Objective. To test the hypothesis that prostate cancer lymph node (LN)
micrometastases, undetected by standard histology, might be found usi
ng sensitive immunohistologic methods and may correlate to preoperativ
e prostate-specific antigen (PSA) levels. Method. Archival paraffin bl
ocks of pelvic lymphadenectomy specimens from radical prostatectomy we
re blindly submitted for immunostaining using pan-cytokeratin monoclon
al antibody SB-3, as well as antibodies directed against PSA. Automate
d immunostaining was performed on a Ventana Medical Systems 320 immuno
stainer. As a positive control, 7 cases with known nodal metastases by
standard histology were blindly analyzed and all has detectable micro
metastases by this methodology. Results. For 13 patients with PSA < 10
, 1 (8%) had LN micrometastases detected. For 10 patients with PSA bet
ween 10 and 20 and for 9 patients with PSA > 20, no occult metastases
were detected. We did find previously undetected prostate cancer (CaP)
LN micrometastases in 1 of 32 (3%) clinically localized prostate canc
er patients who had undergone radical prostatectomy In many LNs, cytok
eratin stains cross-reacted and stained individual plasma cells, where
as in the positive metastatic case, a cluster/nest of CaP cells were r
eactive. To the unfamiliar observer, the pitfall of false-positive res
ults because of nonspecific cytokeratin staining must be considered. T
hese results are in exact agreement with another recent study which al
so found only a 3 percent incidence of unsuspected pelvic lymph node m
icrometastases in clinically localized CaP utilizing similar methods.
Conclusions. Our hypothesis was not substantiated: LN micrometastases
were uncommon and did not correlate to serum PSA. Unlike studies with
breast cancer, occult micrometastatic nodal disease not appreciated by
standard methods appears to be uncommon in clinically localized prost
atic carcinoma.