The cancer detection rate, that is, 6.7% (42 of 623) of the higherrisk groups (IV, V), was 10 times higher than that of the lower-risk groups (1, II, 111), which was 0.67% (10 of 1053). The cancer detection rate of 29.5% of tumor stage V subjects was 25 times higher compared with that of the other tumor stage I-IV subjects, that is, 1.18% (24 of 2031). The average time from the first tumor marker combination assay to the time when clinical cancer appeared was as follows: tumor stage V, 9.4 plus or minus 14.1 months; tumor stage IV, 31.4 plus or minus 18.7 months; tumor stage III, 39.1 plus or minus 19.6 months; tumor stage 11, 12 months. The last case was reported by the patient's family, but we were unable to follow-up this patient.
Tumor Markers and Tumor Stages
The percentages of the subjects in each group who had elevated marker levels for each marker are shown by stage in Figure 3. Elevated values of CEA, HSAP, FT, FT/Fe, IAP, RNase, sialic acid, ALP isoenzyme, CA 19-9, and TPA showed a high correlation with tumor stages. However, no marker specificity for each stage was observed. These correlations were similar to those between tumor markers and tumor size in clinical cancer. 26
Tumor Staging and Risk Assessments
Table 4 shows the cancer incidence of each stage within 0.5-2.0 years after performing the tumor marker combination assay. The rate of incidence of cancer for Stage V was 23.2% (22 of 95). Three of 528 Stage IV subjects (0.57% [3 of 528]) and none of the subjects at the other stages were diagnosed as having cancer. After 5-7 years, the detection rate at tumor stages I, II, III, IV, and V was 0%, 0.4%, 0.7%, 2.7%, and 29.5%, respectively (Table 4). A comparison of the cancer incidence between 5 and 7 years and within 0.5-2.0 years revealed a high correlation between the tumor staging and the risk assessment. The cancer incidence rate from Stage II to Stage V is intimately correlated with risk assessments from low risk (0.4%) to high risk (29.5%). In other words, tumor staging in preclinical cancer can be regarded as being equivalent to risk assessment.
TUMOR STAGES
Figure 3. Percentage of subjects at each stage who had elevated marker levels. (A) CEA: 0 3.5-4.3, 2-4.4 ng/ml. (B) HSAP: 01020, z20 U/L. (C) FT: males, 250-299, females, 150-199; a males 2300, females, 2200 ng/ml. (D) FT/Fe: 0 2-4.99, a Z5. (E) IAP: 0 450-499; 0,--500 rig/ml. (F) RNase, 0110-144, 0 2145 U. (G) Sialic acid: 0 60-65, 0 >-66 mg/dl. (H) ALP,: 01-4; 0 Z5%. (1) CA 19-9: 0 33-36, s a37 U/ml. 0) TPA: 0 85-109, 0 2110 U/l.Discussion
Generally, the method of sampling subjects for screening studies is extremely important so as to avoid biased results. The age distribution in the current study was the same as in conventional screening (Table 2), and the distribution of screenees in each tumor stage by geographic area was similar to the total distribution of total screenees, indicating that our sampling was relatively unbiased. The cancers tath were detected according to each tumor stage were varied. The rate of cancer incidence in various organs was similar to that of general epidemiologic data in Japan. Our tumor marker combination assay, therefore, promises to be applicable to early cancer detection in general.