Prostatic intraepithelial neoplasia and putative precursor lesions of prostate cancer: a clinical perspective A.N. VIS*² and TH.H. VAN DER KWAST* *Department of Pathology, Josephine Nefkens Institute, and ²Department of Urology, Erasmus University and University Hospital Rotterdam, The Netherlands Introduction In recent years, prostate cancer has become an increas- ing health problem in North America and Western Europe, now being the most commonly diagnosed noncutaneous malignancy in men beyond middle age, and the second cause of cancer-related death after lung cancer [1]. The causes of prostate cancer, the target cells of prostatic carcinogenesis, and the histological changes preceding and leading to the initiation and progression of prostate cancer have yet to be elucidated. Many research groups are trying to solve the puzzle of prostatic carcinogenesis, with their attention focused within the morphological continuum between benign glands at one end, to premalignant lesions and invasive disease at the other. Also clinicians are sometimes confronted with morphological features on the diagnostic prostatic needle biopsy that although negative for cancer raise suspicion of concomitant malignancy. These ®ndings present a particular diagnostic challenge (Table 1). This review highlights the current understanding and knowledge of the main putative premalignant lesions of the prostate and of lesions that raise particular suspicion of concomitant malignancy. Their association with clinical variables and incidence rates were assessed in different study groups and in ours, as were the pre- dictive values for prostate cancer on follow-up biopsy. The consequences of ®nding these distinct morphological entities on the diagnostic needle biopsy are set in a wider clinical perspective. Possible target lesions of prostatic carcinogenesis Carcinogenesis is a complex multistep process, involving molecular, cellular and histological changes. It describes the conversion of benign epithelial glands, through premalignant lesions, to invasive carcinoma. Several requirements should be met to consider a lesion pre- malignant (Table 2). An epidemiological relationship must be shown, especially when the development of a premalignant lesion to early stromal invasion and full-blown malignant disease takes months or years. The precursor lesion presents at an earlier age than its malignant equivalent, and the age-adjusted prevalence is expected to rise synchronously with that of histologically con®rmed and/or clinically manifest malignant disease. Typically, the age-adjusted prevalence of the precursor lesion decreases at a particular time while that of cancer continues to increase (Fig. 1). As an epidemiological association does not preclude that premalignant and malignant conditions simply coexist with each other, clear morphological (cellular, histological, architectural) similarities should also be present. In addition, organs harbouring invasive cancer should have a greater fre- quency, severity and extent of the premalignant lesion than organs that have not. In the organ, premalignant lesions should be close to their presumed malignant equivalents, whereas sometimes micro-invasion of the stroma by the precursor lesion may be seen at the microscopic level. The de®nite proof of a relationship between a precursor lesion and malignancy is the clinical evidence of progression into invasive carcinoma. Previously, several morphological lesions have been proposed that may act as potential precursor lesions of prostatic adenocarcinoma. These are the morphologi- cally distinct entities of focal atrophy or post-atrophic hyperplasia (PAH), atypical adenomatous hyperplasia (AAH) or adenosis, and prostatic intraepithelial neoplasia (PIN). Lesions designated as `atypical' or `suspicious' have also been associated with the presence of prostate cancer. As a wide diversity of morphological features is reported, a clear description in histological terms is not possible and consequently `suspicious for malignancy' lesions should not be viewed as premalignant lesions of the prostate in a strict sense, but should be regarded as a separate diagnostic entity associated with concomitant prostate cancer. It is plausible that the aforementioned morphologies may not account for all malignancies of the prostate, and that the human prostate gland may harbour other hitherto unrecognized premalignant lesions. BJU International (2001), 88, 147±157 # 2001 BJU International 147