For those of us who are not computer literate, the sound of the post arriving is still exciting — perhaps it is the remembered pleasure of receiving a letter at boarding school or during military basic training. Most days now, though, my mail includes letters that are far from cheering. Many are from patients with prostate cancer who, by the time treatment was started, found that it was no longer curable but only manageable.

Sometimes the cancer had been without early symptoms, but in many cases the patient’s symptoms, or worries, had been dismissed as no more than a sign of advancing middle age. In other cases, sometimes despite a family history of prostate cancer, a sensible request for a PSA test was dismissed.

Fortunately, not all the letters we receive at The Times about prostates are gloomy. Many express the writer’s gratitude for being told about the PSA test in our reports.

This week some reports — based on a study by the US Department of Veterans’ Affairs in Connecticut — which suggested that the PSA test was not worthwhile have confused patients and angered many urologists.

The urologists were worried that the research team was not led by people with an established history in prostatic medicine, that the sample size was small and that there was inadequate analysis of the patients’ medical history. Consequently, they felt that the lessons drawn were not as conclusive as was subsequently represented.

No diagnostic screening tests are ideal. Although colonoscopy for detecting colorectal carcinoma is more reliable than other diagnostic procedures, occasional cases are still missed. Mammography, now known to save many lives, was notoriously inaccurate before it was refined.

No one has ever denied that the PSA test has faults. There are false positives — tests in which a patient is shown to have an abnormal PSA, though subsequent tests reveal an apparently normal gland. As a result of these apparently misleading results, some patients have had to undergo tiresome, uncomfortable tests. In other cases an early single PSA test may give a false negative result, suggesting that all is normal although a cancer is present.

Fortunately these tests affect only a minority of patients. The mistake is to regard a single PSA test as giving a definitive answer. Repeated PSA tests at regular intervals — which all men over 40 should have if there is a family history, and all men over 50 regardless — will usually reveal false negatives as the PSA, although still low, will be rising abnormally quickly.

The PSA test is like a sieve; it will collect patients who need further investigations, which in turn usually provide a more precise diagnosis and, importantly, prognosis. Likewise, serial PSA testing detects many cancers when there is a false negative PSA.

Unfortunately not all false positive PSAs are later shown to have been unnecessarily gloomy and disturbing. Recent research suggested that men with a false positively raised PSA were more likely to develop cancer of the prostate at a later date than those whose initial tests showed a normal PSA. The supposition must be that cancer was always there but that the subsequent tests failed to detect it.

The standard post-PSA test is the transrectal ultrasound and biopsy; the blame for its failure to detect cancer despite a raised PSA is attributed to an overreliance on the established practice of taking only six biopsies on one occasion. Even in the best hands, ultrasound misses many prostatic tumours, so biopsies are often done without seeing the tumour.

In many units it is now accepted that a patient with an apparently false positive should be kept under review and have further biopsies. To achieve better diagnosis routinely, either more biopsies need to be taken initially, after the first PSA test — uncomfortable for the patient — or the PSA should be repeated early if it doesn’t return to normal.

In false negative PSA results, the cancer cells are often poorly differentiated and of high malignancy. Even though such cells may not secrete as much PSA as well differentiated, less malignant cancer cells, if serial PSAs are carried out the abnormal rate of increase will be detectable. In these cases, repeat PSA tests and early definitive diagnosis is even more important for long-term survival.

There is an intensive search for a better test than a PSA. But whatever that test is, it will still need confirmatory investigations and a more effective biopsy regime.