The sequence of events in haematogenous metastasis from colonic carcinoma was analysed, using 1541 necropsy reports from 16 centres. The findings are consistent with the cascade hypothesis that metastases develop in discrete steps, first in the liver, next in the lungs and finally, in other sites. Deviations of necropsy findings from the cascade model are largely explained on the basis of false negative reports. In only 216 of 1194 cases was there suggestive evidence that metastatic patterns (excluding lymph nodes) were causally related to lymphatic or non-haematogenous pathways. The incidence of metastatic involvement in 'other' (quaternary) sites correlated with target organ blood-flow (ml min-) per g, only when bone marrow and thyroid were excluded. In the thyroid the incidence was lower than expected on the basis of blood flow per g tissue; this may indicate that the thyroid is an unfavourable site for metastatic growth of colonic carcinoma. In the bone marrow it is higher; the latter may be due to delivery of cancer cells via both arterial blood and the vertebral venous plexus. Recognition of this pattern of metastases in the bone marrow could be important with respect of diagnosis and therapy, in patients with colonic carcinoma.

No abstract

The metastatic behaviour of renal cell carcinoma has been studied in a series of 687 necropsies. The observations were consistent with the concept of "metastatic inefficiency", in that in 295 cases, including 25 with renal vein invasion, there were no detectable metastases. In the present series, renal vein involvement was not an important prognostic factor in stage 1 or 2 disease. In 73% of cases without lung metastases there were none in other sites, and in 84% of those with lung metastases there were others elsewhere, consistent with a metastatic "cascade" in which metastases first developed in the lungs and were later detected in other organs. However, the observations did not permit discrimination between anatomic cascades, in which other organs were seeded from metastasizing pulmonary metastases, and temporal cascades, in which the other were seeded at the same time as the lungs, but with fewer cancer cells. The patterns of arterial metastasis were consistent with the "seed-and-soil" hypothesis, and a novel index was developed to quantify differential organ "soils". The contralateral kidney was not the best soil for metastases from renal carcinoma. Given the presence of lymph node metastasis, the probability of heamatogenous metastasis is 90%. However, in the absence of nodal metastasis, approximately half the cases had haematogenous metastasis.

Since the carcinomas of the cardia and the adenocarcinomas of the esophagus show many similarities in their histological and morphological descriptions, a detailed comparative study was attempted on the basis of 66 esophageal carcinomas in adenoid differentiation, 359 carcinomas of the cardia, 1288 gastric carcinomas in infracardial localisation, and 492 squamous carcinomas of the esophagus. The evaluation yielded no significant differences between the adenocarcinomas of the esophagus and the cardia neither in age and sex distribution nor with regard to the classifications of Borrmann, WHO, Ming, and Laurén, but a significant discrimination was possible between esophageal and cardial adenocarcinoma together, on the one hand, and infracardial gastric carcinoma on the other. Furthermore, esophageal adenocarcinomas were localized preferentially in the lower third, unlike squamous carcinomas of the same organ. These results suggest that esophageal adenocarcinoma and carcinoma of the cardia must be considered as one separate entity, probably originating from a common stem cell. They further suggest that the cardia belongs to the esophagus rather than to the stomach.

The current study was carried out on 88 colorectal carcinomas to assess the degree of intratumor heterogeneity as reflected by multiple aneuploid DNA stemlines and their relation to tumor stage and morphologic differentiation. Each tumor was segregated into an average of nine specimens (3-15), which were analyzed separately. DNA aneuploidies were identified in 72 cases (82%), 29 revealing multiple aneuploid DNA stemlines with up to four aneuploid subpopulations. In 10 of the 29 carcinomas with DNA stemline heterogeneity, a ratio of 2:1 was calculated from the different DNA indices, possibly indicating that the additional DNA stemline emerged from the first one by doubling its DNA content. No correlation was found between the overall frequency of DNA aneuploidies or heterogeneous DNA stemlines and the tumor stage according to Dukes' staging. Well-differentiated carcinomas tended to express aneuploid DNA stemlines more frequently than moderately or poorly differentiated tumors, although the morphologic intratumor heterogeneity did not correspond to the appearance of multiple aneuploid DNA stemlines. These data indicate a high degree of intraneoplastic diversity in colorectal cancer and emphasize the usefulness of DNA analyses for the quantitative assessment of tumor heterogeneity.