Geoffrey Keynes Research on Cancer Treatment

In God We Trust. All Others Must Have Data

Radical surgery had undergone an astonishing boom in the 1950s and 1960s. William Halsted had become the patron saint of cancer surgery in the United States. But at St. Bartholomew’s Hospital in London, a doctor named Geoffrey Keynes was not convinced.

In August 1924, Keynes examined a patient with breast cancer. Rather than reaching indiscriminately for a radical procedure, he opted for a much more conservative strategy. He buried fifty milligrams of radium in her breast to irradiate her tumor and monitored her to observe the effect. Surprisingly, he found a marked improvement. Her tumor had reduced so rapidly that Keynes might be able to remove it with a minor surgery.

Over the next five years, Keynes tried other variations of the same strategy. The most successful variation was to remove the tumors with a minor surgery, followed by a small dose of radiation to the breast. Nothing was radical, yet their cancer relapse rate was comparable to those got by using radical surgery in Baltimore and New York. In 1927, Keynes reviewed his experience combining local surgery with radiation in a technical report to his department. But his theory and operation were ignored by American surgeons. They called Keynes’s surgery “lumpectomy.”

In 1953, a colleague of Keynes’s gave a lecture on the history of breast cancer at the Cleveland Clinic in Ohio, focusing on Keynes’s observations on minimal surgery for the breast. In the audience was a young surgeon named George Barney Crile. Crile had learned the radical mastectomy from students of Halsted. But he was having his own doubts about radical mastectomy. As Crile poured through Keynes’s data, the flaw in the logic of radical surgery came to light. If the breast cancer was locally confined, then it could be cured by a small local surgery. Radical surgery could add no benefit. If the tumor had already spread outside the breast, then even the most exhaustive surgery would be useless.

Crile gave up on radical mastectomy and treated breast cancer using an approach similar to Keynes’s. Over six years, he found that the effect of his “simple mastectomy” was remarkably similar to Keynes’s, with patient survival rates similar to those got from radical mastectomy.

A Pennsylvania surgeon named Bernard Fisher had also lost faith in radical mastectomy. In 1971, Fisher organized a clinical trial through the NSABP – National Surgical Adjuvant Breast and Bowel Project – to test the efficacy of radical mastectomy against lumpectomy+radiation and simple mastectomy. It took Fisher 10 years to gather that data. 1,765 patients from 34 centers in the United States and Canada enrolled in the trial. Patients were randomized into three groups: one treated with simple mastectomy, the second with lumpectomy followed by radiation, and the third with radical Mastectomy. The results of the trial were made public in 1981. The rates of breast cancer relapse, death, and metastasis were statistically identical among all three groups.

Radical mastectomy is rarely, if ever, performed by surgeons today.

In 1973, a 22-year-old veterinary student in Indiana named John Cleland was diagnosed with metastatic testicular cancer-cancer of the testes. The cancer had metastasized into his lungs and lymph nodes. In 1973, the survival rate for such a cancer was less than 5 percent. Cleland was under the care of a young oncologist named Larry Einhorn in the cancer ward at Indiana University. Einhorn initially treated Cleland with a three-drug cocktail called ABO, which was found to be marginally effective. In the fall of 1974, Einhorn replaced the ABO regimen with a new chemical called cisplatin. Other researchers had seen transient responses in testicular cancer patients treated with cisplatin. Einhorn wanted to see if he could increase the response rate by combining cisplatin with two other drugs. For Celand, it was a choice between the uncertainty of the new regimen and the certainty of death. He took the gamble and enrolled as “patient zero” for BVP, the new regimen containing bleomycin, vinblastine, and cisplatin. Ten days later, the tumors in Cleland’s lungs had vanished.

By 1975, twenty additional patients had enrolled in the trial – all with remarkable and durable responses similar to Celand’s. By the late winter of 1976, it had become clear that some of these patients would not relapse at all.


Meanwhile, the NCI had turned into a factory of toxins. With money from the National Cancer Act, the institute’s drug-discovery program was testing zillions of chemicals each year to discover new cytotoxic drugs. The money also stimulated enormous, multi-site trials, turning academic centers into drug factories and cancer hospitals into efficient trial-running machines.

It was trial and error on a humongous scale, not targeted research. In one NCI-sponsored trial, known as the eight-in-one study, children with brain tumors were administered eight drugs in a day. Most of the children died soon afterward, having only marginally responded to the chemotherapy.