The stage of a cancer at diagnosis is multiply determined, theoretically depending on such disparate factors as tumor biology, patient education, and physician behaviors. Data from all Hispanic and non-Hispanic white residents of New Mexico diagnosed with a malignancy from 1969 through 1982 were analyzed to determine the relationship between the age of the patient and the stage of cancer at the time of diagnosis. Three general patterns were apparent. For cancers of the bladder, breast, cervix, ovary, thyroid, and uterus, and for melanoma, there were significant linear trends for the cancers to be diagnosed at more advanced stages in older patients. For cancers of the lung, pancreas, rectum, and stomach, there were significant linear trends for cancers to be diagnosed at an earlier stage in older patients. For cancers of the colon, kidney, liver, and prostate, there were no significant linear trends in stage at diagnosis versus age of the patient.

We performed a population-based case-control study to examine the association between uranium mining and lung cancer in Navajo men, a predominantly nonsmoking population. The 32 cases included all those occurring among Navajo men between 1969 and 1982, as ascertained by the New Mexico Tumor Registry. For each case in a Navajo man, two controls with nonrespiratory cancer were selected. Of the 32 Navajo patients, 72 per cent had been employed as uranium miners, whereas no controls had documented experience in this industry. The lower 95 per cent confidence limit for the relative risk of lung cancer associated with uranium mining was 14.4. Information on cigarette smoking was available for 21 of the 23 affected uranium miners; eight were nonsmokers and median consumption by the remainder was one to three cigarettes daily. These results demonstrate that in a rural nonsmoking population most of the lung cancer may be attributable to one hazardous occupation.

Cystosarcoma phyllodes is a very rare tumor which maybe difficult to diagnose clinically. The epidemiology and pathobiology are different from those of breast carcinoma. Risk factors, multicentricity, bilaterality, as associated with breast carcinoma, are not observed in patients with cystosarcoma phyllodes. Although the term "sarcoma" indicates a malignant tumor, only 10%-30% of cystosarcomas are histologically diagnosed as malignant; clinical diagnosis of malignancy does not exceed 10%. Axillary node involvement is rare, but hematogenous spread of cystosarcoma occurs into lung, pleura, bone, and liver. Clinically, cystosarcoma is a large (usually 3-5 cm in diameter) painless tumor with sudden growth acceleration especially during pregnancy. Cystosarcoma is usually circumscribed, containing firm and soft areas. The differential diagnosis has to include fibroadenoma, fibrocystic disease, mastitis, abscess, and medullary carcinoma. Neither clinical, mammographic or sonographic signs exist to predict a benign or malignant tumor. Therapy of cystosarcoma is not uniformly agreed upon. Radical, modified-radical, and simple mastectomy and tumorectomy are typical treatments; therapeutic results are the same for each treatment modality. For histologically diagnosed malignant cystosarcoma, the relative 5-year survival rate is about 80%. Clinically, malignant metastatic cystosarcoma is incurable; radiotherapy, endocrine treatment, and polychemotherapy are all ineffective. Because of the specific tumor pathobiology of cystosarcoma and its rarity, evaluation of treatment modalities and comparison of survival rates are difficult.

To determine the effect of ethnic group on respiratory disease occurrence, average annual sex, ethnic, and disease specific mortality rates for the period of 1969 to 1977 were calculated for New Mexico's American Indian, Hispanic, and Anglo populations. Incidence data were available for respiratory tract cancer. This study corroborates previous findings of reduced mortality from lung cancer in American Indians of both sexes and in Hispanic males.

A review of the population-based New Mexico Tumor Registry data identified 446 patients with nonsimultaneous multiple primary cancers, excluding non-melanoma skin cancers and carcinomas in situ of the uterine cervix. Expected numbers of cases were established by observing the person-years of exposure to the risk of developing a second or subsequent primary cancer and then applying the appropriate locally determined age-, sex-, ethnic-, and site-specific cancer incidence rates. The relative risk (observed/expected) of developing a second primary cancer was elevated for "Anglo" and Spanish American cancer patients in comparison with the risk of developing a first primary cancer in persons who have never had one. Only six cases of nonsimultaneous multiple primary cancer were observed (6.39 expected) in the region's American Indian population. There were differences in site-site associations among the three ethnic groups, but in many categories there were too few cases for analysis.