Testosterone Metabolism and Prostate Cancer

[Updated June 12, 2013]

Introduction

In this post we discuss theories of testosterone metabolism as it relates to prostate cancer. The research discussed here involves, in part, theories which, while based on scientific studies, still require additional investigation in order to establish their validity in a medical context. We will primarily rely on the four pathway model found in Endotext, an online endocrinology textbook, and further modelling efforts based on the highly accessed 2007 [PMID: 17678531] [Full text] and 2005 [PMID: 15777479] [Full Text] papers by Friedman in the journal: Theoretical Biology and Medical Modelling.
[Note that since this page was written Ed Friedman has published a book entitled "The New Testosterone Treatment: How You and Your Doctor Can Fight Breast Cancer, Prostate Cancer, and Alzheimer's". This book was just published so I have not seen it nor is the information below based on it. June 2013]
We start by discussing two naive single pathway models that likely correspond to what many believe to be the case but are too limited to give sufficiently complete understanding of the biochemical dynamics. To overcome this limitation, we then expand the single pathway model to a four pathway model. This is followed by a discussion of the androgen and estrogen receptors that form key components of the four pathway model. To this we discuss a further layer involving certain apoptotic and anti-apoptotic proteins. (Apoptotic proteins cause cancer cells to be killed, which is desirable, whereas anti-apoptotic proteins protect cancer cells which is undesirable.)

Naive Model 1

In 1966 Charles Huggins won the Nobel Prize in Medicine (with Peyton Rous) for his discovery of hormonal treatments for prostate cancer. The naive model of testosterone simply says testosterone leads to cancer growth:


T -> PCa growth

One recent result that would tend to reinforce the idea that testosterone fuels prostate cancer is the finding that in response to drugs that suppress testosterone and other androgens (a common treatment for advanced cancer which tends to work for a period of time but then typically becomes ineffective after a while), metastatic prostate cancers develop genetic pathways enabling them to generate their own testosterone. See [PMID: 18519708] [full text] or [News Release]. This suggests that such testosterone is advantageous to them. In fact, Cougar Biotechnology is in Phase III clinical trials: NCT00485303 [Trial participant blog] of a drug, abiraterone, whose mechanism of action includes preventing prostate cancers from manufacturing their own androgens inhibiting androgen production in the testis, adrenals and prostate by blocking 17-alpha-hydroxylase and an enzyme required for androgen synthesis C17,20-lyase (also known as CYP17 or P450c17). See [Medical News Today] [BBC News and audio interview] [Times News and video] [PMID: 18645193] [Abstract] [ASCO presentation] and [Urotoday ASCO summary] [PCF Video], [Abiraterone Clinical Trials].

An example of using this simple model is Gat's hypothesis [prostatecancerinfolink article] [PMID: 19737278] that an age-related malfunction can allow a very much higher than normal amount of testosterone to reach the prostate from the testes. By performing microsurgery to block this transmission he was apparently was able to reverse prostate cancer in 5 out of 6 patients. (Note that the models discussed below could lead to other explanations.) A brief article on Gat's hypothesis appeared in the Feb 2010 US Too Newsletter.

Problems with Model - Lower T implies Higher Gleason Score. Although the idea that testosterone fuels prostate cancer, is appealing in its simplicity unfortunately there is evidence that seems inconsistent with it. In particular, [PMID: 18692874] found that men with lower levels of testosterone had higher grade, i.e. more aggressive, tumors. If testosterone fuels prostate cancer one would have thought that those with the lowest testosterone would have the least aggressive tumors, not the most. This suggests that there must be more to the model than just testosterone.

Problems with Model - TRT. Also, although reducing testosterone is a major approach taken in prostate cancer therapy experiences with Testosterone Replacement Therapy (TRT) in prostate cancer patients have, contrary to the above models, not consistently shown an adverse effect.

In a July 2008 review [PMID: 18638000] the authors conclude that "In the few available case series describing testosterone replacement after treatment for PCa, no case of clinical or biochemical progression was observed. ... Although further studies are necessary before definitive conclusions can be drawn, the available evidence suggests that TRT can be cautiously considered in selected hypogonadal men treated with curative intent for PCa and without evidence of active disease." A 2008 UrologyTimes article cites four studies: [PMID: 15310998], [PMID: 15643240], [PMID: 17509298], [PMID: 17183557] in which Testosterone replacement therapy was administered to prostate cancer patients beneficially. On the other hand, in the same article Dr. Lamm pointed out: In a "meta-analysis of nine studies, 11 of 987 patients given TRT developed prostate cancer versus zero of 129 control group patients [PMID: 14749457]. While not statistically significant, these figures raise the possibility that TRT fuels prostate cancer development" [reference modified to point to pubmed]

Dr. Morgensterm interprets this as "levels of androgens in the prostate do not reflect levels in the blood". He further says, referring to [PMID: 17169647] which he authored: "This is a new and exciting area that has turned on its head our idea about the relationship between testosterone and prostate cancer," Dr. Morgentaler said. "In fact, the latest data show the opposite of what we used to believe: It's not that high testosterone is a problem with prostate cancer. Actually, low testosterone appears to be a problem."

Summarizing the situation, Judy Foreman writes in the January 5, 2009 Boston Globe that: "In 2006, Morgentaler cowrote a study on 345 men with low testosterone. The study - published in the journal Urology and not industry funded - showed prostate cancer risk was higher in men with the lowest testosterone, a finding supported by a handful of other small-scale studies using human subjects. That was contrary to findings suggested by the Physicians' Health Study in 1996, a discrepancy doctors can not fully explain." On the other hand, another view expressed in the same article was: "To say that testosterone replacement therapy is safe because we have no evidence it's harmful is making an assertion on faith, not facts," said Dr. Ian Thompson, chairman of the department of urology at the University of Texas Health Science Center at San Antonio, echoing the view of other doctors who disagree with Morgentaler." Morgentaler has written a book "Testosterone for Life".
Problems with Model - Calcium. Although removal of testosterone (T) kills most cancer cells, consistent with the model, Friedman points out that this destruction of the cancer cells occurs through calcium ion influx, i.e. calcium ions entering, the cells. If calcium ions are prevented from entering the cancer cells then 70% of the cancer cells survive even in the presence of low T. Thus it seems clear that cancer cells can very well exist without T which again seems contrary to the model. See [PMID: 2235727]. (Recently it was discovered that high blood calcium levels increase the risk of fatal prostate cancer; however, this refers to calcium in the prostate cancer cells and not the blood. [PMID: 18768497] [WebMD]. Also see [Urotoday] where hypotheses related to PTH and calcium are discussed.)

Naive Model 2

A more refined, but still too naive, a model is that testosterone is converted to DHT by 5AR and the DHT that is produced then acts on the androgen receptors.


T -> (5AR) -> DHT -> AR -> PCa growth

Supporting the critical role of 5AR and DHT in the process is that individuals with a certain genetic defect in 5AR exhibit pseudohermaphroditism. They produce high levels of testosterone with low levels of DHT and have never been known to get prostate cancer. This would tend to support the role of DHT as a critical step in the model.

5AR is involved in a number of different prostate cancer and non-prostate cancer examples:

• Fat and Genistein. Even though models one and model two may be incomplete they may still be adequate for explaining certain phenomena. For example, [PMID: 18483578] uses this model, as seen in this figure to hypothesize the effects of fat and genistein on prostate cancer. "A high dietary fat intake, a risk factor of prostate cancer, induces prostate 5a-reductase-2 [i.e. 5AR] gene expression and subsequently stimulates prostate growth." On the other side 5AR and the associated growth can be inhibited by genistein, a phytoestrogen. As the two opposing forces are thought to both act on 5AR this model seems sufficient to describe these dynamics if the hypotheses of the paper are correct.
• BPH. Inhibition of 5a-reductase [i.e. 5AR] activity by medication is used in the treatment of BPH.
• Male-pattern baldness. Inhibition of 5a-reductase [i.e. 5AR] activity by medication is used in the treatment of male pattern baldness.
• Diagnostics. Regions of higher blood flow within the prostate are thought to be higher risk areas of prostate cancer. Thus if we can locate regions of higher blood flow using imaging we could focus biopsy sampling on those regions to increase the likelihood of detecting prostate cancer. Contrast agents are drugs which make it easier to image this blood flow. The imaging itself, is done via ordinary grey scale ultrasound or color doppler ultrasound or power doppler ultrasound. Color is thought to make it easier to detect the blood flow and power doppler is a variation that is thought to be less dependent on the angle of the reflected sound waves. (For comparison of ultrasound types see this table from this 2006 review by Halpern.) A promising new method to enhance this further, which has been investigated [PMID: 16183033] [full text] and is currently the subject of further clinical trials at Thomas Jefferson University, makes use of the fact that 5AR inhibitors appear to reduce blood flow within healthy prostate tissue but do not reduce it in cancerous tissue. This would mean that by administering 5AR agents such as dutasteride two weeks prior to biopsy imaged blood flow is more likely to indicate prostate cancer.
• Prostate Cancer Prevention. The use of 5AR inhibitors "in prostate cancer prevention is still controversial although it can decrease the incidence of prostate cancer." [PMID: 18483578]. In fact in the PCPT trial 18.4% of the men on finasteride vs. 24.4% of the controls developed prostate cancer -- a nearly 25% drop. (There was a greater number of advanced prostate cancer cases but that was thought to be due to the fact that as mentioned above 5AR inhibitors make it easier to detect cancer so we have to correct for the fact that a greater percentage would have been detected in the Finasteride group. See [Urosource] and [New York Times, June 15, 2008].) Finasteride is sold by Merck as Proscar.

Four Pathway Model

The previous model is still insufficient as it does not capture the fact that testosterone is known to exhibit some anti-cancer effects as well as promoting cancer.

Problems with Model - DHT correlates with survival A result that appears to be inconsistent with the above model is the observation, albeit in a small sample study, that the 15 year survival of prostate cancer patients who had higher DHT at diagnosis tended to survive longer than ones with lower DHT at diagnosis. If DHT were truly the key critical step then one would have expected the reverse. See [PMID: 18462534] [Full Text]

This suggests that a better model is required to truly understand what is going on. See [link]. We add additional pathways to the above model. See [this diagram from Endotext] for a better drawn version. Below, in addition to the pathway illustrated in the Endotext diagram we have added PCa, i.e. prostate cancer, growth or death at the end of each pathway to emphasize the typical relationship of each pathway with prostate cancer -- PCa growth is bad and PCa death is good:

1. Amplification pathway: prostate, hair, skin
   
   
T -> (5AR) -> DHT -> AR -> PCa growth
2. Direct pathway: muscle
   
   
T -> AR -> PCa death
3. Diversification pathway: brain, bone
   
   
T -> (Aro) -> E2 -> ER -> PCa growth
4. Inactivation pathway: liver
   
   
T -> excretion

Naive model 2 is just the first of four pathways in the Four Pathway model. As before, the first pathway, now labelled the amplification pathway, says that in the presence of 5AR testosterone (T) is converted to DHT which interacts with androgen receptors (AR) to encourage prostate cancer (PCa) growth.

In the second or direct pathway, testosterone (T) acts directly on the androgen receptors and has an opposite effect from the first pathway. That is in the direct pathway testosterone (T) acts against the prostate cancer (PCa). This is the opposite of what one might expect if one only looked at the first pathway.

The third or diversification pathway converts testosterone (T) to Estradiol (E2) via Aromatase (Aro). This acts on the estrogen receptors (ER) to promote cancer (PCa) growth. Friedman's model suggests that it this pathway that triggers prostate cancer.

The fourth or inactivation pathway is a route by which testosterone (T) is eventually excreted.

Hormone Receptors

Since even the four pathway model does not explain all observations seen in practice, Friedman suggests taking it to an additional level of detail where we focus on the hormone/receptor interactions. Rather than any hormone being good or bad Friedman suggests that we model the system in such a way that each hormone can exert positive and negative effects depending on which receptor is involved. In Friedman's words "testosterone, estrogen and progesterone can be either helpful, harmful, or not do much at all. It all depends on the amount of each hormone receptor within the prostate cancer cell. Each hormone has receptors that acts in contradictory manner (in effect the cells drive with one foot on the gas pedal and one foot on the brake pedal). Since in vivo you are dealing with a heterogenous population, even if 99% of the cancer cells die in the presence of any one hormone, the 1% left will thrive in that environment. Also, for different individuals, one hormone treatment might initially be extremely helpful and for another be extremely hurtful."

1. Estrogen receptors:
   
   There are two estrogen receptors in this model:
   
   • ER-alpha: accelerates prostate cancer
     
   • ER-beta: puts the brakes on prostate cancer.
     It is believed that ER-alpha and ER-beta have a relationship to TMPRSS2-ERG gene fusions. These gene fusions are associated with more aggressive cancers and future diagnostics may use their presence as a marker to distinguish between indolent and aggressive prostate cancer. Also see [PMID: 18505969]
     
     Example: Toremifene. Toremifene is in a class of drug known as a selective estrogen receptor modulator (SERM). Low doses of toremifene act again ER-alpha and to a much lesser extent against ER-beta. Since ER-alpha accelerates the cancer the effect of toremifene is anti-cancer; however, at higher doses toremifene acts against not only ER-alpha but also against ER-beta so at these higher doses the ER-beta no longer counteracts the ER-alpha and so is no longer effective. This gives it an inverse dose response curve: i.e. toremifene is effective at lower dosages where it only knocks out ER-alpha but at higher dosages it is less effective or ineffective since it starts blocking the beneficial ER-beta as well.
     Example: phytoestrogens. Phytoestrogens have an anti-cancer effect via a pathway outside the scope of this model; however, they also bind to ER-beta which could have the effect of disabling ER-beta's moderating influence on prostate cancer and encouraging the formation of bcl-2, a protein which protects cancer cells. Particularly problematic might be if the patient simultaneously increased bcl-2 from multiple sources such as by consuming high amounts of phytoestrogens such as soy and at the same time generated even more bcl-2 by taking 5AR drugs or natural 5AR inhibitors such as saw palmetto and its key ingredient beta sitosterol or with white button mushrooms. See Ed Friedman's comments and more comments. "Green tea catechin (-)-epigallocatechin gallate (EGCG) is a natural AR5 inhibitor. Flavonoids that were potent inhibitors of the type 1 5alpha-reductase include myricetin, quercitin, baicalein, and fisetin. Biochanin A, daidzein, genistein, and kaempferol were much better inhibitors of the type 2 than the type 1 isozyme. Several other natural and synthetic polyphenolic compounds were more effective inhibitors of the type 1 than the type 2 isozyme, including alizarin, anthrarobin, gossypol, nordihydroguaiaretic acid, caffeic acid phenethyl ester, and octyl and dodecyl gallates." (quotes from [PMID: 11931850])
     
     
     Example: Histone Deacetylase Inhibitors. Referred to as simply HDAC or HDI, there is some evidence that these inhibit the detrimental ER-alpha without also inhibiting the beneficial ER-beta and therefore may form a new class of anti-cancer drug in the future. See [PMID: 16158045] [full text]
     References: The use of the terms "accelerate" and "brake" as a mnemonic to remember alpha and beta and the discussion of toremifene comes from page 60 of a March 2006 presentation of Gerald L. Andriole [pdf] [flash] who in turn references Price, AUA 2005. Also see [link] and [link]. Also Sabnis et al (2007) [PMID: 17942301] have created a mouse model which has predicted the outcome of some clinical trials of breast cancer involving aromatase inhibitors and estrogen receptors in the amplification pathway. A review of SERMs focusing mostly on breast cancer is available [PMID: 17117297] [here]. Some discussion of toremifene and prostate cancer in the context of osteoporosis is available [PMID: 17062721] [Full Text]. A clinical trial on prostate cancer prevention with toremefine is discussed: [here].
2. Androgen Receptors. There are androgen receptors on the cell membrane and within the cell:
   
   • membrane androgen receptors (mAR) modulate (acts against) PCa by upregulating calcium which in turn kills prostate cancer cells. See [PMID: 15585562] [Full Text]
   • intra-cellular androgen receptors (iAR) invigorates PCa by counteracting the effects mAR. At the same time iAR also has certain anti-cancer effects. Unfortunately counteracting the mAR is the stronger of the two effects so the net effect of these two opposing forces is to promote the prostate cancer (which is bad).
   
   Example. T and DHT. DHT binds more strongly to iAR than T does to mAR.
   We can summarize this in the following:
   
   
DHT:iAR >> T:iAR

DHT:mAR = T:mAR
   
   where we use : to mean the two sides bind to each other and we use >> to mean the left side's effect outweighs the right side's effect.
   
   This means that the:
   
   • effect of DHT binding to iAR outweighs the effect T of binding to iAR and
   • DHT binds to mAR equally well as T binds to mAR

The above explains a number of phenomenon:

• DHT is pro-cancer
• T is anti-cancer but only in the absence of DHT
• if a subject has impaired iAR so that the DHT:iAR interaction is ineffective then increasing T could have an anti-cancer effect

In addition to T, AR5, DHT, mAR, iAR, aromatase, ER-alpha, ER-beta there are several additional components to the model:

Anti-apoptotic Proteins (Promoting Cancer)

The following proteins promote cancer:

• bcl-2. A small protein which promotes prostate cancer by protecting cancer cells from cell death. bcl-2 is often found in hormone resistant cancer cells. "Bcl-2 is undetectable in about 70% of patients with hormone responsive cancers. In contrast, hormone resistant tumors showed high levels of the protein."
  PCRInsights 6(1) One study found that 65% of patients with androgen independent prostate cancer exhibited high levels of bcl-2. [PMID: 8996359]. "Like the animal model, the amount of bcl-2 found in the remaining cancer increased during the course of hormonal therapy." PCRInsights 6(1).
  A May 2009 study [PMID: 19414838] found "a significant statistical association between patients' eventual death from prostate cancer and abnormal expression (using protein staining) of "bcl-2", which regulates cell death, or of the "p53" tumor suppressor gene. Similarly, high microvessel density (the number of small blood vessels in the tumor) from biopsy specimens taken at diagnosis was also associated with an increased risk of death over 11 to 16 years." [summary]. ER-alpha upregulates bcl-2 (bad) while ER-beta downregulates bcl-2 (good). Progesterone recpetor A (PRA) upregulates bcl-2 (bad) while prosterone receptor B (PRB) downregulates bcl-2 (good). (P) administered with RU-486 downregulate bcl-2 (good). mAR upregulates bcl-2 while iAR downregulates it. Note that there are many anti-apoptotic proteins similar to bcl-2 and bcl-2 may simply be a prototype that refers to them all. In fact, in [link] bax and bcl-xl were more closely related to Gleason score than bcl-2 and in [PMID: 18331646] investigators found that zinc increased bcl-2 which would appear to be bad but it increased pro-apoptotic bax even more, which is good, and the ratio of bax/bcl-2 (higher is better) may be more important than either constituent alone. Several drugs may inhibit bcl-2 formation: (1) In a 1997 paper entitled Bc12 Is the Guardian of Microtubule Integrity investigators hypothesize that the action of the drug taxol is related to reducing bcl-2 and associate it with an increase in bax. (2) Another agent that appears to have the potential to inhibit bcl-2 which, if effective might render the cancer defenseless against attack by further anti-cancer agents, is DCA. See our earlier post on DCA. (3) A third bcl-2 antagonist that is currently under investigation is WL-276 [PMID: 18519699] [full text].
• Calreticulin appears to be a primary androgen-response gene protecting the cancer cells from being destroyed by calcium influx. One reason that T suppression can destroy cancer cells is that the lack of T and DHT increases calcium
  influx while simultaneously downregulating the calreticulin that would have otherwise have protected those cancer cells from the calcium.

Apoptotic Proteins (Anti-Cancer)

The following have anti-cancer effects:

• AS3 is a protein that shuts off cell proliferation (good). Calcitriol, the active form of Vitamin D upregulates AS3. iAR upregulates AS3 (good) while mAR downregulates AS3 (bad).
• apoptotic proteins act against cancer
• calcitriol is the active form of Vitamin D. It inhibits certain anti-apoptotic proteins which would otherwise protect cancer cells from cell death.
• other apoptotic protiens.

The Friedman E-D Model

Putting together all of the above, Friedman summarized the effect of hormones on the hormone receptors in this [table of the 2005 model] based on his earlier paper and then revised and extended the model in this [table of 2007 model]. In this latter table R_D refers to the rate of prostate cancer cell death and R_G refers to the rate of prostate cancer cell growth. The up and down arrows indicate that the effect is to increase or decrease the relevant rate. Of course, increasing prostate cancer cell death and decreasing prostate cancer cell growth are good while decreasing prostate cancer cell death and increasing cancer cell growth are bad.

The idea is that the relative rate of growth and death of cancer cells is the key balance:

Cancer is a heterogeneous mix of cells. Some cells will slowly die off, where their rate of growth is only slightly less than their rate of death. Some cells will slowly increase in number, where their rate of growth is only slightly greater than their rate of death. However, those cells that thrive under the conditions of androgen deprivation are the ones that typically are the cause of death ... There are various reasons why such cells are able to thrive - high levels of the anti-apoptotic protein Bcl-2 is the reason for over 65% of the known androgen independent cells. Other mutations, such as in the pro-apoptotic protein p53 are also bad news.[Friedman]

For each of the hormones (testosterone, progesterone, and estradiol) there are two receptors acting in opposition to each other. In reality, for men with normal genetic makeup, all of the hormones initially increase the rate of cell death. If their levels are high enough, e.g. a typical teenage level, then PCa can not proliferate. As men age, and the hormone levels drop, then PCa can start up. Once it starts up, Darwinian evolution occurs, and each mutation that protects it from the hormones has a selective growth advantage. By the time a man becomes terminal, it is likely that the PCa will increase its growth rate in response to each of the three main hormones. ... in theory it is possible to have PCa cells die solely by using hormones, but in practice in will never happen unless you can start increasing hormone levels at the point that there is only a few cancer cells and no mutations yet.

... the initiating step in causing PCa is high local levels of estradiol (same for breast cancer). Continual high levels of estradiol immortalize those cells by producing oncogenes such as Myb which cause the cells to divide and producing telomerase which allows the cells to divide without shortening their telomeres. However, even though this initiates PCa, the hormone levels (especially T) must be low enough in order for PCa to proliferate. [Friedman]

The BC and PC in the table refer to breast cancer and prostate cancer as the table covers both. We recommend that the reader examine the 2007 E-D model table link carefully since it together with the Endotext diagram link conveniently summarize what we have discussed so far.

In addition to the information in the 2007 E-D model table Friedman has pointed out the importance of Dambaki (2005) et al [PMID: 16293185] [Full Text] who discovered that the mAR increase with disease progression. In particular, since T interacts with mAR to increase bcl-2 (bad) while interacting with iAR to decrease bcl-2 (good), at early stages when there is less mAR the net effect of T is to decrease bcl-2 and so T is anti-cancer but as the ratio of mAR to iAR increases over time the balance tips and the net effect of T becomes pro-cancer.

Example: Effect of DHT. As an example of using the full model together with the Dambaki et al observation of increasing mAR as disease progresses we consider whether DHT has a favorable or unfavorable effect.

In advanced PCa, DHT (and T) is bad because of the increase in mAR that occurs. The DHT downregulates the apoptotic proteins upregulated by mAR, and bcl-2 increases because of the extra mAR that is present. In very early PCa, DHT still downregulates the apoptotic proteins, but now ends up in a decrease in bcl-2 because with fewer mAR the effect of DHT binding to iAR is more effective that the effect of DHT binding to mAR.

Example: Effect of T in Healthy Men. A second example of using the full model is consideration of the effect of testosterone (T) in men without prostate cancer (PCa). As this is somewhat tricky to explain we quote Friedman directly:

Basically, there are two results that have been repeated at least twice each that seem to contradict each other. First, for men with a normal range of T, the higher the free T, the greater the chance of getting PCa. Next, for men with a low range of T, the lower the T, the greater the chance of getting PCa.

Let's look at the first case - that for normal range of T. The higher level of free T means that there is more E2 if Aromatase is turned on. This higher level means that more normal prostate cells will start dividing when they shouldn't and increases the chance of a mutation that will turn this growth cancerous. Although the rate of developing PCa is higher per year the greater the level of free T, the amount of bcl-2 produced also decreases, so that the PCa that results is less aggressive. As the level of T increases, eventually you reach a level (around teenage level) in which bcl-2 is so low that the PCa dies more quickly than it divides.

Next, looking at men with low levels of T, the rate of developing PCa each year will in fact be lower the lower the amount of T present. However, the amount of bcl-2 produced will be higher, so the PCa will be more aggressive the lower the level of bcl-2 present. The researchers are not checking for the rate of PCa developing each year - they typically would look at a bunch of 50 year olds and see how many of them have PCa. Because the lower the T the more aggressive the PCa and the lower the T the earlier in life the sooner PCa has a chance to develop, the result is that lower levels of T will result in a greater chance of PCa having grown to the size that it is capable of being detected for men at the same age.

What is interesting is that all of the above is just an examination of the PCa rate that occurs naturally. For men who don't have any PCa cells in them who take T supplementation, then bringing T to teenage levels with enough Arimidex to keep E2 within the normal range should make it almost impossible to ever get PCa.

Example: . Soy and 5AR Inhibitors. Soy may have drug interactions with 5AR2 inhibitors as mentioned by Friedman in this [here] (which can also be found [here]).

Protocols to Investigate

Friedman discusses several potential protocols. Note that these protocols are of an experimental nature. Future research will be needed to determine the effectiveness and safety of these approaches.

High Testosterone Low DHT (HTLD)

Lowering DHT would allow Testosterone to exhibit its anticancer effect. Testosterone would have to be high enough and DHT low enough for this to occur. Administration of testosterone together with an AR5 inhibitor to depress DHT would be required to effect this protocol. One significant potential problem is that (1) interaction of high T with the mAR receptors and (2) lowering of DHT both increase bcl-2 so some other strategy is required to concurrently lower bcl-2. Friedman discusses the possibility of increasing progesterone in conjunction with RU-486 and also decreasing phytoestrogens to get this effect. He also discusses the possibility of increasing calcitriol, the active form of Vitamin D. (The calcitriol component is not necessarily related to bcl-2.) Friedman summarized the effect of these action in this [HTLD table].

Dambaki (2005) et al [Full Text] discovered that mAR increase with disease progression. Thus it would be expected that the undesirable effect of T combining with mAR to produce bcl-2 is magnified in later stage patients (as there are more mAR available and therefore more bcl-2 being produced) which may restrict the useful range of the HTLD protocol to early stage patients.

Dr. Leibowitz has been using HTLD and his clinical observations seem consistent with those predicted by the models discussed here. See the following discussions on his web site: [soy], [testosterone replacement therapy], [testosterone replacement therapy case reports].

Research by Eggener et al [PMID: 16372330] is also supportive of this approach. In that study, Androgen Deprivation Therapy followed by HTLD was more effective than continuous ADT. Also see [Friedman comment] and also this [Friedman comment].

One scenario under which this protocol might be harmful is if mAR have mutated so as not to moderate PCa despite the high T. Then the hoped for beneficial effect would be absent while bcl-2 levels were increased giving a net unfavorable effect.

Low Testosterone High DHT (LTHD)

One problem with the HTLD approach is that it is pro bcl-2. Thus consider the opposite of that protocol. That is consider low testosterone and high DHT. The aim of this protocol would be to lower bcl-2. That is, since T increases bcl-2 and DHT decreases it we attempt to minimize T and maximize DHT. As with the HTLD approach we would add progestone with RU-486 and calcitriol components. Friedman summarized the effect of these actions in this [LTHD table]. The LTHD protocol, if viable, would be a prevention protocol rather than a treatment protocol. Such a protocol might sequentially follow HTLD and itself be sequentially followed by just high T. All these protocols would include calcitriol as a safeguard.

All mAR No iAR (AMNI)

The existence of mAR and iAR suggest novel therapies. For example Casodex knocks out iAR but not mAR thus a combination of high T plus Casodex (or T plus Casodex plus Proscar) might be effective. See [Friedman comment].
Friedman summarized this protocol in this [AMNI table].

No mAR All iAR (NMAI)

This protocol blocks the mAR which also blocks the generation of bcl-2. Without the protection that bcl-2 affords the cancer cells, this would leave the prostate cancer exposed to destruction by nearly any agent at all while the upregulation of the iAR would generate apoptotic proteins to carry out such destruction. Other agents could also be used to destroy the cancer now that it is no longer protected by bcl-2. This is the protocol most favored by Friedman as a potential cure for prostate cancer but currently cannot be effected due to the lack of agents that selectively block mAR. NMAI is summarized in this [NMAI table].

Thanks to Ed Friedman who commented on an earlier draft of this post.

Biochemical PSA Recurrence

[Updated January 25, 2014]

Introduction

In an particularly useful review article in the February 2008 Canadian Family Physician Wilkinson, Brundage and Siemens write

An increasing PSA level after curative therapy is termed a biochemical recurrence (BCR). Approximately one-third to half of patients will experience BCR during the course of their follow-up, regardless of modality of treatment. [PMID: 10886105] [PMID: 16600730] The significance of a BCR is in itself unclear, as not all men who have experienced BCRs will go on to experience metastatic disease. [PMID: 12605977] In one study, fewer than one-third of patients with BCR after RP developed systemic recurrence.[PMID: 12605977] In those patients who progress, BCR usually predates metastatic disease progression by an average of 7 years and prostate-cancer specific mortality by 15 years.[PMID: 16921049] [Full Text] Therefore it is useful in allowing enough lead time to implement effective salvage therapeutic strategies in those patients whose recurrences are deemed to be local. (See [Table 1] which lists factors that suggest biochemical recurrence (BCR) represents local vs distant disease after primary treatment with surgery or radiotherapy).

(Regarding the 7 year estimate above, Biotech Strategy Blog reports that Alex Haese at the AUA 2011 meeting presented data that the time to mets is 4.7 years in Europe compared to 8 years in the US.)

In [PMID: 18603352] [Full Text] the authors indicate that of those with biochemical recurrence, "25% progress to distant metastases and 11% die of prostate cancer". This same reference also contains information on predicting biochemical recurrence.

A study of "127,236 men of up to 75 years of age for whom relevant information was available in the SEER database, all of whom were treated by radical prostatectomy between 1988 and 2003" is summarized in this [PC Infolink] blog post and the abstract is available at [PMID: 22114813]. The annual hazard (roughly the probability of death in the following year given one is alive at the beginning of the year) was found to be 0.4%, 0.7% and 1% for 5, 10 and 15 years post radical prostatectomy with refined estimates based on risk groups as described in these links. Note that the hazard increases over this time span in contrast to other cancers where it typically decreases.

Several calculators on the Calculators page also can be used to estimate the risk of biochemical recurrence. An excellent presentation on biochemical recurrence dated May 11, 2011 by Niall Corcoran can be found here [pdf] [presentation with audio]. Patient information regarding the treatment of advanced prostate cancer can be found on here [Uptodate on Advanced Prostate Cancer] and in the references at its end most of which are also online.

Ultrasensitive Testing

(1) PSA Less Than 0.01 After Surgery

After prostatectomy the PSA value should go down to close to zero. If it goes below 0.01 then multiple studies with 2 to 5 year follow up have shown that the chance of biochemical recurrence is quite small.

• In a 2000 study of 200 patients over 2 years, Doherty et al [PMID: 11076649] concluded that patients who achieved a PSA of less than 0.01 after surgery had only a 3% chance of subsequent recurrence vs. 76% for those who did not achieve this level. The 0.01 level was achieved at a median of 10.4 weeks after surgery. Biochemical relapse was defined as three successive PSA rises.
• In a study of 225 patients Taylor et al (2006) [PMID: 16925750] followed 225 patients for 5-65 months and concluded that, like the Doherty study, low levels of PSA were also associated with favorable outcomes although they did not quantify the result. Biochemical relapse was defined as two succesive rising after PSA reached 0.20
• In a 2006 study Sakai et al, [PMID: 16601384] followed 127 patients for 6-75 months, and got consistent results with recurrence rates of 6.3%, 25% and 91.7% for patients achieving a PSA of less than 0.01, 0.01-0.05 and greater than or equal to 0.05. Biochemical relapse was defined as PSA persistently above 0.20.

The Taylor et al study cited above indicated that both malignant disease and indolent disease exhibited PSA values in the ultrasensitive range (i.e. PSA values which are below the 0.1 detection limit of a normal PSA test) suggesting that the cutoff for recurrence after RP should be higher than that. This also calls into question the utility of ultrasensitive testing over the long term for detection of recurrence although this and the other studies just listed show that for monitoring in the weeks immediately after surgery it does have predictive value.

(2) PSA Less Than 0.04 Three Years Post Surgery

Malik et al [PMID: 21652145] [PC Infolink writeup] found that over 95% of 765 patients who had PSA less than 0.04 three years post surgery were recurrence-free seven years post surgery. (They defined recurrence as patients who "developed a PSA level ≥0.2 or underwent salvage RT for a persistently rising PSA level after 3 yr of follow-up".)

Conventional PSA Testing

Freedland et al 2003 found the following probabilities of PSA progression:

PSA After Surgery Prob of progression (i.e. sensitivity) 0.11 - 0.2 36% (at 1 year), 67% (at 3 years) 0.2 - 0.3 86% (at 1 year), 100% (at 3 years)

which suggests 0.2 as the level defining recurrence [PMID: 12597949]; however, Stephenson [PMID: 16921049] concludes that using 0.4 as a cutoff to define recurrence would better correlate with eventual metastatic disease. There is some question regarding the interpretation of very low PSA levels but it has been hypothesized that a consistent increase in PSA over time as evidenced by a stable PSA velocity or doubling time, even at a low level of PSA, might be an early warning of disease. On the other hand, according to this 1997 paper [PMID: 9338734] "low but detectable serum PSA levels less than or equal to 30 pg/mL [i.e. 0.03 ng/mL] can be produced by nonmalignant sources of PSA" which could obscure the situation.

Work done in 1999 [PMID: 10235151] found that:

With regard to timing of PSA progression after RP, this commonly occurs within the first 2 years (44.7%), and nearly all progressions occur within 5 years (76.6%) according to Pound et al. Progression as late as 9 years has been noted in only 4% of cases. If PSA progression does occur, then freedom from metastasis at 3 years has been reported in a study at 78% and at 5 and 7 years as 63% and 52%, respectively. Patients with higher Gleason score tumours [Gleason score 8-10] have a lower metastasis-free rate compared with men with low Gleason score tumours [Gleason score 5-7]: 29 vs. 62% at 7 years. [link] based on data from: [PMID: 10235151]

AUA Criteria for Biochemical Recurrence

Need for Different Defintions for Surgery and Radiation

In surgery the primary source of PSA is removed so a level near zero can be expected if no cancer cells remain. The issue is more complex with radiation and cryotherapy since, as Nielsen and Partin, 2007, write:

"The detection of recurrent disease was in fact the original clinical application of PSA in prostate cancer. This is a relatively straightforward task following radical prostatectomy, because the primary source for PSA production has been removed. Biochemical failure among surgically treated patients is generally defined as a measurable or detectable PSA level. As such, patients with biochemical "success" after surgery have been essentially cured of their disease. The definition of failure after radiation therapy is complicated, in that currently available technologies by design incompletely ablate all functioning prostatic epithelium. This limits the definition of a clinically meaningful post-treatment nadir analogous to the postsurgical undetectable PSA." [PMID: 17592538] [Full Text]

Surgery

For post surgical monitoring, in February 2007 a panel of the American Urological Association, AUA, recommended "defining biochemical recurrence as an initial serum prostate specific antigen of greater than or equal to 0.2 ng/mL, with a second confirmatory level of prostate specific antigen of greater than 0.2 ng/mL". See [PMID: 17222629]. A typical protocol might be to start following the patient more closely if PSA reaches 0.05 to build up a history of PSA values and if there is no evidence of systemic disease to proceed to radiation by the time the PSA reaches 0.3 or 0.4 . This would put the patient in the most favourable risk group (represented by the blue line in panel B of Stephenson et al 2007 [Figure 1] of [PMID: 17513807] [full text].

Radiation

At the same time, the panel also recommended that the 2005 ASTRO criteria be used for monitoring patients after radiation: (1) a rise by 2 ng/mL or more above the nadir PSA be considered the standard definition for biochemical failure after EBRT with or without HT; (2) the date of failure be determined "at call" (not backdated). [PMID: 16798415] The ASTRO definition was developed as having the best sensitivity (probability of subsequent failure given the criterion is met) and specificity (probability that subsequent failure does not occur given that criterion is not met) of the considered criteria. See [table]. The probabilities in the last two columns of the table are the fraction of men who will not meet the recurrence criteria (that is they will be free from recurrence) at 5 and 10 years.

One phenomenon to note is that frequently there is a temporary rise in PSA after radiation thought to be due to inlammation induced by the procedure itself. This rise is known as PSA Bounce and is further discussed [here].

Other

A [European Consensus] (summary) (post RP chart) (post RT chart) also agreed on these definitions. The summary and chart links just cited provide useful brief overviews of treatment and it is recommended that all readers review them. We have also added these links to the Links in the right side of this blog under "Guidelines - Europe" to keep them easily available.

For Cryotherapy, Cooperberg and Carroll of UCSF define PSA recurrence in the cryotherapy context as a rise in PSA level of 0.2 ng/mL points after a nadir (i.e. lowest point achieved) of less than 0.5 ng/mL.

Post Treatment Recurrence Calculator

There is a Memorial Sloan Kettering calculator for post treatment recurrence that implements a predictive model based on Stephensen 2007 JCO for PSA progression. This calculator is also provided in nomogram form here. The University of Montreal provides other relevant calculators. Also see the Calculators page on this site. The d'Amico risk category (discussed in the Favorable Outlook section of this post and the subject of an April 2008 validation study at the Mayo Clinic [PMID: 18289596] specifically within the context of biochemical recurrence) has also been used for assessing progression risk.

Local Recurrence vs. Systemic Disease

Local recurrence refers to rising PSA without further spread of the disease (i.e. without metastases). If the disease has spread then it is referred to as systemic. At the 2009 ASCO meeting John Hopkins researchers reported on a 25 year follow up to patients who had surgery and subsequent recurrence. They found that PSA doubling time (PSADT), Gleason score, and time to PSA progression were "strong independent predictors of metastasis-free survival". Of the patients experiencing recurrence, patients with with PSADT of less than 3 months, 9 months and 15 months had 20x, 6.3x and 2.4x the risk of systemic disease relative to those whose PSADT was longer respectively. Also patients with a Gleason score of 8 or more had double the risk of those with a lower Gleason score. Those for whom progression was evidenced within 3 years of surgery had roughly 3x the risk of those whose recurrence occurred later than 3 years. See [abstract] and [Science Daily News]. For more about PSADT see this [4 part post on PSADT]. (Note that the ratios (20x, etc.) described above were actually hazard ratios rather than relative risks. [PMID: 15273082] [Free Full Text]. To take the example of Gleason scores, the hazard ratio of 2 cited means that the odds that a patient with recurrence and Gleason score of at least 8, say, exhibits systemic disease before a patient with recurrence and a lower Gleason score is 2. Hazard ratios measure relative risk but are not necessarily numerically equal to the relative risk. See references just cited for details.)

One caution is that PSA doubling times from ultrasensitive assays can be substantially different from ordinary assays. For example, in [PMID: 22014796] the authors found that "Ultrasensitive prostate specific antigen doubling time was more or less rapid than traditional prostate specific antigen doubling time by more than 15 months in 244 (62%) and 35 (9%) patients, respectively." and they therefore conclude that: "Agreement between prostate specific antigen doubling time calculated using ultrasensitive vs traditional prostate specific antigen values is poor. Ultrasensitive prostate specific antigen doubling time is often significantly more rapid than traditional prostate specific antigen doubling time, potentially overestimating the risk of clinical recurrence. Until the significance of ultrasensitive prostate specific antigen doubling time is better characterized, the decision to proceed with salvage therapy should not be based on prostate specific antigen doubling time calculated using ultrasensitive prostate specific antigen values."

Treatment Options

The Niall Corcoran presentation we referred to previously suggests that salvage radiation would be offered where local recurrence is likely, life expectancy is long and there is no visible sign of mets. A discussion of salvage treatment options for recurrence, including two case studies, was presented at a session of the conference: Challenges in the Management of Urological Cancers (Amsterdam, The Netherlands) June 30, 2006 chaired by Thomas Keane and is summarized [here] and in a subsequent paper [PMID: 18163938].

Trock et al (2008) [PMID: 18560003] [Full Text] present a retrospective study of 635 men with a median follow up of 6 years after recurrence and 9 years after surgery. They found that:

• 22% of men with recurrence and no salvage therapy died in the follow up period vs.
• 11% of those who had salvage radiation and
• 12% of those who received salvage radiation + hormones

After case mix adjustment the group with radiation salvage had 3x the survival rate of the no salvage group. Salvage therapy had to be administered within 2 years of recurrence and PSA doubling time had to be less than 6 months for salvage radiation to be effective. The authors recommended that the results be validated in a randomized trial.

Stephenson published a nomogram for Recurrence after Salvage Radiotherapy that can be found in the section of that name on our Calculators page. The sections on the Memorial Sloan Kettering calculators and the University of Montreal calculators on the same Calculators page provide risk estimates for recurrence after surgery.

Prevention

The WCRF/AICR report produced by a team of researchers around the world details lifestyle changes to reduce the risk of cancer. The same steps are believed to help reduce the risk of recurrence for those who already had cancer.

A 2012 review [Full Text] [PMID: 22218632] of nutrition and chemoprevention provides a [table] listing the candidates, their demonstrated benefit (or lack thereof), and the level of evidence (level 1 is highest level of evidence, 2 is lower, etc.). Using this would, again, be on the assumption that chemoprevention measures are the same measures that would be helpful to those who already have prostate cancer.

Regarding aspirin, which is listed in the table just cited, further evidence in support of aspirin's anticancer effect was published in March 2012 in several papers in the Lancet. See this [Medscape summary] which also provides links to the papers. The effect in reducing the risk of metastasis was quite large for adenocarcinomas (95% of all prostate cancers are adenocarcinomas). GI bleeding is one adverse side effect although it was found that after a period of time that side effect was reduced.

A randomized trial [PMID: 23162860"] followed a median 11.2 years found an 8% reduction in cancers among men who took daily multivitamins. An article on page 3 of UStoo newsletter discusses it further. He points out that although the reduction is modest the fact that its a randomized study rather than observational makes it more reliable. He indicates that the multivitamins used are most similar to children's multivitamins and the actual brand they used currently has a different formulation so that cannot be used to ensure comparability.
Several recent studies by EL Richman and others have suggested some additional intriguing possibilities:

• Eggs may be harmful. "In conclusion, consumption of eggs may increase risk of developing a lethal-form of prostate cancer among healthy men." [PMID: 21930800]. Quoting from the same abstract: "Men who consumed 2.5 or more eggs per week had an 81% increased risk of lethal prostate cancer compared to men who consumed less than 0.5 eggs per week."


• Cruciferous vegetables may be beneficial. "In conclusion, cruciferous vegetable intake after diagnosis may reduce risk of prostate cancer progression." [PMID: 21823116]. The study was based on "1,560 men men diagnosed with non-metastatic prostate cancer taken from a famous US database (known as CaPSURE – 40 sites, mostly community-based clinics). ... Men that reported a regular (about once a day) intake of cruciferous vegetables had a significant 59% reduction in risk of cancer returning compared to men that occasionally consumed these veggies." (quoted from Dr. Moyad article in Sep 2011 US Too). Cruciferous vegetables include brocolli, brussel sprouts, cabbage, cauliflour, bok choy, radishes, daikon, kohlrabi, rutabaga, collard greens, turnip greens, arugula and cress. See Wikipedia for a longer list.


• Brisk walking may be beneficial. "Brisk walking after diagnosis may inhibit or delay prostate cancer progression among men diagnosed with clinically localized prostate cancer." [PMID: 21610110] Quoting from the same abstract: "Men who walked briskly for 3 h/wk or more had a 57% lower rate of progression than men who walked at an easy pace for less than 3 h/wk (HR = 0.43; 95% CI: 0.21-0.91; P = 0.03). Walking pace was associated with decreased risk of progression independent of duration (HR brisk vs. easy pace = 0.52; 95% CI: 0.29-0.91; P(trend) = 0.01). Few men engaged in vigorous activity, but there was a suggestive inverse association (HR ≥3 h/wk vs. none = 0.63; 95% CI: 0.32-1.23; P(trend) = 0.17). Walking duration and total nonvigorous activity were not associated with risk of progression independent of pace or vigorous activity, respectively."

Some additional studies indicated:



• Vegetables may be beneficial and high GI food harmful. In a 2011 study [PMID: 21774611] of 982 men "comparing the highest to lowest quartiles of intake, we found that increasing intakes of leafy vegetables were inversely associated with risk of aggressive prostate cancer [adjusted odds ratio (OR) = 0.66, 95% CI: 0.46, 0.96; P trend = 0.02], as was higher consumption of high carotenoid vegetables (OR = 0.71, 95% CI: 0.48, 1.04; P trend = 0.04). Conversely, increased consumption of high glycemic index foods were positively associated with risk of aggressive disease (OR = 1.64, 95% CI: 1.05, 2.57; P trend = 0.02). These results were driven by a number of specific foods within the food groups. Our findings support the hypothesis that diets high in vegetables and low in high glycemic index foods decrease risk of aggressive prostate cancer."
• Low fat diet with fish oil supplements may be beneficial. In a 2011 study of 55 men with prostate cancer were given a low fat diet with fish oil supplements to achieve a omega6:omega3 ratio of 2:1 (vs. 15:1 for Western diet). There was no change in IGF-1 status but there was a reduction in Ki-67 proliferative index relative to controls on a Western diet. See [abstract]. Note that this is not particularly strong evidence. For example, this study found improved proliferative index from antioxidants yet its conclusions were later reversed in larger studies.


• Coffee. Based on a prospective analysis of 47,911 men in the Health Professionals Follow-up Study who reported intake of regular and decaffeinated coffee in 1986 and every 4 years thereafter researchers found that those who consumed coffee had a lower risk of prostate cancer and a much lower risk of lethal prostate cancer. This study was focused on first time cancer rather than recurrence although its commonly thought that the same factors affect both. Note that this is only an observational study and so is less persuasive than a randomized study with controls; nevertheless, there are a number of supporting aspects to the portion regarding the risk of prostate cancer: There was a dose-response effect for risk of prostate cancer, i.e. the more coffee that was consumed the lower the risk. The risk reduction for 3 cups or less, 4-5 cups and 6+ cups per day was 6%, 7% and 18% (fully adjusted for other risk factors). Also reductions in risk have been found for many other cancers strengthenng the conclusion, e.g. see this metanalysis of several cancers and coffee: [PMID: 21406107]. For lethal prostate cancer the risk reduction was 19%, 14% and 60% reduction for the same categories. Although the risk reduction in fatal prostate cancer was large for the heaviest coffee drinkers, there were only 12 subjects in that category (i.e. small number of observations) and strictly increasing response with dosage was not observed. See [PMID: 21586702] [Full Text] [table] [NY Times] [Environmental News Network]. The last link reviews the pros and cons of various levels of coffee consumption including not only prostate cancer but other diseases.
• CAPE from Honey Bee Propolis. Caffeic acid phenethyl ester, or CAPE, derived from propolis (used by honey bees to construct their hive) arrests the growth of prostate cancer cells in a mouse model. Note that there are many treatments that seem to work in mouse models that fail in humans so this only provides only weak evidence. [Sciencedaily] [Cancer Prev Res 2012. 5(5), 788–97] [PMID: 22347457] [Full text]


• Eggs and poultry WITH skin may be harmful." Our results suggest that the postdiagnostic consumption of processed or unprocessed red meat, fish, or skinless poultry is not associated with prostate cancer recurrence or progression, whereas consumption of eggs and poultry with skin may increase the risk." [PMID: 20042525] Quoting from the same abstract: "Intakes of processed and unprocessed red meat, fish, total poultry, and skinless poultry were not associated with prostate cancer recurrence or progression. Greater consumption of eggs and poultry with skin was associated with 2-fold increases in risk in a comparison of extreme quantiles: eggs [hazard ratio (HR): 2.02; 95% CI: 1.10, 3.72; P for trend = 0.05] and poultry with skin (HR: 2.26; 95% CI: 1.36, 3.76; P for trend = 0.003). An interaction was observed between prognostic risk at diagnosis and poultry. Men with high prognostic risk and a high poultry intake had a 4-fold increased risk of recurrence or progression compared with men with low/intermediate prognostic risk and a low poultry intake (P for interaction = 0.003)."


• Well done, grilled or barbequed red meat may be harmful. A 2011 study of 470 cases and 512 controls found that "Higher consumption of any ground beef or processed meats were positively associated with aggressive prostate cancer, with ground beef showing the strongest association (OR = 2.30, 95% CI:1.39-3.81; P-trend = 0.002). This association primarily reflected intake of grilled or barbequed meat, with more well-done meat conferring a higher risk of aggressive prostate cancer. Comparing high and low consumptions of well/very well cooked ground beef to no consumption gave OR's of 2.04 (95% CI:1.41-2.96) and 1.51 (95% CI:1.06-2.14), respectively. In contrast, consumption of rare/medium cooked ground beef was not associated with aggressive prostate cancer." See [PMID: 22132129]


• Anti-Angiogenesis. As solid tumors require supporting blood vessels to grow Judah Folkman [papers] proposed that one way to combat cancer might be to prevent the growth of blood vessels. About a dozen drugs having this effect are already on the market for other cancers (breast, lung, colon, brain, and kidney) but none for prostate cancer. See [PMID: 20678204] for a review. In a [TED video] Dr. William Li speculates on the possibility that if applied early enough that simply eating foods known to inhibit angiogenesis might prevent cancer or recurrence. See the list of foods [here].


• High Fiber Diet. Research on mice has found that a high fiber diet (believed to be due to the active ingredient compound inositol hexaphosphate, also known as IP6) can slow progression of prostate cancer in mice. See [US Too 3/2013 pg. 5] and [PMID: 23213071]. It is unknown whether this applies to humans. The American Cancer Society summarizes what is known [here]. A set of slides on high fiber foods is available on [WedMB].
• Mediterranean Diet. This study indicates that carnosol may be the active anti-cancer ingredient of the Mediterranean diet. It is found in the spices rosemary and sage. The paper indicates that there is test tube and animal evidence that carnosol has activity against prostate, breast, skin, leukemia and colon cancers. Given that the Mediterranean diet is known to have health benefits in humans this suggests that the test tube and animal studies may apply.
• Carbohydrate Restriction. [PMID: 23038057] showed that restricting carbohydrates to 20% of calories slowed prostate cancer in mice.
• Replacing animal fats and carbohydrates with vegetable fats. In an observational study of over 4000 men, "men who substituted 10 percent of their daily calories from animal fats and carbs with such healthy fats as olive oil, canola oil, nuts, seeds and avocados were 29 percent less likely to die from spreading prostate cancer and 26 percent less likely to die from any other disease when compared to men who did not make this healthy swap" [WebMD Summary] [PMID: 23752662] [video interview with author]
• Pomi-t. A double blinded randomized controlled study of about 200 men found that those who took two Pomi-t capsules per day where each capsule consisted of 100mg Broccoli Powder, 100mg Turmeric Powder, 100mg Pomegranate seed powder and Green Tea 5:1 extract 20mg equivalent to 100mg plus Gelatine in the capsule itself experienced significantly lower rises in PSA vs. the control group (PSA rise of 14% vs. 78% for the control) over a 6 month period. The study had no commercial funding. [ASCO Abstract]. [CancerNet UK Trial Descxription], [Pomi-t web site], [Abstract on Pomi-T web site] Powerhealth (manufacturer) (Note that this is from a presentation at ASCO and not from a journal article so it would not been subject to the scrutiny of peer review.
• Ornish Lifestyle Study. A pilot study published in the Lancet of prostate cancer patients conducted in the lab of Nobel laureate Elizabeth H. Blackburn, by Dean Ornish and others at UCSF found that life style changes were associated with increase telomere length (a measure of aging - longer is more favorable). Article and Ornish interview video. A bit more detail on the life style changes (from the Toronto Star) is here. NPR article cautions that telomeres are not the whole story and even the favorability of longer telomeres is not certain: NPR article The abstract to the study is here: pubmed: 24051140
• A 2013 John Hopkins study found that among 29 men with of metastatic castration-resistant prostate cancer taking 600 mg/day of the drug itraconazole (which is currently approved in the US as an antifungal and is also known as Onmel® or Sporanox®) resulted in no PSA progression in 14 of 29 high dose subjects at 24 weeks. The high dosage seems essential as among 17 men in a low dose arm (200 mg/day) it was found to be ineffective and that part of the study terminated early. This and previous test tube and animal studies (see references 11-16 in paper) found that the mechanism of action is to inhibit angiogenesis and Hedgehog signaling. The researchers concluded that the high dose regimen has "has modest antitumor activity". Side effects included fatigue, nausea, anorexia, rash, and a syndrome of hypokalemia, hypertension, and edema. Additional side effects are listed here. There are a number of drugs which should not be taken with Itraconazole. See Wikipedia. The abstract and full paper are available here: [PMID: 23340005] [Full paper].

Except for large randomized trials the nutrition studies provide only provisional evidence and can be overturned once such large randomized trials are performed. If the effect that they find is large it makes the nutrition study more likely to hold as competing reasons must then also be large to overturn them; however, such alternative explanations (such as those who consume X tend to also have good or bad lifestyle habits in general) are always a possibility. In fact the SELECT trial has already overturned a number of the provisional conclusions in the prostate cancer section of the WCRF/AICR report.

PSA Screening and Early Detection - Part 5. More Diagnostic Testing Concepts

PSA Screening and Early Detection. Part 1 - Guides
PSA Screening and Early Detection. Part 2 - Key Points on PSA
PSA Screening and Early Detection. Part 3. Current Environment
PSA Screening and Early Detection - Part 4. Diagnostic Testing Concepts [previous]
PSA Screening and Early Detection - Part 5. More Diagnostic Testing Concepts [current]

[Updated May 17, 2010]

In the following we give a number of actual examples of press releases and research results about PSA and related testing and show how to analyze these sources in terms of the diagnostic testing concepts that were discussed in the last part of this series. The examples illustrate how to gain insight into these articles and even uncover errors. Although in keeping with the focus of this blog, the examples will all involve prostate cancer the material can also be useful for understanding diagnostic testing, in general, for other diseases, as well.

First we review the material from the last part covering it from additionl perspectives. Some of the discussion could, in principle, involve algebra (but nothing beyond elementary high school level); however, we will avoid even that by making use of the free symbolic algebra program, Mathomatic, that can be used to perform all the calculations with hardly any knowledge of mathematics. It is free and available on Windows, Mac and Linux. It can also be accessed online, i.e. over the internet without installing it on your computer, by using the following line run from the Windows or the UNIX command line:


telnet mathomatic.orgserve.de 63011


Prevalence

The prevalence of prostate cancer is the proportion of men in a population that have the disease at a point in time.

The prevalence is also your probability of having prostate cancer before you know your test results. Because of this, the prevalence is sometimes referred to as the prior probability.

Based on the data derived from a meta study in Vollmer, 2006[PMID: 16613336] the prevalence of prostate cancer among men subject to test is 0.9595% among men aged 55, 5.015% among men aged 65 and 11.947% among men aged 75. We shall use 5% as an overall prevelance when we wish to use a single figure regardless of age.

Sensitivity/Specificity and PPV/NPV

We can divide the subjects under test into two groups in two different ways. We can either divide them into the diseased vs. healthy populations or we can divide them into the groups having a positive test (i.e. indicative of cancer) and negative test (i.e. indicative of no cancer). We consider each of these divisions in turn:

1. Diseased vs. Healthy

The first way to divide the population into two is to divide subjects into the diseased and healthy populations. The fraction of correct tests in each of these populations is called the sensitivity and specificity, respectively. For example, in the PCPT trial discussed in the last part of this series out of every 5 diseased subjects 2 had positive tests (assuming a cutoff of 2.5) so the sensitivity is 2/5 = 40%. Also from the PCPT out of 95 healthy men we had 77 negative tests which is a specificity of 77/95 = 81%.

Tests with high sensitivity are good at detecting that disease is present.

Tests with high specificity are good at avoiding needless treatment on healthy individuals. In the case of prostate cancer if one gets a positive PSA test one will proceed to have a biopsy so the higher the specificity of the PSA test the less likely a healthy individual will needlessly get a biopsy.

The sensitivity and specificity depend on the cutoff but not on the prevalence. They are measures of the test but not of the population under test (all other factors being the same). Thus if we wish to decide which of several alternative tests to undergo we would want one with the higher sensitivity and higher specificity.

2. Those Testing Positive vs. Those Testing Negative

The second way of dividing the population into two is to divide it into those who test positive and those who test negative. (A positive test is indicative of disease and negative test is indicative of being healhty. This is different than actually having or not having cancer because tests are not perfect. Having a positive test result does not necessarily mean you have cancer and having a negative test does not necessasarily mean you do not have cancer.) The fraction of correct tests in each of these populations is called the positive predictive value (PPV) and negative predictive value (NPV). That is the PPV is the number of men with cancer and positive tests as a fraction of those with positive tests. The NPV is the number of men who do not have cancer and have negative tests as a fraction of all men who have negative tests.

In the PCPT trial discussed in the last part of this series out of every 20 positive tests 2 actually had cancer which yields a PPV of 2/20 = 10%. Out of every 80 negative tests 77 did not have cancer which gives an NPV of 77/80 = 96%.

Like sensitivity and specificity, the PPV and NPV depend on the cutoff used for the test but unlike sensitivity and specificity they also depend on the prevalence of disease in the population. Since PPV and NPV depend on the prevalence they are not pure measures of the test. Their values will change by applying the test to a different population. In particular, if the prevalence is higher the PPV will be higher and the NPV will be lower.

The PPV and NPV answer the questions: if I have a positive test what is my chance of having cancer (PPV) and if I have a negative test what is my chance of being free of cancer (NPV)? Thus the PPV and 1-NPV are the prevalences of cancer in the positive and negative test groups respectively. As a result they are sometimes referred to as the posterior probabilities in contrast to the prevalence in the entire population which is called the prior probability. Thus before you are tested your probability of cancer equals the prevalence and after you are tested your probability of cancer is PPV or 1-NPV depending on whether your test was positive or negative, respectively.

Stated in terms of prevalence, the prevalence is the fraction of cancer in the entire population while the PPV is the prevalence of cancer in the subgroup that tested positive (and 1-NPV is the prevalence of cancer in the subgroup that tested negative).

Since anyone with a positive test is sent for a biopsy PPV is also the proportion of biopsies that were really needed while 1-PPV is the proportion of biopsies that were done needlessly (since the subjects tested positive but did not have cancer).

Since PPV rises as prevalence rises and since men with a family history of prostate cancer have a higher prevalence they would have a higher PPV. In fact [PMID: 16515992] found that the PPV in a group of men with prostate cancer family history was 32.2% vs. 23.6% in a group with no family history.

Comparison

We use prevalence to decide whether to get tested, sensitivity and specificity to decide which test to undergo and we use PPV and NPV to assess our likelihood of being diseased after we get back our test results. We deal with each of these cases:

1. Prevalence is used to decide whether to get tested in the first place. If the prevalence is very low, i.e. the disease is rare, then you probably don't need to get tested. Prostate cancer is sufficiently prevalent past a certain age that you probably want to get tested (at least according to nearly every patient and physician group but not according to governnment groups as discussed in Part 3 of this series). If you are in a high risk group (family member with prostate cancer, African American, carrier of BRCA1 and BRAC2 genes) then it would be even more important as the prevalence among your group is even higher. The idea of screening high risk groups is called targeted screening and is discussed in this Nov 2006 paper by Mitra et al.


2. Once you have decided to get tested the sensitivity and specificity can be used to allow you to determine which test to take. You want the test with the highest sensitivity and specificity. You want high sensitivity so that it detects disease if its there and you want high specificity so you are not subject to a biopsy if you don't have disease. Of course, since PSA testing is the only widely available test for screening you do not really have a choice here but in the future there may be new tests and in that case if one of the future tests has higher sensitivity and higher specificity than the PSA test then you would want that one.


3. Once you are tested, if you test positive the PPV is your probability of having cancer and if you test negative the NPV is your probability of not having cancer.

Data on PSA Testing

The following table summarizes the 2x2 table for each of the three data sources in the last part of this series together with the sensitivity, specificity, PPV, NPV and prevalence derived from each 2x2 table. The cutoff for the PCPT data was 2.5 and is unknown for the other two data sets.

Data Source		Sensitivity	Specificity	PPV	NPV	Prevalence
PCPT	PC no PC PSA +ve 2 18 20 PSA -ve 3 77 80 5 95 100	2/5 = 40.0	77/95 = 81.1	2/20 = 10.0	77/80 = 96.3	5/100 = 5
Ontario	PC no PC PSA +ve 3 5 8 PSA -ve 2 90 92 5 95 100	60.0	94.7	37.5	97.8	5
AAFP	PC no PC PSA +ve 3 7 10 PSA -ve 1 89 90 4 96 100	75.0	93.0	30.0	98.9	4

A 2009 meta analysis of PSA testing [PMID: 1974436] summarized various studies on PSA accuracy with the following table:

	Study	Year	No	TP	FP	FN	TN	Sensitivity	Specificity
1	Aragona	2005	3171	1073	1695	98	305	0.92	0.15
2	Beneduce	2007	101	42	31	8	20	0.84	0.39
3	Ciatto	2004	410	167	171	18	54	0.90	0.24
4	Espana	1998	170	53	96	15	6	0.78	0.06
5	Fischer	2005	178	61	76	13	28	0.82	0.27
6	Hofer	2000	184	67	81	7	33	0.91	0.29
7	McArdle	2004	171	93	52	10	16	0.90	0.24
8	Ryden	2007	361	180	146	8	27	0.96	0.16
9	Unal	2000	59	30	10	0	19	1.00	0.66
10	Wymenga	2000	716	253	228	68	15	0.79	0.06

In terms of our previous 2x2 tables, for each study the TP (true positive) and FP (false positive) entries form the first row of the 2x2 table for that study and the FN (false negative) and TN (true negative) form the bottom row. In this notation Sensitivity = TP / (TP + FN) and Specificity = TN /(TN + FP) so that for example in the first row we have Sensitivity = TP / (TP + FN) = 1073 / (1073 + 98) = 0.92 and Specificity = TN /(TN + FP) = 305 / ( 305 + 1695 ) = 0.15 .


Formulas

Because the sensitivity and specificity are pure measures of the test whereas PPV and NPV partly measure the test and partly measure the population, the sensitivity and specificity are the two numbers that are typically shown for a diagnostic test. Once we are tested we will be interested in the PPV if we test positive and NPV if we test negative. To get PPV and NPV given the sensitivity, specificity and prevalence use these formulas (where p = prevalence, Sn = Sensitivity, Sp = Specificity, PPV = positive predictive value, NPV = negative predictive value).


PPV = p * Sn / ((p * Sn) + ((1-p) * (1-Sp)))
NPV = (1-p) * Sp / ((1-p) * Sp + p * (1-Sn))


We will mainly use the first of these two formulas. That formula relates PPV to Sn, Sp and p. By solving the formula for p, Sn or Sp we can get a new formula which gives p, Sn or Sp in terms of the other three variables. When we need to do that we will use the free Mathomatic software to calculate the solution in order to avoid explicit algebraic manipulation. Similar comments apply to the NPV formula.

Examples

Example 1. PPV of the ECPA-2 Prostate Cancer Test

According to this June 26, 2007 ABC News article the ECPA-2 "test has been shown to correctly identify which male patients did not have cancer 97 percent of the time, and which men did have prostate cancer 94 percent of the time.

By comparison, figures from the National Cancer Institute and others have shown the PSA test has specificity rate as low as 15 to 30 percent - meaning that on average, for every four or six men who test positive, only one actually has prostate cancer.

'For men with elevated PSA levels, only one in six who get biopsies today actually have prostate cancer,' he explained. 'This amounts to 1.3 million to 1.6 million men being biopsied to find the 230,000 or so who actually have prostate cancer.'"

The first paragraph says that the sensitivity of the ECPA-2 test is 94% and the specificity is 97%.

The second paragraph illustrates a confusion on the writer's part. He refers to the term specificity but then defines it as the percentage of positive tests for which the subject has cancer. That is the definition of PPV, not specificity.

Its already clear that the ECPA-2 test is superior to the ordinary PSA test (if the numbers in the article are correct) since ECPA-2 has both sensitivity and specificity in the 90's; however, since the article intends to compare the PPV of the ECPA-2 with the PSA test, let us make such a comparison ourselves.

Answer:

We have:


p = 5% (assumed)
Sn = 94%
Sp = 97%

PPV = p * Sn / (p * Sn + (1-p) * (1-Sp))
= 0.05 * 0.94 / (0.05 * 0.94 + (1-0.05) * (1-0.97))
= 62.25%


Note that this PPV of 62.25% for the ECPA-2 is the one that should be compared to the PSA PPV. The PSA PPV of the PCPT trial, the Ontario data and the AAFP data were 10%, 37.5% and 30%. Thus, while EPCA-2 continues to show a significant advantage for the ECPA-2 test, the advantage is completely misrepresented by the article by wrongly comparing the 15 - 30% number to the 94% or 97% numbers.

Also aside from any data and calculation issues there is the issue that the study is not independent. See [comment] and [PMID: 11829700].

(Aside: The EPCA-2 may also be usable to determine how aggressive the cancer is as well as whether there is cancer or not. See 2007 ASCO presentation by Dr. Getzenberg. Also according to the Schlotz article starting on page 8 of [link] there are indications that EPCA-2 may be able to distinguish between organ confined and non-organ confined disease.)

Example 2. Miraculins has a press release in which says that 70% of biopsies done as a result of a positive PSA test are needless -- the patient never had prostate cancer. Assuming the sensitivity and specificity for the PSA test are as in each of the 3 data sets (PCPT, Ontario, AFFP) in our table above what is the prevalence?

Answer:

We have three of the 4 variables in the PPV equation above, i.e. we have PPV, Sn and Sp, so we can get the remaining one, i.e. p, by implication. An easy way to do that is to use the free Mathomatic software which is available for Windows, Mac and Linux. Download, unzip it and double click it to start it up or just run it online via telnet as described near the beginning of this post. Then copy the following three lines to the clipboard and paste them into the running instance of Mathomatic. The first line is the PPV equation we showed previously, the second line tells it to solve for p and the third line tells it to calculate the result. After entering calculate it will prompt us for the values of the other variables and then display the value of p implied by them:


PPV = p * Sn / ((p * Sn) + ((1-p) * (1-Sp)))
p
calculate


The mathomatic session looks like this. Lines that begin with 1-> or the word Enter were entered by us and all other lines were output by Mathomatic.

1-> PPV = p * Sn / ((p * Sn) + ((1-p) * (1-Sp)))

                      p*Sn
#1: PPV = -----------------------------
          ((p*Sn) + ((1 - p)*(1 - Sp)))

1-> p

               PPV*(1 - Sp)
#1: p = --------------------------
        (Sn - (PPV*(Sn + Sp - 1)))

1-> calculate ; PCPT data
Enter PPV: .3
Enter Sn: .4
Enter Sp: .811
 p = 0.168399168399
1-> calculate ; Ontario data
Enter PPV: .3
Enter Sn: .6
Enter Sp: .947
 p = 0.036476256022
1-> calculate ; AAFP data
Enter PPV: .3
Enter Sn: .75
Enter Sp: .93
 p = 0.0384615384615

From the above we see that the implied prevalence is 17% based on the PCPT trial data or slightly less than 4% based on the other two data sets. 17% seems to be a bit high but the 4% seems within the ballpark we would expect so we accept the figures as reasonable.

Note that the formulas generated by Mathomatic can be used directly (as opposed to using Mathomatic calculate command) so if, in the future, we have another situation where we wish to calculate prevalence from sensitivity, specificity and PPV we could just use the above formula that Mathomatic came up with directly without access to Mathomatic at all. For example, in the case of the PCPT trial we could redo the calculation above like this:


p = PPV*(1 - Sp) / (Sn - PPV*(Sn + Sp - 1))
= .3 * (1 - .811) / (.4 - .3 * (.4 + .811 - 1))
= .168


Example 3. For this example we will calculate the specificity of the Miraculins test using a prevalence of 5%, the sensitivity given in the article and the PPV implied by the article. The article says that the fraction of all biopsies that are needless with the Miraculins test is 20% less than the needless biopsies based on the PSA test alone. Also the the sensivity of their test is 96%. Let us calculate the specificity of their test.

Answer. The fraction of all biopsies that are needless is 1-PPV and we know that this is 20% less than the 1-PPV of the PSA test. For the PSA test, the PCPT trial yielded a PPV of 10% so 1-PPV is 90% and if the Miraculins' 1-PPV is 20% less than that then the Miraculins' 1-PPV is (1 - 0.20) * 0.90 = 0.72 so the Miraculins' PPV is 1-0.72 = 0.28. Assuming a prevalence of p = 0.05 we have:

- PPV = 0.28
- p = 0.05
- Sn = 0.96

We have previously entered the PPV equation into Mathomatic so we need not enter it again. We can just enter:


Sp
calculate


and we get the following. When it prompts for the variable values we enter them. Below we show the actual session including the output from Mathomatic. We had previously entered the PPV equation so we did not have to enter an equation again. We simply entered Sp which caused Mathomatic to solve the equation alrady entered so as to express Sp in terms of the other variables. Then we enter calculate and it prompts for the value of each of those variables finally giving us a value of specificity of 0.87 for the Miraculins test (based on our assumptions and the data in the news article).

1-> Sp

             p*Sn*(1 - PPV)
#1: Sp = -1*(-------------- - 1)
             (PPV*(1 - p))

1-> calculate
Enter PPV: 0.28
Enter p: 0.05
Enter Sn: 0.96
 Sp = 0.87007518797

Example 4. Sensitivity

Suppose that in the last example we did not know the sensitivity but did know values of the other three variables. Then we could solve for sensitivity as shown below.

1-> Sn

         (1 - Sp)*PPV*(1 - p)
#1: Sn = --------------------
            (p*(1 - PPV))

1-> calculate
Enter PPV: 0.28
Enter p: 0.05
Enter Sp: 0.87
 Sn = 0.960555555556

We get a sensitivity of 0.96 which agrees with the prior exammple.

Example 5. Specificity of PSA Nanotest

This article on an experimental PSA Nanotest indicates that the test has a sensitivity of Sn = 100% and 1-PPV of 24% so PPV = 76%. Assuming the prevalence is p = 5% what is the specificity?

Answer:

Using the equation for specificity that we got from Mathomatic in Example 3, we have:


Sp = (p*((Sn*(1 - PPV)) + PPV) - PPV) / (PPV*(p - 1))
= (.05 * ((1 * (1-.76)) + .76) - .76) / (.76 * (.05 - 1))
= .983


Example 6. Color Doppler

In Kuligowska et al, 2001 the authors write: "Color Doppler US alone had a sensitivity of 43.2%, a specificity of 66.4%, a PPV of 40.8%, an NPV of 68.5%, and an accuracy of 58.3%." Let us determine what prevalence the paper is assuming to get this PPV.

In Example 2 we already computed the formula for prevalence using Mathomatic so instead of repeating it let us just write it down again from Example 2:


p = PPV*(1 - Sp) / (Sn - (PPV*(Sn + Sp - 1)))
= .408 * (1 - .664) / (.432 - .408 * (.432 + .664 - 1))
= 0.35


Assuming that the population under test is those who have a positive PSA test, the prevalence of such a population would be the PPV of the PSA test. This is higher than the PPV from the PCPT trial and the AAFP data but is consistent with the Ontario data.

If we were to use the PPV of the PCPT trial as the prevalence then the PPV of color doppler would be much less than the 40% claimed above:


PPV = p * Sn / (p * Sn + (1-p) * (1-Sp))
= 0.10 * 0.432 / ((0.10 * 0.432 + (1 - 0.10) * (1 - 0.664)))
= 0.125


Thus based on the PPV of the PSA test from the PCPT data if you have a positive color doppler there is a 12.5% chance of having prostate cancer. The corresponding figures for the Ontario and AAFP data are 37.3% and 30.8%. Thus if we accept the Ontario data the PPV calculated for color doppler is about right but it seems high relative to the PCPT and AAFP data.

Interestingly the PPV of color doppler is higher than for PSA even though both the sensitivity and the specificity are lower than for PSA. That is because the color doppler is used on a population with higher prevalence of prostate cancer, namely the population of patients with positive PSA tests.

Example 7. Power Doppler

Power doppler is a type of color doppler that uses the amplitude of the echo rather than its frequency shift and is believed to be better at detecting cancer.

In a 1998 investigation reported in [PMID: 9586699] there were 23 patients with prostate cancer and 19 of them
were successfully detected with power Doppler sonography. Also there were 17 who did not have prostate cancer and 4 of them had positive tests anyways. Thus we have this table:

	PCa	No PCa
Test +ve	19	4	23
Test -ve	4	13	17
	23	17	40

The diagonal elements divided by their column totals ive the sensitivity, 19/23 = 82.6%, and specificity, 13/17 = 76.5% and the upper diagonal element divided by its row total gives the positive predictive value, PPV, which is 19/23 = 82.6% -- in this case is coincidentally the same as the sensitivity. The prevalence is the ratio of the first column total to the grand total which is 23/40 = 57.5% . This is much higher than the prevalence of prostate cancer in the general population but such imaging would likely only be done on patients who already had some positive indication from a PSA test or DRE. At any rate the PPV but not the sensitivity and specificity depend on the prevalence.

A second 1998 power doppler investigation was described like this: [PMID: 9772875]

OBJECTIVE: To determine the role of transrectal power Doppler ultrasonography (PDU) in the diagnosis of prostate cancer. PATIENTS AND METHODS: Thirty-six patients (mean age 66.4 years, SD 7.7, range 59-82) with possible prostate cancer, suspected from an abnormal digital rectal examination or elevated prostate specific antigen level, underwent transrectal ultrasonography, transrectal PDU and biopsy. The vascularity on PDU was graded on a scale of 0-2, where grade 1-2 was considered positive and grade 0 negative. RESULTS: The vascularity was grade 2 in 11 patients, grade 1 in 11 and grade 0 in 14; 20 of the 36 (56%) patients had prostate cancer. Of the 22 patients positive on PDU, 18 had malignant disease and four benign; two of 20 patients with histopathologically confirmed malignancy had a normal PDU. The sensitivity of PDU was 90%, the specificity 75% and the positive predictive value 82%. CONCLUSION: Focal hypervascularity on PDU was associated with an increased
likelihood of prostate cancer. Although ultrasonography alone cannot detect all
cancers, even using PDU, the technique appears to increase the sensitivity and
to help identify appropriate sites for biopsy.

The reader may wish to try filling out the table prior to looking:

	PCa	No PCa
Test +ve	18	4	22
Test -ve	2	12	14
	20	16	36

and then calculate the sensitivity, specificity, PPV and prevalence (which should be 18/20 = 90%, 12/16 = 75%, 18/22 = 81.8% and and 20/36 = 56%).

These numbers are reasonably close to the first investigation so their consistency seems favorable.

A third study found that targeting biopsies at areas of high blood flow resulted in higher detection rates: [Science Daily].

More information on power Doppler sonography is summarized on the web site of Dr. Bard here: [here].

Other definitions

The cancer detection rate (or just detection rate) is used to refer to the fraction of actual cancers found. That is it is the number of subjects that tested positive and had cancer divided by the number of subjects. It is also called the true positive rate.

The term test prevalence is used to refer to the fraction of positive tests among the entire population tested. This statistic does not take into account whether the test is correct or not. It just uses the number of all positive tests as the numerator.

In these terms the PPV is the cancer detection rate divided by the test prevalence while the sensitivity is the cancer detection rate divided by the true prevalence.

Cutoff Value

As one varies the cutoff value that separates positive from negative test scores the sensitivity and specificity change. Similarly the PPV and NPV change. In fact, one can arrange for the PSA test to have any sensitivity desired by modifying the cutoff level sufficiently. Similarly one can arrange for the PSA test to have any specificity desired; however, if we fix either the sensitivity or the specificity then the other will be determined. We cannot fix both at once.

ROC Curve

If for each cutoff value we plot the fraction of positive tests among diseased individuals (along the vertical axis) against the fraction of positive tests among the healthy individuals (along the horizonal axis) we get a curve known as the receiver operator curve (ROC). In terms of sensitivity and specificity this amounts to plotting the sensitivity against 1-specificity for each possible cutoff value. The curve will start at (0,0) which corresponds to a very high cutoff value and rises to (1,1) which corresponds to a cutoff value of 0 or at least a very low cutoff value. More information on ROC curves can be found in these articles: Wikipedia and anaesthetist.

AUC

The area under the curve (AUC) is the fraction of sensitivity/specificity combinations that are worse than that of the test under consideration. The larger the AUC the better.

Comparing Tests

There are a number of ways of comparing tests:

1. ROC. If we plot the ROC curves for two tests on the same graph and if one lies entirely above the other than that one has a higher sensitivity than the other for every specificity and so the higher curve represents a uniformly superior test.


2. Sn+Sp-1. A test with sensitivity plus specificity equal to 1 is no better than random. For example, suppose we flip a coin and assign subjects a positive test if it comes up heads. That test has a sensitivity and specificity each of 0.5 so sensitivity plus specificity equal 1. If the coin is biased, e.g. .9 heads, then we get a sensitivity of .9 and specificity of .1. By varying the bias we can get any desired combination of sensitivity and specificity that sum to 1. Thus we can use sensitivity + specificity - 1 as a gauge of how much better a particular test at a particular cutoff is relative to a random assignment.


3. AUC. As mentioned previously, the area under an ROC curve (AUC) is the proportion of sensitivity, specificity pairs that are less than the test in question. This can be used as a measure of test desirability. A larger number is better.


4. Utilities. Except for #1 these methods do not explicitly take into account the seriousness of the two sorts of errors: not detecting that someone has prostate cancer vs. subjecting patients to needless further testing. Clearly the first error (missing someone with cancer) is the more serious; however, its probably not feasible to subject everyone to biopsy so some tradeoff needs to be made. If one could assign costs (not necessarily monetary) to the two errors a tradeoff analysis could be made.

Bayes Odds Form

The Bayes odds form for PPV and 1-NPV can be entered into Mathomatic instead of the explicit forms given before. They are entirely equivalent so its a matter of taste which one we use. If we were solving the equations by hand the odds form do have the advantage of being easier to solve for the component variables. The odds forms also have a certain attractive compactness and elegance to them.

The Bayes Odds form of the equations for PPV and NPV can be written as shown:


PPV / (1-PPV) = Sn / (1-Sp) * p/(1-p)
(1-NPV) / NPV = (1-Sn) / Sp * p/(1-p)


Here the quantities Sn / (1-Sp) and (1-Sn) / Sp are known as the positive and negative likelihoods. The left sides are the odds of PPV and the odds of 1-NPV. The p/(1-p) on the right hand side is the odds of p. The term odds is used here in the same sense as in racetrack betting.

Odds and probabilities are equivalent. In fact, the following two formulas can be used for translating between odds and probabilities:


odds = probability / (1 - probability)
probability = odds / (1 + odds)


The Bayes odds forms are easier to solve for Sn and Sp so instead of solving directly for the probability of PPV or p one solves for their odds and then uses the odds to probability conversion formula.

See: [link] and [link].

Tabular Form

We can form a 2x2 table in which diseased and healthy correspond to columns one and two and postiive test (i.e. indicative of disease) and negative test (i.e. indicative of healthy) correspond to rows. If p is prevalence, pT is test prevalence, Sn is sensitivity, Sp is specificity, PPV is positive predictive value and NPV is negative predictive value we can fill it in two ways like this:

        Diseased         Healthy         Total     
+Test   p * Sn           (1-p) * (1-Sp)  pT
-Test   p * (1-Sn)       (1-p) * Sp      1-pT
Total   p                1-p             1

        Diseased         Healthy         Total     
+Test   pT * PPV         pT * (1-PPV)    pT
-Test   (1-pT) * (1-NPV) (1-pT) * NPV    1-pT
Total   p                1-p             1

The first form is particularly useful since it allows us to form the entire body of the table table given just p, Sn and Sp. The totals can then be calculated by summing the rows and columns. If we were given the PPV, NPV and pT then we could calculate the body of the table using the second formulation although this is less common.

The table can also be used to create the formulas that we have used Mathomatic to derive. For example, with reference to the first table, since PPV is the detection rate, i.e. the upper left cell, divided by the detection prevalence, i.e. the first row total, we have:


PPV = detection rate / detection prevalence
= [upper left cell] / [first row total]
= [upper left cell] / ([upper left cell] + [upper right cell in body])
= [p * Sn] / ([p * Sn] + [(1-p) * (1-Sp)])


which indeed is the formula we started out with when discussing PPV.

PSA Screening and Early Detection. Part 1 - Guides
PSA Screening and Early Detection. Part 2 - Key Points on PSA
PSA Screening and Early Detection. Part 3. Current Environment
PSA Screening and Early Detection - Part 4. Diagnostic Testing Concepts [previous]
PSA Screening and Early Detection - Part 5. More Diagnostic Testing Concepts [current]