The third and most serious problem is prostate cancer, a medical oddity that is a long way from being understood. If all men lived to be ninety, almost all would harbor prostate cancer. It is especially widespread among African-Americans, who have the highest rates of death from prostate cancer in the world. According to John T. Isaacs, a medical oncologist at Johns Hopkins, data from autopsies suggest that about 11 million Americans have prostate cancer in some form. Lung cancer is thought to afflict about 300,000 Americans. The disparity in the prevalence of different types of cancer is baffling, and especially so when comparing prostate cancer with other kinds. "Cancer of the colon is extraordinarily common," says John McNeal, a pathologist in the department of urology at Stanford, "whereas cancer of the small intestine is very uncommon. How come? Here's two parts of the same organ, hooked together, and one appears to be quite susceptible to cancer and the other appears not to be." Yet more puzzling, he says, is the seminal vesicle, the small gland that produces most of the constituents of seminal fluid. "The seminal vesicle is attached to the prostate, and cancer of the seminal vesicle is, I would guess, the rarest of all human cancers, and it's only millimeters away from the organ that has one of the commonest cancers."
The American Cancer Society projects that no more than 165,000 prostate carcinomas will be discovered this year, whereas it expects 170,000 new cases of lung cancer and 182,000 of breast cancer. The difference is due in part to the relative newness of the PSA test, which for the first time makes diagnosis likely. But more important is the propensity for this cancer to lurk for years or even decades without apparent ill effects. As a rule it grows extremely slowly, taking four years or more to double in size. Breast carcinomas, in contrast, often double in less than three months; a tumor the size of a pinprick will fill a cubic centimeter in a couple of years and be ready to spread. According to McNeal, only about one out of five prostate carcinomas ever becomes clinically significant. The rest are sometimes called histologic carcinomas, meaning that tissue at the disease site fits the technical definition of cancer but no more. Many histologic carcinomas are microscopic; some are barely distinguishable from their surroundings. In some sense, he says, this dormancy is what one would expect. "A cancer so common as this would have to be very slowgrowing or it would have wiped out the population at the upper end," he points out. "It would not be something that people are just becoming aware of if it were like other cancers."
At present no one can distinguish at the outset between those carcinomas that will cause no trouble in a man's lifetime and thus should be ignored and those that will spread in time and thus should be treated, if at all possible. Scientists are not likely to solve this quandary in the near future. Cancer cells are generally thought to be created by a sequence of genetic mishaps. Every time a cell divides, there is a tiny chance -- typically one in a billion - that it will not copy its genetic material precisely and will instead produce a slightly abnormal daughter cell. (Such failures occur randomly, although their incidence can be increased by exposing the body to carcinogens, chemical compounds that in some way interfere with cell division.) In turn this imperfect daughter cell divides. There is a slight probability that it, too, will fail to copy itself exactly - actually, a slightly greater probability, because precancerous cells seem to be accidentprone. If the cell makes the "right" error, it will be one step further toward full malignancy. In colon cancer the mistakes usually occur in sequence. In prostate cancer, though, each mistake may be independent of the others, so that, say, step No. 4 could occur before step No. 2 and after step No. 6. All that matters is that over the years one cell or set of cells passes through every step. Nobody knows why this happens to one man rather than another, or why it happens especially often to AfricanAmericans.
The genetic material in a cell is deoxyribonucleic acid, or DNA, long, threadlike molecules twisted into forty-six bodies called chromosomes, which resemble the soft Hs in alphabet soup. Failures take place either when entire chunks of a chromosome snap off, fall away, or switch places, or when small bits of the strand get lost or mixed up. The former are easier to spot than the latter. For example, cancer of the colon is associated with gross chromosomal abnormalities, one reason that scientists have recently nailed down the sequence of events that leads to it. No such obvious changes have been spotted in prostatecancer cells, and hence the thinking is that the accidents involved are small and affect individual bits of DNA. If it were stretched out, the DNA in each human cell would be about six feet long; errors may involve snippets shorter than a millionth of an inch. The implication for researchers, unpleasant to contemplate, is a long period of slogging. Thus far the only clue is the possibility, turned up over the past three years by a group of researchers in Sweden and a second group at Johns Hopkins led by John Isaacs's brother William, that particular bits of chromosomes 8, 10, and 16 may be involved. Is a sudden clinical breakthrough possible? William Isaacs thinks it unlikely. "No one is going to shout 'Eureka!' on this one," he says. "The champagne is safe."
Imagine a blindfolded man with an unlimited supply of darts. He is standing before, say, six dart boards. (The exact number of steps necessary for prostate cancer to occur is unknown; colon cancer, as it happens, involves six major chromosomal changes, so I chose that number.) Blindly tossing away, he will eventually hit a bull'seye. That is equivalent to one of the "right" genetic accidents. After, say, two or more bull'seyes he will have histological cancer. After a long time he will probably hit all six bull'seyes. When he does, the cancer will be fully developed and ready to spread. The question is when treatment should occur. Surgery for prostate cancer involves the possibility of impotence, incontinence, and death. It also might not work: the cancer might already have spread, unnoticed, beyond the prostate. Considering these odds, the operation should take place immediately before he strikes the last bull'seye. But that's impossible. By definition, one can never predict when a random event will occur. A more realistic goal is needed. One of the bull'seyes - imagine that it is painted green - corresponds to the genetic change that triggers the propensity to spread that is cancer's most deadly feature. The other five bull'seyes can be hit without any real menace. The problem is that no one can say when a dart has struck the green bull'seye.
To be an ideal diagnostic test, the PSA would be positive only when the player has struck the green bull'seye. Instead it is like an alarm that rings when the player has hit two or more. It does not tell him which two bull'seyes he has hit; and because BPH, too, can increase the level of PSA, he may in fact have hit none at all. Are we better off for having installed this imperfect alarm nationwide? The answer depends on whether its use will lead to what social scientists call increased wellbeing. Everyone I have spoken with has an opinion on the use of the PSA test, favorable or un . But no one says that his or her opinion was empirically established, based on the sort of controlled data that biomedical researchers like. No scientific study has demonstrated that the PSA test saves lives, though it might seem logical to suppose it would. Indeed, none has shown definitively that surgery leaves people better off than doing nothing. As a result, John Wasson says, the United States is rapidly putting the cart before the horse. "We're using at an explosive rate a technology that has not proven itself," he says. "The Europeans at the international meetings I go to are aghast. They say, 'What are you people doing over there?"