One of the ultimate aims of our research has been to try to utilize in cancer the knowledge obtained from investigations of the general problems of pathology and therapy. Encouraged by the results of biologically guided therapy in many other conditions, we have applied it to the treatment of malignancy.

As we have mentioned before, differences between animal cancers, both experimental and spontaneous, and human cancers represent one reason why an agent, however good its results in animals, may not apply to human malignancy. Another factor, conduct of treatment, is no less important. The main characteristic of our therapeutic approach resides in the fact that treatment is continuously guided by data representing the actual condition of the subject. At least for the moment, it appears impossible to recognize, through suitable tests, the patterns present in animals so as to apply them to guided therapy. Therefore, we have been obliged to do our therapeutic research in humans, reserving animal studies for limited problems. This situation had led us to emphasize, always, the experimental nature of our therapeutic efforts in humans. Although we started with desperate terminal cases, frank immediate subjective and objective benefits, even though temporary, were obtained frequently enough to encourage us to go on. Together with the above mentioned considerations, they seemed to justify the continuation of therapeutic research in human patients. We will try to review as objectively as possible the results obtained with therapeutic methods and agents evolved over the years.

In 1927, a 33-year old woman with typical preterminal cancer of the stomach came under our care. In the highly emaciated patient, a hard, irregular mass filling the entire epigaster was palpable. Radiological examination indicated a prepyloric gastric tumor. Laparotomy showed an inoperable tumor of the stomach, with the omentum and lymphatic glands greatly involved and evidence of multiple peritoneal and liver metastases. In view of the patient's general condition and the fact that the pylorus was only partially obstructed, no surgical procedure was performed other than biopsy of one of the metastases in the omentum. The biopsy showed an adenocarcinoma Grade III of gastric origin. Treatment was not prescribed.

I saw the patient two years later in apparently good health. Clinical and radiological examination at that time showed no tumor. The patient attested to receiving no treatment. At the time of the operation she had been two months' pregnant. We had attributed her amenorrhea at that time to the advanced cachectic condition. She had given birth at term to a normal girl. Hers was one of those cases usually catalogued as "spontaneous remission."

Since then, I have analyzed many of the published observations of cases of so called "spontaneous remission" of cancer always to find a turning point that coincided with the intervention of some event usually considered to have no possible significance for malignancy. The fact that such events have not induced similar changes in other cancer patients has made them seem unimportant to many investigators.

While not regarding them as the only cause of favorable changes, we have not eliminated the possibility that such events may have a contributory role. We must recognize that, if such events in themselves appear to be powerless to change the course of cancer, they may intervene in conjunction with, and potentiate, another factor also powerless in itself to induce a change.

It was with this concept in mind that we reviewed the case of the woman with stomach cancer. We considered the possible effects of two factors which apparently intervened concomitantly: pregnancy and surgery. We then began a series of experiments.

Ehrlich mammary carcinoma was grafted in two groups of female mice, one pregnant and the other not. In each group, half of the mice were kept as controls while the other half was submitted to a sham surgical procedure consisting of a laparotomy in which multiple ligatures were performed. Growth of the tumors and survival times were noted. Compared with non operated, nonpregnant mice serving as controls, both the pregnant mice and the surgically treated mice showed a slowing down in the evolution of cancer lesions. It was in the group of mice, both pregnant and surgically treated, that a temporary arrest in tumor evolution was seen. In some animals even temporary regression was noted; in 1/20 the tumor regressed entirely.