Tissue polypeptide-specific antigen (TPS) is a marker of cytokeratin 18 ( I ) . Almost 40 years ago Bjorklund & Bjorklund (2) described an insoluble, heat-labile tumour antigen, tissue polypeptide antigen (TPA). Recently, mapping of TPA has revealed that it contains 35 different epitopes (3). One of these, M3, is localized to cytokeratin 18 and is mainly released during malignant cell mitosis. The TPS turnour marker assay operates on the basis of a monoclonal antibody raised against M3 (4). Measurement of TPS in serum may. therefore, give useful information about the proliferation rate of tumour cells and TPS is a prognostic parameter in patients with advanced gastrointestinal cancer (5) and in breast cancer patients (6). In the first mentioned group of patients serial TPS measurements identified patients who will not benefit from the planned treatment. CA 125 is an antigenic determinant on a high-molecular weight glycoprotein which can be recognized by a monoclonal antibody, OC125. CA 125 is apparently at present the best marker for identifying changes in the tumour mass of patients with epithelial ovarian cancer. The use of TPS as a supplement to CA 125 may provide additional information about the proliferation rate of residual tumour cells during treatment and control. The preliminary study reported here is aimed at measuring and comparing the clinical usefulness of information obtained by TPS and CA 125 in two groups of ovarian cancer patients. Material and merhods. Two separate groups of patients were investigated: a second-look group and a follow-up group. All patients underwent post-therapeutic exploratory laparotomy (second-look operation), including microscopic examination of biopsies from the operation area, scrapings from the abdominal diaphragm, and liquid from peritoneal lavage. The second-look group (n= 20) included 12 serous, 4 endometrioid, 3 undifferentiated, and 1 mucinons ovarian carcinoma(s). Serum samples for TPS and CA 125 analyses were drawn 1 and 7 days after the last course of chemotherapy in 2 patients and in 18 patients the samples were drawn median 28 days (range: 26-44 d) after the end of therapy. Serum was drawn median 21 days (range: 11-49 d) prior to the second-look operation. The follow-up group included 15 patients (5 serous, 6 endometrioid, and 4 undifferentiated ovarian carcinomas) from whom a total of 125 serum samples were drawn. All patients had complete pathologic response (CPR) at the preceding second-look laparotomy. TPS and CA 125 were scheduled to be measured every second month during the follow-up period. The result of the marker analyses was compared with the result of a gynaecologic examination scheduled to be performed every third month during the first year of control and twice during the second year. Serum samples for tumour marker analyses were collected in the departments of oncology at the university hospitals of Herlev and Aarhus from August 1985 to May 1990. TPS was measured by an enzyme immunoassay (BEKI TPSTM-EIA monoclonal; BEKI Diagnostics AB, Bromma, Sweden) according to the manufacturer’s specifications. In the present study TPS values above 80 U/l (95% percentile of TPS levels in 195 completely healthy test subjects, TPSTM-User Information) were considered to announce dividing tumour cells. Serum CA 125 was measured by a commercially available enzyme immunoassay (Abbott CA 125TM-EIA Monoclonal, Abbott Laboratories, Chicago) as previously described (7). CA 125 values above 35 U/ml were considered abnormal. All measurements were performed in duplicate. Serum was stored in plastic cryotubes at -80°C until analysis. At this temperature the stability of the TPS and CA 125 antigen is at least 3 years. Results. The second-look group (Table I ) . TPS was elevated in 7 (35%) patients of whom 3 were free of tumour and 4 had macroscopic disease. Thirteen patients had normal TPS values; 5 had no tumour, 5 had microscopic and 3 macroscopic residual disease. CA 125 was above the cut-off level in 3 (15%) patients; one had microscopic and 2 macroscopic tumour. The last 2 patients also had increased TPS values. Seventeen patients had normal CA 125, 8 had no tumour, 4 had microscopic and 5 macroscopic tumour. The fol1ou-up group (Table 2). Six patients remained free of symptoms during the observation period. Marker values were measured for 19-25 months (median: 22 months) and clinical examinations were performed for 19-60 months (median: 51 months) after the second-look operation. TPS was increased in three patients and was within normal levels (14-74 Ujl) in another three. All 6 patients had normal CA 125 values ( 2 1 8 Ujml). Nine patients experienced recurrent disease 4-30 months after the second-look operation. Five of these patients had increased TPS values. In 1 patient the TPS values constantly lay above 80 U/l and the first increase was found 10 months before relapse was established by clinical diagnosis. The antigen level fluctuated