growth-related tumor markers, respectively, we thought that they might play an important role in the tumor marker diagnosis of cancer. According to our model, in cancer development, the growth-related tumor marker appears first, followed by associated tumor markers and specific tumor markers. The presence of these three substances and the order of their appearance are of major significance in cancer screening.

Thus, we can classify the tumor stages and use them to make a risk assessment. Our system of staging the tumor development (and the corresponding developmental stage of the actual tumor) is based on the cutoff values of the tumor marker responses, as described below.

With our tumor marker combination assay, we can continuously follow the time course of cancer development and discover not only early cancer in a defined organ, but also early cancer at any site. Thus, by identifying a high-risk group, we can decrease the number of patients with cancer by implementing a cancer prevention program, thus achieving true cancer prevention.7-10

In the current paper, we report the detection rate obtained in each developmental stage (from tumor stage I to stage V) and the relationship with risk assessment as well as the question of the reasonableness of our proposed model of natural history of preclinical cancer and its tumor stage classification.

Subjects and Methods

Subjects Screened

From 1984 to 1986, we performed mass screening for cancer using a combination assay of tumor markers at 15 randomly chosen defined geographic areas in Japan, with a total of 2126 randomly chosen residents (1091 males and 1035 females). Geographic areas, environmental conditions, and the number of screenees are shown in Table 1. The subjects had already been mailed a medical history form requesting such clinical information as past illnesses, present symptoms, and pregnancy. The completed form was handed in on the day of the screening. The subjects were presumably healthy, mostly between the ages of 23 and 80 years, and asymptomatic for cancer (Table 2).

Approximately 18 ml of blood was taken from each subject. Serum was separated from the blood by centrifuging and stored at -20°C until the assay.

Combination Assay of Tumor Markers

The serum samples were used for a tumor marker combination assay. The following tumor markers were assayed simultaneously.

Carcinoembryonic antigen Carcinoembryonic antigen (CEA) was assayed by enzyme immunoassay using a CEA-EIA Kit (Hoffmann-Roche Co., Ltd., Basel, Switzerland).

Heat-stable alkaline phosphatase. Heat-stable alkaline phosphatase (HSAP) was first reported by Fishman et al. as the Regan isoenzyme.4 It has been found in tumor tissue and in the serum of patients with various cancers. HSAP determinations were performed by a modification of the method described by Maslow et al." After 50 ul serum had been heat-treated at 65°C for 7 minutes, it was allowed to react with a fluorescent substrate (naphthol AS-MX phosphate) for 20 minutes at 37°C; it was deproteinized with acetone and the intensity of the fluorescence in the supernatant was measured with a fluoroessence spectrophotometer (Hitachi 650-10, Tokyo, Japan).

Ferritin. Ferritin (FT) was assayed by radioimmunoassay using an RIA-gnost Ferritin Kit (Hoechst Behringswerke Aktion Geselshaft, Marlburg, Germany). From clinical experience, we decided to use the ratio of FT to serum iron (FT/Fe) as another tumor marker.

Immunosuppressive acidic protein. Immunosuppressive acidic protein (IAP) is characterized by inhibiting both phytohemagglutinin-induced lymphocyte blast formation and mixed lymphocyte reaction in vitro." IAP was determined by means of single radial immunodiffusion using an IAP plate kit (Sanko Jyunyaku Co., Ltd., Tokyo, Japan). Five microliters of sample were applied in each well of an agarose gel plate containing anti-IAP serum, and after incubation for 48 hours at 37°C, the diameter of the precipitation ring was measured.