Breast Cancer Screening: All Women Benefit From MRI

Data from the last decade shows that MRI beats mammography for breast cancer screening in almost every aspect. However, mammography remains the number one screening method. Christiane Kuhl attempted to answer why.

In the 1970ies, when MRI mammography first came up, the first idea was “to find out whether screening per se works,” said Christiane Kuhl, Aachen University, Germany. It did: “No one thought early diagnosis would make an impact on survival, but many randomized controlled trials back then showed that mammography reduces mortality by at least 25%.”

It has since become clear that individual women have different screening needs, just as their cancer risk differs. “We now know how to evaluate risk,” said Kuhl. Not only lifetime risk, but also the risk for the next five years. This has led to risk categorizations:

  • Low risk equals an expected breast cancer incidence within five years of less than 1%,
  • Average means a risk between 1 and less than 2%
  • Increased risk ranges between 2 to less than 6%, and
  • High risk equals a risk of over 6%.

So far, guidelines only recommend breast MRI for the high-risk group. These recommendations remain, although there is enough data showing that screening mammography will fail early diagnosis in patients with high breast density and the risk to develop triple-negative, i.e. highly aggressive breast cancer. “It is a one size fits all approach, while we know that one size does not fit all at all,” stated Kuhl.

Genomics Have Changed the Game

First of all, breast cancer is a heterogeneous disease. Its genotype determines how a cancer will present itself in imaging, i.e. its imaging phenotype. For example:

  • Luminal-A cancers, which make up for about 40% of all cases, are HER negative and have low Ki-67. Their prognosis is fairly good and their likelihood to metastasize is fairly low. Yet, in mammographic terms, they resemble a malignant structure with the spicular and stellant edges.
  • A triple-negative (ER-, HER-, and any Ki-67) tumor, which has a high likelihood of metastases, resembles a benign structure in mammography.

Mammography reflects regressive changes of slow-growing cancers, i.e. architectural changes due to hypoxia, necrosis and fibrosis. These cancers usually have a better prognosis. “If I could choose, I would like the tumor found on mammography, it finds the nice cancers”, said Kuhl. This type of overdiagnosis is a “length-time bias put to an extreme,” she noted. What causes the overdiagnosis? “It is the way we screen,” noted Kuhl.

Two Problems – Overdiagnosis and Underdiagnosis

The root of the mammography problem is overdiagnosis of irrelevant tumors versus an underdiagnosis of relevant tumors, explained Kuhl.

Study estimates for breast cancer overdiagnosis range between 10 to 52%. “With a range like this, this means no one knows for sure”, commented Kuhl. However, it is likely that there is overdiagnosis.

Mammographic underdiagnosis also exists, and it is both patient-related and tumor related: Patient related, because dense breasts hide cancers. Tumor-related, because “what looks benign is often related to aggressive cancer phenotypes in any breast.

Screening Could Do More

Despite long-term screening efforts, evidence from cancer epidemiology shows that breast cancer remains the most frequent cancer death in women below 50 years of age and continues to be then main cause of life years lost in women. “We may use an inappropriate method to find cancer cases,” said Kuhl. The ideal method should offer a sensitivity profile fitting oncological needs: This means high sensitivity for biologically significant disease and desirably low sensitivity for biologically in-significant disease.

MRI Wins Screening Comparison

Kuhl showed comprehensive data that allows comparing screening methods. The main parameter she used is the additional cancer yield reached by each method. This parameter quantifies how many more correct breast cancers have been detected by that method in 1000 women in one year.

  • Digital Tomosynthesis leads to an additional cancer yield of 2.25 women per 1000.
  • With ultrasound an additional cancer yield of 4.1 per 1000 can be achieved, however, the false-positives are high and the exam takes an average 20 minutes. “Taking this into consideration, this is the most expensive method we have,” commented Kuhl.
  • For MRI, the additional cancer yield was 14.6 per 1000. This was the result of a study (Berg 2012) that added MRI or ultrasound in a cohort pre-screened with mammography.

“The jump in cancer yield was caused by MRI only, whatever else you add does not matter,” noted Kuhl.

Interval Cancer Rate is the Right Metric

The actual metric for assessing the success of screening would be interval cancers, said Kuhl. She presented data from prospective clinical trials between 2004 and 2017 for screening women at increased breast cancer risk: “Between twenty and thirty percent of all cancers are interval cancers, which means they are found between mammography screenings,” said Kuhl. These interval cancers have a significantly worse prognosis compared to mammographically visible cancers (Houssami 2017). Interval cancers indicate that mammography often fails to show the right thing.

A recent study (Kuhl 2017) shows that MRI imaging detects interval cancers early, even in average risk women. “This means MRI improves breast cancer detection and effectively avoids underdiagnosis of interval cancers, no matter what the personal risk of the woman is,” concluded Kuhl.

Discussion

Being asked why breast screening MRT implementation is so slow, she said: – is it still medical or already political: “There is no good answer, but the resistance certainly does not come from patients, the radiological community itself might be somewhat resistant.” Session chair Lale Umutlu, Essen University, Germany, asked why there was no resistance in prostate cancer MRI. “That is because we do not have a ‘man-o-gram’, there was a complete absence of other tools to compare MRI to,” answered Kuhl.

Another audience member commented that in Australia and the US, resistance comes from the surgeons, as contrast enhanced MRI seems to introduce complexity.

References

Berg W et al. Detection of breast cancer with addition of annual screening ultrasound or a single screening MRI to mammography in women with elevated breast cancer risk.

JAMA. 2012;307(13):1394-404.

https://doi.org/10.1001/jama.2012.388

Houssami N, Hunter K. The epidemiology, radiology and biological characteristics of interval breast cancers in population mammography screening. NPJ Breast Cancer 2017;13(3):12.

https://doi.org/10.1038/s41523-017-0014-x

Kuhl CK et al. Abbreviated breast magnetic resonance imaging (MRI): first postcontrast subtracted images and maximum-intensity projection-a novel approach to breast cancer screening with MRI. J Clin Oncol. 2014 Aug 1;32(22):2304-10.

https://doi.org/10.1200/JCO.2013.52.5386

Kuhl CK et al. Supplemental Breast MR Imaging Screening of Women with Average Risk of Breast Cancer. Radiology. 2017;283(2):361-370.

https://doi.org/10.1148/radiol.2016161444

Presentation Title: Using Abbreviated Protocols
Speaker: Christiane Kuhl, Aachen University, Germany
Date: Wednesday, 27th February 2019
Session Code:
SY 1