Many people believe that cancer is a relatively recent phenomenon caused by the stresses of modern life. Dietary changes, behavioral changes and manmade changes to our environment have subjected humans to toxins that contribute to cancers.

But new findings from researchers at South Africa's University of the Witwatersrand published in the South African Journal of Science show that it's a bit more complicated.

Paleontologists found a benign tumor in a 12- or 13-year-old boy specimen that dates back almost 2 million years. They also found a malignant tumor on a little toe bone of a left foot that's 1.7 million years old.

Previously, the oldest discovered human cancer was between 780,000 and 120,000 years old.

Much of the reason that few cancers have been found in ancient specimens is that people just didn't live long enough to develop tumors. Still, the discoveries, found by using new 3D imaging methods, are leading some scientists to reevaluate the role of tumors in the history of early human ancestors.

The earliest discussion of cancer, according to National Geographic, came from Egyptian physician Imhotep, who described in his writings a "bulging at the breast" that did not respond to any known remedies.

Things go wrong inside our bodies, the study says, that are unrelated to stresses brought on by society, like pollution and smoking. Even millions of years ago, they sometimes resulted in tumors or growths that may or may not have been cancerous.

"The evidence is out there that these conditions have been with us a long time and we've been kind of hoodwinked that cancer is a modernity," said Patrick S. Randolph-Quinney, one of the study's authors. "These things are ancient."

The greatest predictor of cancer, the study argues, is longevity. The longer we live, the more chances something in our bodies goes wrong, the more chances that something is a tumor.

And yes, the incidence of cancer has increased over time because our habits have changed, but longevity provides more opportunities for things to go wrong.