Mel,
The MSK nomograms are also known as the "Kattan nomogram", because they are based on the initial work done by Michael Kattan, PhD (statistician) and Peter Scardino, MD. Scardino's practice out of Bayor University provided the inital data, but the database now includes widespread results from other institutions...as a result, there are enormous quantities of detaied information.
The information was transformed into a sophisticated mathematical model which is able, with considerable accuracy, to predict "lilelihood" of outcomes. Statistical probabilities, as you certainly know, deal with large populations, not specific individiual results.
At the bottom of the MSK nomogram page, there are two links to published papers which describe the data and the results. The first link is from the original publication of the nomogram results (published by Kattan, Scardino, Wheeler), and the second link is from the update, which includes Scardino's expanded work (he is now working at MSK), plus additional contributions from well-known doctors from elsewhere.
The Kattan, Scardino, Wheeler paper defines treatment failure by detectibale PSA this way:
"...the earliest date that the postoperative serum PSA level rose to 0.4 ng/mL or higher, confirmed by a second PSA measurement that was higher than the first by any amount..."
There are also other means of defining treatment failure in the paper, but they are not related to detectable PSA.
See the links at the bottom of the MSK page.
The other thing to keep in mind is that patient micro-analysis of ultra-sensitive PSA test results is a relatively new phenomon which is pre-dated by the bulk of this nomogram work.
hope this helps...