Researchers raise more concerns about mammo trials in Canada

A team led by Dr. Martin Yaffe from the University of Toronto shared new eyewitness accounts that reported flaws in randomization in the Canadian National Breast Cancer Screening Studies (CNBSS), which found no benefit. routine mammography. These eyewitnesses have revealed cases in which symptomatic women moved from the control group to the mammogram group, the researchers said.

Dr Daniel Kopans.

“The studies violated most basic rules of randomized controlled trials and should not be used to determine screening guidelines,” said commentary co-author Dr. Daniel Kopans of Harvard University.

The two breast cancer screening trials began in Canada in 1980 with the aim of determining whether routine screening of women aged 40 to 49 by film-screen mammography and clinical breast examination reduces mortality from breast cancer. compared to usual care, as well as whether such screening and examination in women aged 50 to 59 reduced mortality beyond that obtained with examination alone.

Neither study found any benefit, and breast cancer screening experts over the years have raised concerns about these trials, including the process of randomization, poor image quality, and method flaws. recruitment, among other factors. Now, new eyewitness accounts have emerged that detail what researchers are calling “serious protocol violations” that reversed the randomization and made the trial results unreliable.

“For the results of randomized controlled trials to be valid, blind attribution is essential,” Kopans wrote in a separate overview. “Unfortunately, this basic rule has been ignored by studies.”

The researchers said women with symptoms of breast cancer were recruited into the trial. However, these trials aimed to focus on screening, which would benefit asymptomatic women. The team wrote that women with symptoms are unlikely to benefit from screening and require diagnostic testing, including imaging.

They also noted that some screening centers that worked with the trials saw a drop in the number of recruitments, which led to recruiting women to breast surgeon offices.

“The likelihood of these women being asymptomatic would have been very low,” they said.

Additionally, researchers said they spoke with healthcare professionals at these sites and 17 of those staff said they were aware of recruiting symptomatic women. With some of these women having “no doubt” breast cancer, the team wrote that these numbers may have contributed to deaths in the studies, as well as a systematic bias.

Yaffe et al also said that medical staff have described women moving from the control group to the mammography group, and that randomizing even a few women with advanced breast cancer could affect the effectiveness of the measured screening.

The team also pointed out that with the exception of one of the 15 study sites, all of the women underwent a clinical breast exam by a qualified nurse examiner enrolled in the studies. Citing previous investigations, the group said this could have resulted in the reassignment of women to a part of the study that measured both mammography and clinical breast exam.

A person the researchers said they spoke with in March after a randomization conference confirmed the compromised randomization of the trials. The person, a mammography technologist, worked at the study site at St Michael’s Hospital. His name was omitted from the commentary.

“She reported that some women with breast lumps were deliberately assigned to the study’s mammogram arm, rather than randomized blind,” wrote the commentators.

They also wrote that the technologist’s testimony was supported by a “very high” cancer detection rate in the hospital mammography arm, which was reportedly higher than at other sites and 12 times higher than the hospital mammogram arm. clinical breast examination. The benign-to-malignant biopsy ratio at this site was also low compared to other sites, which the researchers said suggests the prevalence of larger and easily detected cancers.

“Subsequently, another eyewitness, a radiologist at a separate screening site in another city, provided information about serious protocol violations at that site,” they added.

Although the researchers said they believed coordinators and nurses with no previous experience in randomized controlled trials had good intentions, the study design protocols allowed for non-randomization.

“The data has always suggested that it happened; and now we have first-hand testimony that it happened,” Kopans said. “These violations have hopelessly corrupted the results. It is now clear that the CNBSS trials have been fundamentally compromised and the trial results are unreliable.”

Copyright © 2021

Comments are closed.