After the CD8 cells did their job, they remained in the blood but, unexpectedly, they turned into CD4 cells. And when the Penn investigators removed CD4 cells from the blood of Mr. Ludwig and Mr. Olson, they saw that those cells could kill B cells in the laboratory. The CD4 cells had turned into assassins or, Dr. DiPersio noted, “at least guardians that can keep the tumor cells at bay and undetectable in the patient for years.”

Could the CD4 cells remain in the blood with no cancer cells to kill? Or were they there because the leukemia was not really gone but instead kept trying to return, only to be attacked by CD4 cells?

“We can’t find any leukemia cells in Doug,” Dr. June said. But, he added, perhaps they are still there in tiny quantities and emerging, only to be knocked back by CD4 cells, “like whack-a-mole,” he said.

He suspects, though, that the CD4 cells are more like guards.

“The leukemia is gone, but they stay on the job,” he said.

Whatever the mechanism, Dr. Porter said, the result “is beyond my wildest imagination.”

“Oncologists don’t use words like ‘cure’ lightly or easily or, frankly, very often,” he said. “I guarantee that it’s not being used lightly. The patients we treated had far advanced disease,” he noted, adding, “the biggest disappointment is that it doesn’t work all the time.”

“Historically, if these cancers don’t recur in two to five years the likelihood of relapse is low,” said Dr. Hagop M. Kantarjian, chairman of the department of leukemia at the University of Texas’s M.D. Anderson Cancer Center.

For Mr. Olson, now 75 and living in Pleasanton, Calif., life is good. He still shakes his head over the amazing coincidence that his oncologist just happened to be an investigator in that clinical trial a decade ago.

“I’m a lucky man,” he said.