By Deborah Kotz Globe Staff

Researchers estimated Wednesday that mammograms caused more than 1 million American women over the past three decades to be diagnosed with early-stage breast cancers that would not have proved fatal if left undetected and untreated -- a controversial finding sure to provoke more heated debate over the benefits of the screenings.

The study, published in the New England Journal of Medicine, concluded that these women were “overdiagnosed” and “overtreated” because there was not a corresponding decline in the diagnosis of tough-to-cure breast cancers that have already spread. The authors said mammography has saved few lives.

“There’s been a dramatic increase in early-stage breast cancers that coincides with an increased use of mammograms,” said study co-author Dr. H. Gilbert Welch, a professor of medicine at the Dartmouth Institute of Health Policy and Clinical Practice in Hanover, N.H. “But it’s not enough to find early-stage breast cancer; it must translate into fewer women being diagnosed at a later stage, and we found this decrease to be remarkably small.”

But breast cancer specialists heatedly disputed the findings, with some calling the study “junk science” and others questioning the size of the overdiagnosis harms estimated by the researchers.

Dr. Leonard Lichtenfeld, deputy chief medical officer of the American Cancer Society, said that the findings “must be viewed with caution” and that the society stands by its recommendation for women to have annual screening mammograms beginning at age 40. “We find that the evidence supports the conclusion that mammography saves lives, and that the benefits of screening mammography outweigh the risks and harms.”