Cancer family history was studied for 43,117 subjects aged 35-79 at the occasion of stomach cancer mass screening and were followed up for 5 years. 376 cases of cancer were notified out of them by the linkage with records of population based cancer registry covering the prefecture. Significantly elevated risk among those with family history was found only for breast cancer (observed 3, expected 0.82).

During the last 6 years, we examined 402 cases (including 5 cases in their 20's and 17 cases in their 30's) of colorectal cancer. We compared three groups (cases in their 20's, cases in their 30's, and cases over 40 years old), and had the following results. Cases in their 20's had a higher rate of familial history of colorectal cancer, and the cancer was located mainly in the proximal colon. Well differentiated and moderately differentiated adenocarcinoma was dominant in the cases in their 30's and in those over 40 years old. On the other hand mucinous adenocarcinoma and signet ring cell carcinoma was dominant in the cases in their 20's. A case of cancer family syndrome is also presented.

Cell-mediated immunity of the aging and immune response in the family of gastric cancer have been investigated and the following results were obtained. Immune response particularly of cell-mediated immunity was found decreased in paralled with aging; decreases in peripheral lymphocyte and T-cell counts, rate of Con A and PHA induced blastoid transformation, and the tendency of increase in suppressor T-cell were noted. Investigations on the immune response in the family of gastric cancer revealed that cell-mediated immune response was recreased in the compared to age-matched healthy volunteers. These findings suggest that latent immunological failure may exist in the above family.

Chromosome instability is considered to be a probable candidate for a genetic predisposition to cancer. In order to confirm association of heritable fragile sites with cancer, an epidemiologic survey study was conducted which compared their incidence between patients of leukemia and allied diseases, and healthy subjects. The total incidence was 3.2% and 6.0% in patients and controls, which seemed to indicate that the carrier state of a fragile site is not a risk factor for cancer development. However the cases detected in the study were individually quite interesting: One was a coincident case of fra (16) (q22) and inversion of chromosome 16, and another was a coincidence of homozygous fra (17) (p12) and familial clustering of cancers. Chromosome instability, including the fragile sites, should be paid further attentions with respect to its role in the etiology of cancer.

Recent studies of genetic epidemiology of retinoblastoma are reviewed, and major conclusions drawn are as follows: Retinoblastoma occurs in heritable or nonheritable form. The primary genetic change in the heritable form is a point mutation or deletion at a locus on 13 q 14. For the genesis of a tumor in either form, the loss or inactivation of both alleles at this locus is a prerequisite, although additional events may be needed for its development. Information about environmental risk factors for the occurrence of heritable or nonheritable retinoblastoma is still meager. Although viral etiology for nonheritable tumor cannot be excluded, an extensive epidemiologic study revealed no seasonal variation in the births of 675 patients with sporadic unilateral cases, a great majority of which may be regarded as due to somatic mutation. Moreover, there was no paternal age effect at all on the occurrence of 225 sporadic bilateral cases. Paternal exposure to ionizing radiation or chemical mutagens, which should have an accumulated effect with advancing age, does not seem to play a major role in the production of germinal mutation at the Rb locus. Family studies show that host resistance genes at other loci can modify the process of tumor development when the primary genetic change is already present in all the target cells. Unaffected gene carriers may be regarded as inherently resistant to tumor formation, whereas persons who present early onset of bilateral tumors are the most susceptible. In certain families, however, non-expression in the carriers may be due to chromosomal rearrangement in a balanced state.

During the past decade enormous progress has been achieved in the investigation of cytogenetics and moleculargenetics of leukemia and lymphoma, although there have been few systemic studies which clarify genetic etiology leukemia. At this standpoint it is important to review the recent literature which concerns etiology of leukemia. Recently it has been proved that of the incidence of childhood acute lymphocytic leukemia is same in all over the world by Greaves and his collegue. So the major cause of childhood leukemia can be explained to be spontaneous mutation. Nevertheless genetic factors are thought to be a part of another cause of leukemia as well as certain environmental factors. Heterogeneity of genetic mechanism in familial aggregation in each case probably makes difficult to analyse the cause of hereditary mechanism in leukemia.

Previous reports on twin cancer have been reviewed. Especially, cancer in later adulthood has been focused. There have been done several large scale twin studies on cancer in Europe and the U.S.A., but not in Japan. Although, it is not easy to establish twin registry in Japan, the importance of twin study in epidemiology and preventive medicine has been emphasized. Methodology of twin research in epidemiology has been described. Epidemiological research methods in Japan have been described through the authors' past experience.

The basic mechanisms of oncogene activation are gene amplification, increased expression of a single copy gene and somatic mutation, resulting in a genetic product with increased oncogenic potential. Gene amplification appears to be a more common mechanism of oncogene overexpression and its clinical significance is beginning to be appreciated. In human neuroblastomas, N-myc amplification is highly correlated with advanced stages of disease and rapid disease progression, suggesting that the N-myc gene plays a crucial part in determining the degree of malignancy of neuroblastomas. Chromosome aberrations such as double minutes and homogeneously staining region which observed frequently in neuroblastomas are the cytological manifestation of N-myc amplification and its surrounding DNA fragments.

Association with family history of cancer of each site was examined for each type of childhood malignancies using data of National Childhood Malignancy Registry in Japan as materials (n = 16,555). 2,926 cases were with cancer family history. Family history of same type of malignancy was found significantly in excess for leukemia, a malignant lymphoma, brain tumor, neuroblastoma and retinoblastoma. When observed by cell type, association with family history of leukemia was most striking in acute myeloid leukemia. Median age at first diagnosis of retinoblastoma was 11 months earlier when family history of retinoblastoma existed. Family history of leukemia and history of exposure to prenatal radiation exposure was found to enhance relative risk for childhood leukemia when combined, suggesting existence of genetic environment interaction. Mode of interaction was interpreted as multiplicative.

Hereditary large bowel cancer is roughly divided into the hereditary gastrointestinal polyposis (HGIP) that is including adenomatosis coli (AC), Peutz-Jeghers syndrome and juvenile polyposis, and non-polyposis familial large bowel cancer (NPF) which is including cancer family syndrome and its related conditions. In the HGIP, AC is most frequent and of highest risk of large bowel cancer, and epidemiological method of approach to patients has been well studied. NPF, however, has been insufficiently studied because of lacking of distinctive marker as seen in the polyposis in spite of that it may be actually present much more frequently than AC. In order to extract NPF from the general large bowel cancer group, the marker or criteria for it has been studied. The factors besides family history presently considered to be useful for detecting NPF are multiple cancers, proximal cancer, association of multiple adenoma and young (50 years or younger) onset of the disease.

Chromosome alterations, which are directly visible changes in the DNA, have close associations to cancer development, non-specific ageing, and heritable genetic status. Human lymphocyte cultures can be used for cytogenetic monitoring of genetic health because many cancers and genetic effects are caused by long-term unhealthy life-styles. We have investigated the sensitivities of lymphocytes from inherited-cancer-prone diseases to the induction of the chromosome alterations by mutagens and carcinogens, and the correlations between the frequency of sister chromatid exchanges (SCEs) in peripheral lymphocytes and life-styles or daily ways of living. Lymphocytes from patients with Down syndrome, Fanconi anemia, xeroderma pigmentosum, ataxia telangiectasia, and Bloom syndromes showed altered (usually enhanced) susceptibilities to the induction of chromosome aberrations and SCEs by mutagens and carcinogens in our environments. Mean frequencies of baseline SCEs in lymphocytes from normal men with poor life-styles have also been shown to be significantly higher than those in cells from men having good life-styles. The former cells have further been shown to have hyper sensitivities to the induction of SCEs by mitomycin-C' treatment compared to latter cells. Unhealthy life-styles also make the lymphocytes to be more sensitive to ara-C's enhancement of radiation-induced chromosome aberrations.

Familial clustering of cancer causes a problem whether significant differences exist from one person to the next in cancer susceptibility. Such clustering could occur simply by the chance occurrence of common environmental exposure, but in rare families, clustering is caused by discrete genetic factors, some being recessive and some being dominant inheritance with mendelian segregation. In this review article, heritable cancers are listed and discussed. Recent laboratory developments, such as the study on chromosome fragile sites and restriction enzyme-fragment length-polymorphism, to find the predisposition for cancer was introduced. Other analysis for individual difference by the genetic epidemiology was also summarized.

MEN-2 is a syndrome which is characterized by association of medullary thyroid carcinoma with pheochromocytoma. This syndrome is transmitted in an autosomal dominant fashion with very high penetrance. Linkage analysis of the kindreds at high risk and comparison of leukocytes- and tumor-DNAs of the patients using restriction fragment length polymorphisms (RFLPs) will be useful to find out the locus of the gene which predisposes to the syndrome. The environment factors play little role in the tumorigenesis in MEN-2.

Sister chromatid exchanges (SCEs) induced by four anticancer drugs, 3-(4-amino-2-methyl-5-pyrimidyl) methyl-1-(2-chloroethyl)-1-nitrosourea (ACNU), 1-3-bis (2-chloroethyl)-1-nitrosourea (BCNU), nitrogen mustard (HN2), cis-diamminedichloroplatinum (II) (cis-Pt) were examined on five cell lines derived from human malignant glioma biopsy specimens, and compared to results obtained with colony-forming efficiency (CFE) assay. Treatment of the five cell lines with these four drugs produced concentration-dependent increases in SCEs. Treatment with ACNU induced the most SCEs in SF-126 cells decreasing in SF-268 cells followed by SF-210 cells, SF-295 cells, and the least SCEs in SF-188 cells. The results of the SCE assay with BCNU in these cell lines were similar with ones with ACNU. In contrast to results obtained with nitrosoureas, the most SCEs were induced in SF-188 cells and the least were induced in SF-126 cells by the treatment of HN2. The frequency of SCEs induced with cis-Pt was almost similar in the five cell lines. The number of SCEs induced by the treatment of ACNU, BCNU, HN2 and cis-Pt in five cells lines showed a good correlation with cytotoxicity measured by CFE assays, and induction of SCEs occurred at much lower concentrations of these anticancer drugs than those required to induced cell kill. These results suggest that measurement of induced SCEs in human brain tumor cells treated with some anticancer drugs provide a more sensitive indicator of drug action than CFE assay and that SCE assays may be a useful method of the in vitro sensitivity test to some anticancer drugs.