Henry Eisenberg

Geographic pathology of thyroid carcinoma

Pathologic and epidemiologic studies of carcinoma of the thyroid were carried out in an endemic and a nonendemic goiter area. Two hundred twenty‐nine cases from the State of Connecticut, USA, and 217 from Cali, Colombia, were histologically reviewed and classified using a single nomenclature and previously agreed criteria. Tumors were classified in five categories: papillary, follicular, anaplastic, medullary, and others. A detailed pathologic description for papillary and follicular varieties was made, since most of the confusion in nomenclature seems to be due to the absence of well‐defined criteria for these two types of tumor. Incidence of specific histologic types in the two localities were also studied. A statistically significant excess of follicular carcinoma was found in Cali in comparison with Connecticut. An excess of ana‐plastic carcinoma in Cali was also seen, but the number of cases was too small to be significant. No differences were found in the incidence of papillary carcinoma in the two localities. Anaplastic and follicular carcinoma were more often associated with nodular goiter than papillary carcinoma.

EPIDERMOID CARCINOMA OF THE PERIANAL SKIN AND ANAL CANAL.

Cancer of the anus is a rare lesion. Two hundred cases registered in the Connecticut Tumor Registry are studied and treatment methods assessed. Therapeutic groin dissections are recommended for inguinal node metastases. The authors are now doing preliminary laparotomies to determine intra‐abdominal spread of the disease prior to radical groin dissection.

CANCER IN THE AGED.

The management of cancer in aged persons is a problem of increasing importance. The tumor registry maintained by the Connecticut Department of Health has been registering and following virtually every diagnosed case of cancer in that state since 1935. It is thus possible to trace the increasing magnitude of this problem and to describe trends in the management of cancer in elderly patients. The number of cancer patients who were 75 years of age or older at diagnosis increased from 1,221 during the five-year period 1935-39 to 4,995 in 1950-54; a 300 per cent increase (TABLE 1). During the same period the total number of cancer patients increased by 150 per cent, while the total population of the state increased by 50 per cent. There has been a marked shift in the age distribution of cancer patients. During the period 1935-39, 13 per cent of male cancer patients and eight per cent of female cancer patients were 75 years of age and older. Twenty years later, in the period 1955-59,* the corresponding percentages were 20 and 16 per cent. During the latter period, one-half of all male cancer patients and two-fifths of all female cancer patients were 65 years of age and older (TABLE 2 ) . Thus, in dealing with the problem of cancer case management, the clinician is confronted with the problem of treating elderly persons to an increasing extent. The organs in which cancers develop most frequently in elderly male and female patients are listed in TABLES 3 and 4. The prostate accounts for one-fourth of all cancers in males 75 years of age and older; the stomach, large intestine, and rectum account for an additional fourth. Among females in this age group, the breast accounts for 21 per cent of all cancers; the stomach, large intestine, and rectum account for an additional 28 per cent. Thus, half of all cancers in both male and female elderly patients originate in only four organs of the body. The percentage of cancer diagnoses that are microscopically confirmed provides a rough index of the overall quality of medical care. Although there is no direct relationship, there is a positive association between the percentage of cancer diagnoses that are microscopically confirmed and the survival of cancer patients. This is so because microscopic confirmation of cancer diagnoses depends on the availability of skilled surgeons and pathologists and on the extent of the disease when the patient seeks medical care. A histological specimen of the tumor is less likely to be obtained when a patient comes to a doctor in the terminal stages of cancer. There has been a marked increase in the percentage of microscopically confirmed cancer diagnoses in the 25 years under study (FIGURE 1). In the period 1935-39, 73 per cent of all registered cancer cases were confirmed. Among patients 65 to 75 years of age the percentage was 65, and among patients 75 years

Radiation therapy for cervical cancer: Late effects on life span and on leukemia incidence

To investigate life‐shortening after partial body exposure, survival of 497 patients whose primary treatment was exclusively with x‐rays and radium for squamous cell cancer of the cervix (localized and regional stages) is under continuing study. To eliminate early deaths due to cancer and to provide maximal follow‐up time, the group includes only those who survived at least 5 years and who were under age 55 at time of diagnosis. The patients were treated in 1932–1951 and were drawn from the University of California and Connecticut State Tumor Registries. Pelvic regional dose was calculated for each patient. Three dose‐groups were formed: < 23 Mgm‐rads, 23–31 Mgm‐rads, and 32–54 Mgm‐rads. When dose groups were compared with one another within stage and registry, and also after pooling stages and registries, no statistically significant differences in survival were found. When uterine cancer deaths were excluded to adjust for the possibility that unsuccessful treatment might have masked the life‐shortening effect under study, again no association was found between radiation dosage and survival. In excess of 7 cases of leukemia were predicted for the 10‐year period of 5 to 14 years after treatment. To this time, no deaths have been attributed to leukemia.

Cancer of the Bronchus and Lung: Connecticut 1935-1959.

Abstract A marked increase has been noted in the incidence of cancer of the bronchus and lung in Connecticut. The incidence of this malignancy in males has risen from 8.7 per 100,000 population in 1935, to 46.7 per 100,000 in 1959; the corresponding figures for females demonstrate a rise from 4.1 to 7.3 per 100,000 population. A detailed 22-year survival analysis by stage, age, histological type, calendar period of diagnosis, and type of treatment of patients with cancer of the bronchus and lung diagnosed in the state of Connecticut between 1935 and 1956 is presented. Observed survival is measured relative to expected survival in the general population. Among patients diagnosed during 1935–1949, only 1 3 of those who were expected to survive the first 6 months after diagnosis, did in fact, survive; less than 20 per cent of these lived 1 year and less than 5 per cent lived 5 years. Patients diagnosed during 1950–1956, experienced only slightly higher 5-year survival rates than those diagnosed in 1935–1949; the relative 10-year survival rate for lung cancer is 4 per cent for females and 2 per cent for males. There does appear to be a slightly higher survival rate for females having bronchogenic cancer. The increase in survival rates from the 1935–1949 period to the 1950–1956 period was possibly related to the increase in the percentage of patients treated by surgical intervention. This study suggests that there is an increase in the survival time of the surgically treated patients. The survival figures for the state of Connecticut are similar to the average figures obtained in other treatment centers of the United States.

The epidemiology of coronary heart disease in Middlesex County, Connecticut. I. A preliminary report on methodology and the incidence of primary myocardial infarction.

Abstract 1. 1. This report contains a description of the methodology of voluntary reporting of coronary heart disease by private physicians in a general population group in Middlesex County, Connecticut, during an eighteen-month reporting period. 2. 2. Under the most favorable conditions of cooperation and motivation of the practicing physicians and by the effective coordination of a full-time professional staff, such an approach did, as expected, produce valuable information. 3. 3. Population estimates made from the United States Bureau of the Census interview data collected on a probability sample of the population of the County made it possible to determine incidence rates for primary myocardial infarction among white males, aged 35 to 64, for the year April, 1957, through March, 1958. It was concluded that a minimal incidence rate (for myocardial infarction among white males aged 35 to 64 for Middlesex County, Connecticut) was 5.2 per 1,000 and that the actual rate was somewhat larger—approximately 6 per 1,000. 4. 4. Comparisons with similarly obtained data in North Dakota were presented. The rate in Middlesex County for cases comparable to those cases included in the North Dakota rate is significantly higher (5.4 to 3.3) when all the six-county area population in North Dakota is included and not significantly higher (5.3 to 4.4) when the farmers are excluded from the North Dakota population.

Time between pairs of leukemia cases

A study of the time separations from one case of leukemia to another in the same small geographic area revealed no clear evidence that such cases tend to occur at fixed time intervals. Such spacing of cases was not evident even when separate analyses were made by leukemia age group (childhood or adult leukemia) or by urbanization of area (urban or rural).

ABO blood groups and gastric cancer

Abstract 1. 1. A review of the foreign and domestic literature has shown a greatly increased interest in the relationship of the ABO blood groups and gastric cancer. 2. 2. A study of stomach cancer cases reported to the Connecticut Cancer Record Registry from 1950 to 1953 was undertaken. A total of 1,146 cases were found and information as to blood type was obtained on 892 of these cases. 3. 3. Data on a total of 75,904 Connecticut Red Cross blood donors was collected over a period of almost one year and used as controls. This group was drawn from the same ethnologic and geographic areas as were the cancer patients. 4. 4. The relative frequencies of blood groups O and A was shown to be inversed to a statistically significant degree for cancer patients and controls. No significant change was observed in the relative frequency of blood groups B and AB. 5. 5. Several hypotheses were presented and further investigations were suggested to study some of the possible explanations for this statistical phenomenon.

Geographic variation in the incidence of esophageal cancer

Abstract There are significant inter-area variations in the incidence of cancer of the esophagus within the State of Connecticut. Incidence rates were highest in towns of high population density, and considerably lower in areas of low population density. Within areas of similar population density no correlation with the extent or type of industrialization could be demonstrated. No consistent patterns of high or low incidence could be demonstrated in contiguous towns. Survival is poor in cancer of the esophagus and does not vary with population density.

Incidence and Mortality Rates for Leukemia and Lymphoma

T HE VARIOUS FORMS of leukemia and lymphoma are generally considered progressive and uniformly fatal. It may therefore be expected that, in any given population, incidence rates for these diseases will be close to the corresponding mortality rates. However, as shown by Shimkin (1) and others, this is not generally true. Table 1 shows the incidence rates for leukemia and lymphoma reported by two large-scale cancer surveys and one cancer register and compares them with mortality rates reported in the same areas and for the sanle years. All three incidence rates for leukemia and two of the three rates for lymphoma were substantially higher than the corresponding mortality rates. An incidence-mortality ratio of, for example, 1.50 (equivalent to a mortality-incidence ratio of 0.67) seems to imply that one-third of the patients reported to have leukemia or lymphoma are not dying of these diseases. This is not necessarily true. For chronic disease registers receiving both case reports and death certificates, there are several ways in which a person could be included in incidence data for any given disease but could be omitted from the corresponding mortality data, or vice versa.