Browse Day: December 11, 2007

This is the old clinical staging system. Prostate cancer

This is the old clinical staging system, which is worth looking over historically because it tells something about the way prostate cancer used to be diagnosed. This is all in the handout, as is the new staging system. I just want you to understand the concepts here. In the old days a very common entity was so-called stage A cancer. These were cancers that were detected by transurethral resection. These rarely occur these days because medical therapy for BPH has increased and because of the PSA most people are being diagnosed as a result of a biopsy. B cancers are those that are palpable, C cancers are those that extend beyond the prostate, and D are metastatic tumors. So A has gone away almost completely, B is now T2, C is now T3. What now has emerged is the so-called T1C tumors which are tumors that are detected solely as a result of an elevated PSA. These represent, as I said before, 60-70% of the tumors that we now see.

Now if a patient is diagnosed with prostate cancer what does one do to stage the disease? Bone scans are rarely necessary in the individual with a PSA of less than 10. Six studies have now demonstrated that less than 1% of patients will have a true positive bone scan. The only exceptions, I would say, are those individuals with PSA’s less than 10 that have T3 tumors or high grade tumors where the likelihood of having a true positive bone scan is higher. So routine bone scans are not necessary in routine prostate cancer. CT scans are probably even less useful and we only reserve this test if an endorectal coil MRI is not done in men who have high PSA’s, high grade tumors or T3 tumors. Endorectal coil MRI remains the best imaging tool for imaging the prostate but its utility is probably only in individuals with PSA’s in the intermediate range, between 10-20. These are patients who may benefit from an endorectal coil MRI as a staging tool in addition to all the other tools that we have.

Now we know something now about risk stratification based on PSA, grade and clinical stage. We can now tell patients what their likelihood of being cured with different treatments is, as a function of these three factors. I’m not going to go into this in detail but it’s in your handout. Just to illustrate to you, good risk patients; PSA is less than 10, Gleason scores of 6 or less. These patients are cured the majority of time no matter what form of treatment they undergo. Intermediate risk patients; PSA is between 10-20, Gleason is 7, cancers. About 50% of them are cured with different forms of therapy. Poor risk patients; no matter what you do, PSA is greater than 20, high grade tumors. Most of these patients will ultimately relapse with single modality therapy; either radiation or surgery. As I alluded to before, in a study that will be published soon, the number of biopsies that contain cancer helps us out a little bit more in terms of stratifying patients, in terms of curability. This is a large study of 960 men that Anthony D’Amico and his colleagues performed looking at PSA outcomes as a function of percent positive biopsies. I’m not going to go into this in detail but this particular factor, percent positive biopsies, adds further information to PSA, clinical stage and grade which further stratifies patients into different subgroups. I would hold that at the present time that we can do a pretty good job at telling patients in advance their likelihood of being cured of their prostate cancer with local therapy.

So what are the problems with PSA screening?

So what are the problems with PSA screening? Well, the biggest problem I think is over-diagnosis. There are clearly people who have been diagnosed with prostate cancer that didn’t need to know, that would never had died of their disease if left untreated. What we debate is what the frequency of that phenomenon is. Whether it’s 5% or 50%, we just don’t know. The other problem with PSA-based screening is the high false-positive rate. As I showed before, only 25% of patients who have a PSA between 4-10 actually have prostate cancer. How have we improved on that? I don’t know that we have improved that much over the past 5-10 years. We clearly now know that PSA’s mean different things in different age groups. As a man gets older the size of his gland will increase, and in the population the prevalence of benign prostatic hyperplasia will increase and therefore the PSA will increase as a function of age. So a PSA of, let’s say, 4 has a different implication for a man who is 45 than it does for a man who is 75. If one takes into account these changes, one can improve the specificity of PSA screening. PSA density, which is the PSA level as a function of the size of the gland by ultrasound is useful in that people with higher PSA densities are more likely to have prostate cancer. The problem with this concept is that it requires an ultrasound and the modeling of the prostate volume is very problematic in large multi-institution studies. Therefore there is not very much consistency in terms of the numerical values of cut-offs for when people need to have a biopsy, from those when you don’t need a biopsy.

PSA velocity is an important concept in that retrospectively, if you look at the rate change of PSA over time in people who have prostate cancer, it’s much faster than those individuals who have BPH or have a completely normal prostate. The problem is that in real time in a patient who comes in with an abnormal PSA on his first visit, it’s difficult to use PSA velocity to sort out those individuals who need biopsies from those that don’t, because there is so much fluctuation in the level of the PSA over the course of time. So PSA velocity tends not to be very useful as a way of reducing the false-positive rate. Probably the most useful test is the free PSA. PSA in the blood is either complexed to alpha 1 chymotrypsin or is free floating. The proportion of free PSA in a patient with an abnormal PSA, when it increases, the likelihood of having prostate cancer increases as well. So for a slightly elevated PSA and a particularly low free PSA, this test is a useful adjunct to routine screening.

So where do we stand on screening in the United States? It is not accepted, from the standpoint of recommendation, because one has not yet documented a reduction in mortality rate. There are groups that have recommended screening, some groups that have not recommended screening, but it is generally practiced in the United States. What is recommended if one undergoes screening, and if a patient understands all the uncertainties and ramifications of screening, is yearly PSA’s and digital rectal exams beginning at the age of 50 and ending when a patient has a life expectancy under ten years. If a person is in a high risk group, that is, they have a positive family history or they are African-American, then screening should begin at the age of 40.

Conceptually staging of prostate cancer really looks like this. This is not how we really stage prostate cancer, but this is how we should think about staging prostate cancer and this is where we ultimately want to go. There are those patients who have prostate cancer that don’t need to be treated. Either because their disease is so indolent that it is not progressing or has a very low likelihood of metastasizing, or their life expectancy is limited. There are those individuals that should be cured in that they have a more aggressive phenotype or they have a longer life expectancy and they still have organ-confined and curable prostate cancer. This is the most important subgroup. The last subgroup are those individuals that we would like to cure but cannot be cured because they have microscopic metastatic disease. We are improving our ability to separate these entities but we are still not quite there yet. I’ll go into that in a little bit of detail.

Now this is a very important study that Peter Albertson and his colleagues published a year ago in JAMA, looking at the natural history of untreated prostate cancer. In this study there were 3,000 men who were registered within the Connecticut cancer registry who were diagnosed with prostate cancer who, either because they chose not to get treated or their physicians didn’t recommend treatment, were ultimately not treated and followed without treatment for 15 years or longer. And the question being asked in this particular study is, “What is the likelihood of dying of the disease and what are the predictors of dying of the disease?” Common sense would lead you to believe that clearly age is an important factor; the older you are the less likely you are going to die of the disease. The other important factor was the grade of the original tumor. So we learned something about the predictors, we also learned something about the natural history. A very common scenario that we see in our clinic nowadays is a 70-year-old man with a Gleason 3+3 cancer. The likelihood of that individual dying of prostate cancer by 15 years if left untreated is only about 25%. Obviously a younger man has a higher likelihood of dying of the disease and a more aggressive tumor, a higher Gleason score, higher as well. But this is an interesting paper that you should all take a look at to understand what untreated prostate cancer looks like.

Now an important part of how we deal with patients is determining their likelihood of cure. This is something that we actually know a fair amount about right now but it isn’t always practiced in the clinic. These are the factors which determine the likelihood of having organ-confined prostate cancer and likelihood of cure: serum PSA, the lower the level the more likely it is organ-confined, grade, clinical stage, and the number of biopsies that contain prostate cancer. Now just as a background, the most common grading system that we use in the United States right now is the Gleason grading system, which is not a cytological grading system but a histological grading system in that it only takes into account the glandular patterns. In this scheme the arrangements are scaled on a scale of 1-5 where the primary pathology is first, secondary pathology is second, and the Gleason score or sum is the addition of these two subsets. The problem with this – although it’s highly predictive of the likelihood of having metastatic disease and overall behavior – is that most of the tumors that we now see in the clinic are Gleason 6 and 7. Probably 70-80% of the tumors fit into those categories. So there is a narrow range and there are very few tumors now that are a 2 or 3 or 4 and only 10% that are a 9 or 10.

Something about the epidemiology of prostate cancer

Now if we know something about the epidemiology of prostate cancer, maybe we can get some hypotheses regarding the prevention of this disease. Suffice it to say that there have been no prospective randomized studies with intervention that have confirmed that any of these hypotheses work. Although there are some suggestions that dietary fat is important, selenium ingestion, vitamin E, soy and lycopene are important. Thus far one cannot make a strong recommendation to do these things to prevent the disease.

Now let’s move on to some more clinically relevant subjects, particularly screening for prostate cancer. So let me summarize 10-15 years of work that’s been done in screening in one slide. I think this summarizes much of what we know right now. In the old days the screening test that we had for prostate cancer was the digital rectal exam. A lot of inter-observer variability. When one detects an abnormality by digital rectal exam, more often than not is the tumor outside of the prostate. When you combine the digital rectal exam with the prostate specific antigen you can detect three times as many cancers as one did by the digital rectal exam alone. From the Physicians Health Study we learned that when one uses PSA one can detect clinically undetectable cancer on average 5-6 years earlier than one would detect by digital rectal exam or the development of symptomatology. In the big screening studies the majority of cases are diagnosed within the first couple of years. With subsequent screening that the incidence rate falls off significantly. One of the big concerns with PSA screening was that what was being detected were so-called autopsy or latent cancers, these low grade, microscopic foci of prostate cancer that one dies with rather than of. Now it turns out that by pathological criteria, greater than 90% of the PSA-detected cancers are not autopsy cancers. They are not low grade, low volume cancers but are cancers of significant volume and higher grade. Now whether these cancers are truly clinically significant in the sense that these cancers, if left untreated would ultimately cause mortality, remains uncertain. In these screening studies about 5% of men over the age of 50 who come in for a digital rectal exam and PSA will be found to have an abnormality in one or both of those tests and ultimately will be found to have prostate cancer. Although PSA is not a perfect screening test, there is a relationship with the likelihood of having prostate cancer. The level of the PSA does predict the likelihood. Where 25% of patients or people who have a PSA between 4-10 have prostate cancer, you can see the problem here is that 75% of these people do not. The vast majority of people with PSA’s over 10 will turn out to have prostate cancer.

Now the big question with screening for prostate cancer is whether the strategy of PSA-based screening will reduce the mortality from the disease. There is one screening study that was presented at ASCO two years ago that was so methodologically flawed that we don’t use it as evidence for screening, but it did show some benefit for screening. Some people have looked at the slight reduction in mortality of this disease in the United States over the past few years as evidence that the screening strategy in the United States has reduced mortality, but it may be a little too early to declare victory at this point.

What we have noticed clinically in the United States – this is once again from the SERE cancer registry in the United States – these are the rates of stages C and D prostate cancer at presentation in Caucasian and African-Americans. What we do see over the past five to ten years not surprisingly, stage migration where there are fewer and fewer patients who come in the door with stages C and D. The disease has shifted to earlier disease.

What’s the cause of prostate cancer?

What’s the cause of prostate cancer? Well, we really don’t know. We do have some associations, some important risk factors. Obviously male gender is the most important one. Aging is important as well. As you can see, this is just one population. This is in the inner-city Detroit population in African-Americans and in Caucasians. The increasing incidence of prostate cancer as a function of age. Nowadays about 10-15% of patients will be diagnosed under the age of 50. There has been a tendency for younger patients to be diagnosed, obviously because of the use of PSA. Probably more than any other malignancy there is a relationship, a strong relationship, with age. Race and ethnic background is important as well. This is a slide that demonstrates the incidence patterns of prostate cancer around the world. The highest rates of prostate cancer are seen in African-Americans, particularly in inner-city African-Americans, where the rates may be as high as 100-fold greater than that seen in Japan or China. Part of this is genetic but part of it is probably environmental as well because migration studies from areas of low incidence to areas of high incidence have given rise to incidence rates that are intermediate. Either in that generation or subsequent generations, suggesting that environmental exposures are important as well. Leading the pack there are dietary factors, most importantly dietary fat. Not proven but we suspect very strongly that dietary fat, and in particular animal fat, is important in the development or progression of prostate cancer. Endogenous hormones and growth factors are important as well, most importantly testosterone – male hormone. You need testosterone around for the prostate to develop. It’s permissive for the development of prostate cancer. What’s been debated for many years is whether different levels of testosterone in the blood give rise to different likelihood’s of developing prostate cancer subsequently. The answer to that question is yes. There is probably a relationship between levels of testosterone and the likelihood of developing prostate cancer; as described in the Physician’s Health Study where there was a relationship between free testosterone levels and the subsequent development five, ten, fifteen, twenty-years later of the development of prostate cancer.

Growth factors are important as well. The most important of which so far has been insulin-like growth factor 1. There is a relationship between the level of IGF1 and likelihood of developing subsequent prostate cancer. Genetic factors, much like other epithelial adult tumors, are important as well. This is best described within the context of the Health Professional Follow-up Study. A cohort study of 50,000 health professionals asked at the beginning of the study, “Do you have a close family member with prostate cancer? A father or a brother?” The people were subsequently followed and as you can see, those individuals with a positive family history had a two to three-fold higher likelihood of developing prostate cancer as they were observed.

Now why is genetics important? In this study from the Cleveland Clinic of 1,000 men who came in with prostate cancer that were treated for their prostate cancer, those individuals who gave a positive family history did poorer with treatment than did patients with no family history of prostate cancer. Perhaps there are some genes involved in familial prostate cancer that give rise to a more aggressive phenotype. That is the hypothesis that is suggested by this data. Now genetics is a complex thing. Part of the story here is familial prostate cancer, much like familial breast cancer, familial colon cancer. Familial prostate cancer probably represents 5-10% of prostate cancers overall and this is what it looks like. This is a man who came to me at the age of 68 with prostate cancer. We took a family history; he had three brothers with prostate cancer, three cousins with prostate cancer. This is likely the result of inheritance of a rare allele in the population that gives rise to a high likelihood of developing the disease. There are a number of loci thus far that have been associated with familial prostate cancer, one of which is on chromosome 1Q. We don’t yet know what the gene is that’s involved in this particular subset of familial prostate cancer. Now there are other genetic factors as well, and one of the things that we’ve studied has been variations in the androgen receptor gene, specifically a repeat sequence within the androgen receptor gene which gives rise to variable likelihood’s of developing prostate cancer. In this particular case, individuals who have fewer glutamine repeats within the androgen receptor have a higher likelihood of developing prostate cancer than those that have many glutamine repeats within the androgen receptor. This is one of multiple genetic factors that are involved in the complex genetic framework of this heterogeneous disease.

Prostate Cancer

This year there will be almost 180,000 cases of prostate cancer diagnosed, and 37,000 deaths from the disease. There has been a very interesting pattern of incidence of prostate cancer in the United States over the past 30 years, which is illustrated by this slide, in the African-American population which is the top curve and in the Caucasian population in the United States. As you can see, over the past 20-30 years there was a steady but slow rise in the incidence of prostate cancer in these two populations, beginning in the later 1980’s, 1990’s, a rapid increase in incidence rates. This we attribute to several factors; increasing awareness, greater ease of doing biopsies, the development of the biopsy gun, but most importantly the introduction and widespread use of the prostate-specific antigen, which had detected many clinically undetected cancers in the population. What’s most interesting about these curves is the decline in incidence over the past few years, which we think is attributable to the fact that what we have done is cull out of the population many of the prevalence cases of prostate cancer that were not diagnosed because they were clinically silent, and now we are decreasing to another true incidence rate of prostate cancer as would be detected largely by elevations of PSA. So the majority of the patients that we are now seeing in the clinic are patients who are detected by PSA alone.

Now this is how we think of prostate cancer. Prostate cancer is the disease that arises in the epithelium of the prostate. The first recognizable pathological entity is known as prostatic intraepithelial neoplasia, which is the prostatic equivalent of carcinoma in situ in breast cancer; which in contrast to bladder cancer, is not a nasty entity but a disease that may coexist with invasive prostate cancer or may ultimately become prostate cancer. If you see this in a patient with an elevated PSA on biopsy, what is required is a repeat biopsy in the next few months. Because at high frequency, these patients actually have concomitant prostate cancer. The evolution of PIN to invasive prostate cancer probably takes place over years but the natural history of this disease is not really understood. Ultimately invasive prostate cancer will develop and invasive prostate cancer stretches all the way from so-called latent or autopsy prostate cancer – little microscopic foci of low-grade, low-volume prostate cancer – to bulky locally advanced prostate cancer.

Ultimately prostate cancer will metastasize and generally it consists of a heterogenous group of cells which are both androgen sensitive and androgen insensitive. So the strategy in patients with advanced disease is to remove androgen and remove the androgen sensitive population, reduce tumor bulk. Patients will almost always go into a remission but ultimately what emerges is a pure population of androgen insensitive cells and almost always this is the reason why a patient with prostate cancer will die, is the growth and spread of the hormone-refractory tumor cell population.