This is Part 4 of a 4 part series on PSA Doubling Time (PSADT):
- Part 1. Introduction & Use
- Part 2. Calculating PSADT
- Part 3. DIY Formulas with Google
- Part 4. Online Calculators
These calculators are convenient because they are online but all have defects. PSADT calculations are really only valid in the presence of exponential growth and such growth can best be assessed by noting that it corresponds to straight lines on a semilog chart. Unfortunately few of the calculators provide such diagnostic graphics. In the main, those that do provide graphics provide ordinary charts rather than semilog charts. Ordinary charts are not useful in assessing the assumption of exponential growth since its largely impossible for the human eye to discern exponential behavior on such a chart. In addition most of the these pages are poorly designed. Earlier parts of this series of posts have pointed out solutions that are not online calculators but are methodologically preferable -- see [Part 2].
- Memorial Sloan Kettering calculator described more fully [in this post] and available at the [http://www.nomograms.org] site or [directly here] includes a doubling time calculator. This calculator will not accept PSA values less than 0.1 so if you are using an ultrasensitive assay that can detect PSA levels that low then enter all your PSA values as 10 times or 100 times the actual value. The doubling time calculated will still be correct. There are no graphics provided.
- University of Montreal calculator pages described more fully [in this post] and available [here] includes a doubling time calculator. Like the Sloan Kettering calculator it will not accept PSA values less than 0.1 so if you are using an ultrasensitive assay that can detect PSA levels that low then enter all your PSA values as 10 times or 100 times the actual value. The doubling calculated will still be correct. There are no graphics provided. I have had problems using its PSADT calculator under the Google Chrome browser but was able to successfully use it with Internet Explorer and Firefox. One irritating aspect to the PSADT calculator is that after you enter the first date the meaning of the fields is obscured so its relatively easy to make a mistake in entering your data.
- Kevin's Online calculator. http://kevin.phys.unm.edu/psa/. This calculator is restricted to 2-6 PSA values. It gives the doubling time and provides confidence intervals. It also shows the PSAs on a graph but the graph does not use a semilog scale making it less useful.
- Sunnybrook. This calculator is intended to be used for determining when to exit active surveillance and is based on a model of such patients. The model may not be applicable to other applications. The model is different than the prior calculators discussed in that it models log(PSA) as quadratic, rather than linear, in time and further depends on both age and Gleason Score. It is described further at this page. The online calculator lets one enter a series of PSA values and also allows one to easily calculate the PSADT and PSAV on various subsets without re-entering the numbers. On the same graph as the data it shows typical progression for a high risk group as a line and a low risk group as a second line (both assuming the same baseline PSA as the data entered). If the data entered more closely resembles the high risk group or lies above it then active surveillance would not be advisable according to this model. Jon Nowlin has re-implemented in Excel this calculator and improved on it by using a semilog scale. (That site seems to have disappeared. 2010/10/10)
- Anton Ponholzer, Günther Zauner & Nikolas Popper Calculator. This [calculator] can only handle 3 values (in fact you must enter three values -- it will not work with two or four values). It produces a graph but the graph is not on a semilog scale which makes the graph less useful. It provides projections assuming that future PSA continues to double at the PSADT calculated. This may or may not have meaning. It requires a "variation" input which can be left at the default and ignored if you are not interested in the error bounds that it produces. Unfortunately to use the error bounds feature is not really practical since one has to know the value of "variation" and this would generally not be known. There is an article about this calculator [here].