Should we scrap routine mammography?

An analysis of breast cancer registry data in Victoria has found routine mammographic screening did not result in a drop in advanced cancer diagnoses nor an increase in early cancer diagnoses

---

A new analysis suggesting treatment is primarily responsible for the decline in breast cancer deaths in Victoria has reignited debate over whether to scrap routine mammography. 

An analysis of breast cancer registry data in Victoria found that routine mammographic screening, BreastScreen, did not result in a drop in advanced cancer diagnoses nor an increase in early cancer diagnoses.

Study author Professor Robert Burton, epidemiologist and preventative health expert at Monash University, said that the lack of benefit, coupled with the harms of overdiagnosis, meant the program should be scrapped. 

“We found that adjuvant therapy accounted for the observed 30% mortality decline,” Professor Burton said. “Given this finding, we propose that BreastScreen should be terminated.”

Professor Burton, and his co-author, Associate Professor Christopher Stevenson, an epidemiologist at Deakin University, analysed health data from more than 75,000 Victorian women who were diagnosed with breast cancer since 1982. 

The paper, published in JAMA Network Open, showed that breast cancer mortality dropped by 30% in the 20 years following 1994, from 34 women per 100,000 to 24 women per 100,000. 

This corresponded with a big boom of interest in mammography screening around the world in the 1990s.  But for breast cancer screening to be effective, mammograms should result in more early stage breast cancer diagnoses and fewer advanced stage diagnoses. 

Research from NSW, the US, Norway and the Netherlands indicates that the incidence of advanced breast cancer has either remained stable or increased. 

In this analysis, Professor Burton found that advanced stage breast cancer actually doubled, from 12 to 24 women per 100,000, between 1986 to 2013.   Meanwhile, they found that adjuvant therapy for early breast cancer, which included tamoxifen and chemotherapy, more than tripled 1986 and 1999.

According to their calculations, this could account for the entire 30% reduction in crude breast cancer deaths after 1994. 

In contrast, regular mammograms exposed women to the risks of overdiagnosis and treatment, and at least one in three women diagnosed with mammography screening were overdiagnosed, meaning the cancer would have never gone on to cause harm, Professor Burton said. 

A large portion of women with early breast cancer receive postoperative external beam radiation therapy (EBRT), which comes with risks of potentially fatal cardiac events. 

This could have resulted in 54 extra deaths in 2004 and 78 extra deaths in the five years after 2013, said Professor Burton. 

But the 30% figure for overdiagnosis was much higher than Cancer Australia figures, which suggested it might be more like 11 to 19%, Professor Bruce Mann member of BreastScreen Victoria’s Clinical Reference Group and tumour stream director at the Victorian Comprehensive Cancer Centre, said.

He hit back at the latest paper, saying it was based on a number of flawed assumptions. 

“The reason that the outcomes from breast cancer have improved over the last 20 to 25 years is a combination of early diagnosis and treatment,” he said. 

“The idea that it’s one or the other is a false dichotomy.” 

Professor Mann thinks the apparent increase in advanced cancers is due to a change in the definition of “stage 3”, which now requires four or more involved lymph glands rather than “matted” axillary lymph glands prior to 2006. 

Professor Mann said that BreastScreen helped women identify cancers earlier, allowing them to be treated less invasively than if they were caught later. He pointed to research showing women diagnosed with the screening program had half as many mastectomies (35% vs. 17%), fewer axillary dissections (43% vs. 21%) and less chemotherapy (65% vs. 41%). 

Focusing on death rates alone missed this reduction in morbidity, he said. 

He also pushed back on the level of harm the paper claimed women were exposed to from treatment, saying radiation technology had improved “dramatically”. 

“Saying that the morbidity from radiation from the 1970s or the 1980s still applies to those diagnosed and treated now is misleading,” he said. 

Similarly, those overdiagnosis figures didn’t tell the whole story, because the low risk women wouldn’t be recommended to have chemotherapy and mastectomy, he said. 

“This is a dangerous paper, because it may discourage women from participating,” Professor Mann said. 

“Sure, there are ways breast screening could be further improved, but to suggest that it is harmful is mischievous at best.”   

However, Professor Burton said there was a growing body of evidence casting doubt on the benefits of population screening in asymptomatic women. 

Conflicting information was discarded by proponents of BreastScreen because it had become a big industry, supporting many people’s livelihoods, Professor Burton said, likening it to the military-industrial complex. 

JAMA Network Open 2020, 24 June