One of the posters to be presented at the upcoming annual meeting of the American Society of Clinical Oncology offers an interesting model permitting replicative estimation of the numbers of patients living with different clinical stages of prostate cancer.
Solo et al. have constructed a model that allows them to estimate the prevalence (number of men living with) differing stages of localized and more advanced prostate cancer in the USA as of 2009, and then to further estimate the numbers of patients progressing from one clinical stage to the next. Such a model allows the user to evaluate how new drugs and other treatments might impact the prevalence of patients with different clinical status over time.
The model is based on a simulation starting in 1990 and using published SEER data. For the following 19 years (through 2009), additional, annual, incident cohorts of patients have then been layered into the model.
The table below shows the estimates generated by the model. (Click on the image to see a larger version.) For example, as of 2009, Solo et al. estimate that there are 2,179,270 men living in the USA after a diagnosis of localized or locally advanced prostate cancer (clinical status L1 through L3), and that these men constitute 95.5 percent of all men living after a diagnosis of prostate cancer. Similarly, they estimate that only 4.5 percent of all men living in the USA after a diagnosis of prostate cancer (N = 102,464) have metastatic forms of the disease (M1 through M4).
According to the authors, this model takes account of clinical data on currently available therapies for each clinical state and can easily be adapted to account for changing benefits from other more recently approved and potential future therapies.
Filed under: Living with Prostate Cancer, Management | Tagged: clinical status, epidemiology, prevalence, stage |
THE "NEW" PROSTATE CANCER INFOLINK 
If the model is adequately summarized in the table, I’m disheartened and rather shocked at what seems to be shoddy science.
Taking the table at face value, the text in cell B10 (“Total of all living patients diagnosed with prostate cancer”) is arrogant and wrong: All that cell C10 actually contains is the sum of the numbers in cells C2:C9, which clearly do NOT account for even one of the following six groups:
a. Men with localized PCa and stable PSA despite nondefinitive therapy. (I was in this category for half a year from 2007-2008. Did I not exist then, or was I not worth counting?)
b. Men with rising PSA who are on non-ADT treatment. (I was in this category for half a year in 2008. Did I not exist then, or was I not worth counting?)
c. Men with hormone-sensitive metastatic disease who are not on ADT. (I may find myself in this category next month. Will I then cease to exist, or no longer be worth counting?)
d. Men with mCRPC who are not receiving chemo. (I’ve corresponded with three men in this category, even though there supposedly are no such men.)
e. Men with metastatic disease who are not on treatment. (A man in my support group is in this category. I assure you, he exists and is worth counting.)
Any census that counts each of these categories as 0 has inadequate sources of data. Any model that fails to account for these categories should justify their exclusion.
Furthermore, I disbelieve the first number (“160,670”), if it purports to model how many men in 2009 would have been diagnosed without having received treatment. I suspect that the figure 160,670 represents sample bias — that the SEER data overrepresented 1990-era men who were rushed to treatment within weeks of diagnosis and underrepresented 2000-era men with Gleason 6 or 7 disease who realize they should take many months educating themselves before making a decision about how or whether to treat.
This seems to me like either a badly flawed census of the years 1975-2008 or else a wildly incorrect model for the 2010s.
How would a layman contact the paper’s authors to get a reply to such questions/objections?
PaulC in Los Angeles
Dear Paul:
You may not have noted the footnote that refers to the column in the table headed “Active disease.” It clear;y states, “Exclusive of patients who are in remission and/or not on active treatment.”
My comments concerned the text in column B (“Definition”) and numbers in column C (“Prevalence in 2009”). Did I misunderstand something?
I didn’t and don’t know what to make of the numbers in column D (“Active disease”), and so I didn’t address them.
Paul: I think you are over-analyzing the model. It’s goal is to deliver estimates of the numbers of men who are receiving treatment at a point in time, not the ones who aren’t. And it is only a model.
Greater detail is likely to be available on the actual poster at the meeting. If I am able to, I will see if I can stop by that particular poster when I am at the meeting … but I can’t promise, it can be very difficult to get to all the sessions one might like to (along with all the other things one is trying to do).