Background. There is no reliable method for the early diagnosis of ova
rian cancer. Radiolabeled monoclonal antibodies have potential to assi
st in early diagnosis, but they are limited by problems that include a
ntibody specificity, stability, and immunoreactivity, as well as patie
nt reactions to the antibodies used. Methods. Methods were developed t
o Au-198-label a human monoclonal antibody (TC5 antibody), developed a
gainst an ovarian cancer cell surface antigen. Antigen binding sites o
n the TC5 antibody were protected with sepharose 4B affinity chromatog
raphy before Au-198-labeling. The Au-198-labeled TC5 antibody was eval
uated with biopsy specimens in a blind study. The immunoreactivity of
radiolabeled TC5 antibody also was evaluated in slot-blot experiments
with extracts of the biopsy specimens. Results. The Au-198-labeled TC5
antibody had high binding reaction to all biopsy specimens (six of si
x) pathologically diagnosed as ovarian cancer (serous and endometrioid
adenocarcinoma). The radiolabeled TC5 antibody did not bind to any no
rmal (non-neoplastic) specimens (zero in ten), with one exception. One
''normal'' ovary specimen had high binding of radiolabeled TC5 antibo
dy, and metastatic ovarian cancer was diagnosed 4 months later. The TC
5 antibody labeled with Au-198, without protecting antigen-binding sit
es, did not bind to any biopsy specimens. Conclusions. The affinity-la
beling method was necessary to protect antigen-binding sites and prese
rve the immunoreactivity of the TC5 antibody. The Au-198-labeling meth
od may be an ideal technique to evaluate monoclonal antibodies in vitr
o. The TC5 antibody had high sensitivity and specificity for detecting
ovarian cancer.