you a good answer.

It is my impression, perhaps it is just my hope,  that while things
like avoiding impotence are highly dependent on the skill of the
surgeons, the likelihood of a cure, after the surgeon reaches a
certain level of competence, is not so dependent.   You don't really
need a "star" to be cured.   I certainly hope that I am right about
that since my surgeon, while quite good, is not a star.  Also, I think
the skill of the surgeon already shows up in the post surgical data,
such as positive margins, used in the nomogram.