Spread of prostate carcinoma to the perirectal lymph node basin: analysis of 112 rectal resections over a 10-year span for primary rectal adenocarcinoma.

We recently identified metastatic prostate carcinoma (PCA) within perirectal lymph nodes (PLNs) from 2 patients undergoing abdominoperineal resection (APR) for rectal adenocarcinoma (RA). As this phenomenon has not been addressed by any studies in the literature and because these positive PLNs had the potential to be mistakenly diagnosed as metastatic RA, we were prompted to undertake a retrospective study of rectal resections for RA to determine the frequency of PCA metastasizing to the PLNs in this patient population. The laboratory information system of the Department of Pathology, Capital Health, Halifax, Nova Scotia was searched for lymph node (LN)-positive RAs resected by low anterior resection or APR in male patients between January 1, 1992 and December 31, 2002. The hematoxylin and eosin slides were retrieved and reviewed, comparing the histology of the primary rectal tumor with that of the LN metastases in each case. Metastases having a different histologic appearance than the primary rectal tumor or having a pattern suggestive of metastatic PCA were analyzed by immunohistochemistry to detect prostate specific antigen (PSA), prostatic acid phosphatase (PAP), cytokeratin 7 (CK7), cytokeratin 20 (CK20), and carcinoembryonic antigen in LN metastases and in each RA. The presence or absence of mucin in the tumors was assessed by staining with Alcian blue, periodic acid-Schiff (PAS) +/- diastase, and modified PANFOPAS (2-hydroxy-3-naphthoic acid hydrazide/fast black B/saponification/periodic acid-Schiff). The study identified 112 cases of RA with positive LNs. Of those, 5 of 112 (4.5%) were identified as having metastatic PCA within the PLNs. All five were positive for PSA and PAP and only one case had rare CK20-positive tumor cells. The primary RAs were all diffusely positive for CK20 and carcinoembryonic antigen. Two cases of metastatic PCA expressed colonic type/acetylsialomucin, which was also seen in well-differentiated primary RAs. These 5 patients had a mean age of 76.8 years (range, 68-82 years). Four (80%) underwent APR while one (20%) underwent a low anterior resection. The mean number of LNs identified per case was 14 (range, 5-26). The mean number of LNs per case with metastatic PCA was 7.6 (range, 1-18). The majority of the LNs were under 1.0 cm in diameter. Two cases (40%) were associated with significant extranodal extension of PCA. The PLNs were mistakenly diagnosed as being involved by RA 40% of the time. On follow-up, 2 patients (40%) had died with progressive pelvic tumor, while 3 patients (60%) were alive, including 2 who, as a result of the study, were referred to the urology service for management. Both of these patients were subsequently started on medical (anti-androgen) therapy and 1 additionally had bilateral orchiectomy. In 4 of the patients, the serum PSA after the rectal resection ranged from 2.5 to 13.5 ng/mL. In 1 patient, the serum PSA was markedly elevated (5961 ng/mL), and this patient was subsequently identified on bone scan as having extensive skeletal metastases. This study identified a subset (4.5%) of patients with RA and PLNs positive for PCA. PCA may extend to the PLN basin and therefore influence the management of patients with rectal tumors and the staging, LN dissection, and management of patients with PCA. Moreover, the LNs were incorrectly diagnosed as metastatic RA 40% of the time, emphasizing the need to consider the differential diagnosis of metastatic PCA when evaluating PLNs. The diagnosis of metastatic PCA can be confirmed using an immunohistochemical panel consisting of PSA, PAP, CK20, and carcinoembryonic antigen along with mucin stains in some cases. In patients undergoing APR for RA, there should be preoperative screening for PCA as it will not be possible to do a digital rectal examination or transrectal ultrasound post-APR, and a prostate biopsy, if necessary, would have to be done via the more difficult transperineal approach.