© 2 0 0 5 B J U I N T E R N A T I O N A L | 9 5 , 3 1 – 3 3 | doi:10.1111/j.1464-410X.2005.05243.x 31 PREVALENCE OF RCC DIAGNOSED AT AUTOPSY MINDRUP et al. The prevalence of renal cell carcinoma diagnosed at autopsy STEVEN R. MINDRUP, JESSICA S. PIERRE, LAILA DAHMOUSH* and BADRINATH R. KONETY† Departments of Urology, *Pathology and †Epidemiology, University of Iowa, Iowa City, IA, USA Accepted for publication 2 August 2004 OBJECTIVE To compare the rate of renal cell carcinoma (RCC) detected only at autopsy, from two periods, to determine if the apparent recent increase in RCC in the USA is a true increase or mainly a result of improved imaging techniques, as a true increase in the clinical incidence of RCC should not affect the number of previously undiscovered RCC found only at autopsy. PATIENTS AND METHODS We identified all individuals who underwent autopsy in two periods, 1955–60 (3307) and 1991–2001 (2938), and who were indentified with renal tumour either before or after death. Information was obtained on age, sex, histological type and size of tumour, and whether the tumour was identified before or only after death. All cases of RCC detected at autopsy were reviewed for this analysis by one pathologist (L.D.). RESULTS Between 1955–60 and 1991–2001 there was a 55% reduction in the mean annual number of autopsies. The proportion of malignant renal masses (RCC) did not change significantly (35.4% to 32.8%). The rate of previously unsuspected RCC detected only at autopsy did not change significantly (0.91 vs 0.72 per 100 autopsies). CONCLUSION The number of renal masses, both benign and malignant, discovered only at autopsy is declining, possibly because of better detection before death. However, the rate of occult kidney cancer per 100 autopsies did not change significantly between the two periods, suggesting a true increase in the frequency of clinically detected kidney cancer. KEYWORDS carcinoma, renal cell, incidence, diagnosis, autopsy INTRODUCTION Tumours of the kidney and renal pelvis account for 3% of all new cancer diagnoses and 3% of cancer deaths. The are an estimated 35 710 new cases of renal cancer diagnosed in 2004, with an estimated 12 480 resulting deaths. Survival rates have increased from 56% in 1983–85 to 62% in 1992–97 [1]. Current reports show an increasing incidence of renal cancer. Several earlier series postulated this increasing incidence to be a result of more widespread and aggressive imaging techniques [2–4], but more recent series show that the incidence at all stages has increased [5]. In an analysis of the Surveillance, Epidemiology and End Results database, Hock et al. [6] confirmed that the incidence of localized, regional, and distant metastatic renal cancer had increased over the last 20 years. Some authors have called for a re-evaluation of incidence data, to refute or support these observations [7]. We sought to address this issue by comparing the incidence of occult renal cancer diagnosed at autopsy to the incidence of clinically diagnosed RCC during life over two periods separated by 30 years. The hypothesis was that if the observed increase in the incidence of asymptomatic RCC was a result of the widespread use of sophisticated imaging techniques, the rate of occult RCC detected only at autopsy should be declining, as most renal tumours would more likely be detected during a patient’s lifetime. Conversely, if there were a true increase in the incidence of RCC between these periods, the rate of detection only at autopsy would be unchanged or increase. PATIENTS AND METHODS The authors’ institutional autopsy database was searched to identify all individuals found to have a renal tumour of any type either before or after death in the two periods 1955– 60 (3307 patients) and 1991–2001 (2938). Analysis of a longer, more recent period was needed to obtain a comparable number of autopsies. We also selected a comparative period of study in the 1950s to avoid sample contamination by the possible early use of imaging techniques, e.g. ultrasonography and CT. Standard autopsy procedures at our institution mandate that all kidneys are bi- valved and inspected grossly for visible lesions. All observed lesions are step- sectioned. In kidneys with no lesions, or if the patient had no history suggestive of renal disease, a representative transverse section of tissue encompassing the cortex, medulla and renal pelvis is submitted for histological examination. All autopsy reports were reviewed by two authors (J.S.P. and B.R.K.) and information extracted on age, sex, histological type and size of tumour, and if the tumour was clinically identified before death or at autopsy only. Data on the incidence of clinically identified RCC at our institution during the corresponding periods were obtained from the University of Iowa institutional cancer registry. All of the slides of the original haematoxylin and eosin- stained sections of the kidneys from all of the autopsy specimens reported to have RCC or adenoma were reviewed again by one pathologist (L.D.). Tumour size, Fuhrman grade and histological subtype of RCC were assessed. All cases of occult RCC were identified and confirmed through the pathological review. Student’s t-test and a chi-square test were used to compare tumour size, tumour histological subtype distribution