Metastatic Patterns of Prostate Cancer: An Autopsy Study of 1,589 Patients LUKAS BUBENDORF, MD, ALAIN SCHOPFER, MD, URS WAGNER, MD, GUIDO SAUTER, MD, HOLGER MOCH, MD, NIELSWILLI, MD, THOMAS C. GASSER, MD, AND MICHAEL J. MIHATSCH, MD The prognosis of prostate cancer is mainly determined by the presence or absence of metastases. Nevertheless, the metastatic pathways in prostate cancer are not entirely understood. Among 19,316 routine autopsies performed from 1967 to 1995 on men older than 40 years of age, the reports from those 1,589 (8.2%) with prostate cancer were analyzed. Hematogeneous metastases were present in 35% of 1,589 patients with prostate cancer, with most frequent involvement being bone (90%), lung (46%), fiver (25%), pleura (21% ), and adrenals (13 %). Several lines o f evidence suggested the existence of a backward metastatic pathway through veins from the prostate to the spine in addition to classical hematogeneous tumor spread via the vena cava. F'LrSt, there was an inverse relationship between spine and lung metastases, suggesting that metastasis to the spine is independent of lung metastasis. Second, the maximmn frequency of spine involvement occurred in smaller tumors (4 to 6 cm) as compared with the maximum spread to lung (6 to 8 cm) and fiver (> 8 cm), suggesting that spine metastases precede lung and fiver metastases in many prostate cancers. Third, there was a gradual decrease in spine involvement from the lumbar to the cervical level (97% v 38%), which is consistent with a subsequent upward metastatic spread along spinal veins after initial lumbar metastasis. The results of this study show that bone, lung, and fiver are the most frequent sites of distant prostate cancer metastases. Besides the cava-type of metastasis through lung passage, there are strong arguments for the existence and clinical significance of a backward venous spread to the spine, which is likely to occur early in the metastatic process. HUM PATHOL31:578-583. Copyright © 2000 by W.B. Saunders Company Key words: prostate cancer, autopsy, metastasis. Despite the widespread use of prostate-specific antigen screening for early detection, prostate cancer remains the second leading cause of cancer-related death in Western countries, a There is a considerable discrepancy between the histological prevalence and clinical disease from prostate cancer, reflecting the broad spectrum of tumor behavior. 2 Although more than 30% of all men older than 50 years have been shown to harbor prostate cancer, only 9% develop clinical disease. 2 The outcome of prostate cancer is mainly determined by metastases. Hormone withdrawal therapy can initially relieve symptoms in a large propor- tion of patients with metastases, but long-term cure is rare because the tumors become hormone refractory after a few months or years, and efficient additional systemic therapies are not available. 3 Despite the clini- cal significance of metastasis in prostate cancer patients as well as the crucial role of the preoperative detection ofmetastases for treatment selection, the metastatic pathways are not yet fully understood. It has been suggested that besides the cava type of metastasis through the lung, alternative pathways may exist for hematogeneous tumor spread through periprostafic to prespinal veins into the spine. 4,5 Autopsies offer a unique opportunity to study the distribution of metasta- sis. The aim of this study was to evaluate the patterns and pathways of metastasis in a large number of rou- From the Institute of Pathology and Urologic Clinics, University of Basel, Sch6nbeinstrasse 40, Basel, Switzerland; and the Cantonal Institute of Pathology, Liestal, Switzerland. Accepted for publication January 27, 2000. Address correspondence and reprint requests to Lukas Buben- doff, MD, Institute for Pathology, University of Basel, Sch6nbein- strasse 40, CH-4003 Basel, Switzerland. Copyright © 2000 by W.B. Saunders Company 0046-8177/00/3105-0008510.00/0 doi:10.1053/hp.2000.6698 finely processed autopsies from patients with prostate cancer. MATERIALS AND METHODS Patients In a consecutive series of 19,316 autopsies from men older than 40 years of age performed at the Institute of Pathology of the University of Basel between 1967 and 1995, there were 1,589 men (8.2% of all autopsies) having either prostate cancer at autopsy or a history of previously treated prostate cancer. Eight hundred thirty-seven (52.7%) of these tumors were clinically known, and 741 (46.7%) were unsuspected (latent). Ten patients (0.6%) had suffered from known metastatic disease with clinically unknown (occult) primary tumor. The average age of tumor patients at autopsy was 78 _+ 8.6 years (range, 40 to 100 years). Residual tumor could not be detected after local treat- ment in 87 of the 837 patients with clinically known disease. In most cases, histological analysis was done only in case of suspicious macroscopic findings or gross abnormality. Morphological Features Tumor stage and maximum tumor diameter had been assessed macroscopically in most tumors. If not indicated, the local stage according to the TNM classifi- cation was reconstructed from the tumor extension described in the autopsy report. 6 Histological grading had subjectively been performed by different patholo- gists at the time of autopsy as high, intermediate, or low grade based on growth pattern and nuclear atypia. Data on local stage, histological grade, and tumor diameter were available in 1,393 (88%), 1,029 (65%), and 891 (56.1%) patients, respectively. 578