The Metastatic Spread of Cancers of the Digestive System in Man

Abstract
This investigation was based on the analysis of 58 autopsy records of patients with pancreas adenocarcinomas, 143 patients with stomach cancer, 165 patients with colon cancer and 150 patients with rectal cancer, respectively. These records belong to a series of 4,728 autopsy records collected by the Department of Pathology at Roswell Park Memorial Institute from 1956 through 1965. The primary purpose of this paper was to elucidate how metastases spread in the human body, i. e., whether they spread directly from the primary adenocarcinoma throughout the body or by a cascade or multistep process. The basic definition of the cascade spread is that the previous seeding of one or two key organs is required for the disease to progress from the primary malignant tumor throughout the body. It was found, in the absence of liver metastases, that the lungs could either be seeded directly by the primary tumor with no metastases both in the abdominal and pelvic lymph nodes (blood-borne metastases) or that cancer cells can reach the lungs either from the liver (blood-borne metastases) or from the pelvic and abdominal lymph nodes, (in the absence of liver metastases), presumably by invasion of the "cysterna chyli". It was found also that extra-pelvic and extra-abdominal organs are overwhelmingly seeded either by lung metastases or by liver metastases via the lungs. This remark holds true for the thyroid and pituitary glands, the central nervous system, the vertebrae and all other bones. For organs located in the abdomen and pelvis, it appeared that either the lungs alone or the liver through the lungs disseminate metastases to both the abdomen and pelvis. This remark holds particularly true for metastases in the kidneys, adrenal glands and spleen. It was also suggested that other organs, such as the ureters, uterus, pancreas, ovaries and extra-hepatic biliary tract were also seeded by the lungs or by the liver and lungs but these two routes seemed to be less important than they were in the extra-abdominal and extra-pelvic organs. Sites totally enfolded by the peritoneum showed two different metastatic patterns. In the absence of liver metastases no disseminating route was detected between the lungs, and both the small and large intestines, whereby suggesting that both the lungs and the small and large intestines are seeded independently of each other by the primary cancer, very likely via the pelvic and abdominal lymph nodes. With a more advanced disease (presence of liver metastases), it was suggested that the cancer cells may reach the abovementioned organs via the blood-stream (origin of metastases in the liver and lungs). Probabilities were calculated to show how often specific sites could be seeded either directly by the primary cancer or seeded by the lungs alone or by the liver and lungs together. It was found that very often metastases in extra-abdominal and extra-pelvic organs originated both in the liver and lungs. Odds ratios were calculated to show how much more frequent metastases were in specific organs, given that both the lungs and liver, or lungs alone, were seeded relative to the chance that the same organs were seeded, when the lungs were metastases-free.