Is there an easier way to determine whether early detection of prostate cancer reduces mortality?
Is there an easier way to determine whether early detection of prostate cancer reduces mortality?
复制标题
有没有更简单的方法来确定前列腺癌的早期检测是否可以降低死亡率？
DOI：
10.1046/j.1525-1497.1997.07129.x
复制
发表时间：
1997
影响因子：
5.7
通讯作者：
Barry,MJ
中科院分区：
文献类型：
--
作者：
Barry,MJ
当直肠指检（DRES），特别是血清前列腺特异性抗原（PSA）检测用于筛查前列腺癌时，在早期阶段发现更多的癌症。这一阶段的转变表明，筛查可以识别出更容易治愈的癌症，并提高了降低前列腺特异性癌症死亡率的可能性。另外，许多通过筛查发现的早期癌症可能比其他癌症恶性程度低，这将限制它们对死亡率的影响，从而限制我们降低死亡率的机会。究竟会有什么样的结果，还有待观察。1，2美国和欧洲正在进行临床试验来回答这个问题，但由于大多数前列腺癌的倍增时间较长且预后良好，因此需要数千名受试者和十年或更长时间的随访。与此同时，大多数美国医生和他们的病人已经接受了早期发现前列腺癌的目标，等待临床研究的结果。一种更简单、更快速的方法来确定筛查是否利大于弊是最受欢迎的。病例对照研究可能会提供一些答案。如果这些研究发现晚期或致命性前列腺癌患者的筛查频率低于对照组，我们可能会得出结论，筛查是保护性的。然而，在目前被称为“PSA时代”的实践中，几乎每个新诊断为前列腺癌的患者在诊断前都会进行DRE和PSA测试。因此，病例对照研究中的一个关键方法学问题是如何将应计数的筛选研究与不应计数的诊断研究分开。在这个问题上，孔卡托博士证明了为什么这项任务并不容易。4传统上，在无症状患者中进行的测试是筛查测试，在有症状患者中进行的测试是诊断测试。当症状与测试目的无关时，就会出现异常。例如，为患有头痛的男性进行的PSA测试仍然是前列腺癌的筛查测试，因为头痛不会增加前列腺癌的可能性。对于涉及前列腺疾病的病例来说，这种传统的区分是模糊的。最重要的问题是，与良性前列腺增生一致的下尿路症状是否会增加前列腺癌的可能性，调整年龄的影响，这与这些症状和前列腺癌有关。如果他们这样做，那么几乎所有的DREs和PSA测试都是诊断测试，因为大多数50岁以上的男性至少有一个下尿路症状，如偶尔尿失禁。早期发现的一个支持者认为情况确实如此。然而，一些研究表明，这些症状并不会增加前列腺癌的可能性，至少在DREs和PSA检测中是如此。6-8如果这些研究是正确的，那么所有的DREs和PSA测试都是筛查测试，无论是否存在下尿路症状。孔卡托认为，真相介于两者之间。他仔细开发了一个系统，根据研究人员对前列腺癌筛查的确定程度，将其分为五类。他的系统是基于下尿路症状的存在和稳定性以及任何早期前列腺癌测试的结果。Concato使用该系统对少数前列腺癌患者的DRE进行分类。然后，他进行了一系列假设对照的病例对照研究。他发现，当DRE被分为不同类别时，优势比也不同。虽然孔卡托可能高估了效果...
When digital rectal examinations (DREs) and particularly serum prostate-specific antigen (PSA) tests are used to screen for prostate cancer, more cancers are found in earlier stages. This stage shift suggests that screening identifies more readily curable cancers and raises the possibility of reducing prostate-specific cancer mortality rates. Alternatively, many early-stage cancers detected by screening may be less malignant than other cancers, which would limit their effect on mortality rates and thus our opportunity for reducing them. Which outcome will occur remains to be seen. 1, 2 Clinical trials are under way in the United States and Europe to answer this question, but they will require thousands of subjects and a decade or more of follow-up because of the long doubling times and favorable outcomes of most prostate cancers. Meanwhile, most US physicians and their patients have embraced the goal of early detection of prostate cancer, pending the results of clinical studies. 3 An easier, faster way to determine whether screening does more good than harm would be most welcome.Case-control studies may provide some answers. If these studies found that patients with advanced or fatal prostate cancer had been screened less frequently than control subjects, we might conclude that screening is protective. In the current practice of what some have called the “PSA era,” however, nearly every patient with a new diagnosis of prostate cancer will have had a DRE and PSA test before the diagnosis. Therefore, a key methodologic issue in case-control studies is how to separate screening studies, which should be counted, from diagnostic studies, which should not be counted. In this issue Dr. Concato demonstrates why this task is not easy. 4 Traditionally, a test is a screening test when it is done among asymptomatic patients, and it is a diagnostic test when it is done among symptomatic patients. An exception occurs when a symptom is unrelated to the purpose of the test. For example, a PSA test done for a man with a headache is still a screening test for prostate cancer because the headache does not raise the probability of prostate cancer. This traditional distinction is blurred for cases involving prostatic diseases. The most important issue is whether lower-urinary-tract symptoms that are consistent with benign prostatic hyperplasia raise the probability of prostate cancer, adjusting for the effect of age, which is related to both these symptoms and prostate cancer. If they do, then almost all DREs and PSA tests are diagnostic tests, because most men over age 50 have at least one lower-urinary-tract symptom such as occasional nocturia. One proponent of early detection has suggested that this is the case. 5 Several studies suggest, however, that these symptoms do not raise the probability of prostate cancer, at least as detected by DREs and PSA tests. 6–8 If these studies are correct, then all DREs and PSA tests are screening tests, despite the present or absence of lower-urinary-tract symptoms. Concato suggests that the truth lies somewhere in between. He has carefully developed a system for classifying prostate cancer tests into five categories according to how certain the investigator is that they were screening tests. His system is based on the presence and stability of lower-urinary-tract symptoms and on the results of any earlier tests for prostate cancer. Concato used the system to classify DREs in a small number of patients with prostate cancer. He then conducted a series of case-control studies with hypothetical controls. He found different odds ratios when DREs were classified into different categories. Although Concato may have overestimated the effect by …