Patient Outcome: Not Quite Right for Breast MRI

Some studies on MRI for breast cancer diagnosis may use the wrong parameters, warned German breast MRI expert Christiane Kuhl: Outcome is the correct measurement for treatment success, not for diagnostic tools.

Using outcome measures is regarded highly in ‘value based medicine’, but it might have harmful consequences for the radiological community, explained Kuhl.

Studies using patient outcomes such as lower excision rates or rates of mastectomies mix the quality of the diagnostic tool with the quality of patient management and surgery. All the Oxford Centre for Evidence-Based Medicine 2011 levels of evidence demand for a new diagnostic test is high diagnostic accuracy. Patient outcomes are reserved for treatment.

Surgeon Effect Trumps MRI Effect

A lot of data supports the fact that MRI is far more accurate than mammography or ultrasound for delineating disease extent. “This impacts surgical management in a substantial number of cases,” said Kuhl.

While radiology can influence the imaging tool, it is, however, unable to influence what surgeons do with this information. Surgical practice is individual, as for a study on re-excision rates after breast conservation surgery shows (McCahill 2012). It evaluated re-excision rates and the correlation of these rates with the individual case load of 54 surgeons from four large US cancer centers. The excision rates ranged between 0 and 70 percent. The rates were unrelated to the surgeon’s practical experience or case load. “It was just a matter of individual type,” commented Kuhl.

This means that huge variations of individual practice patterns will confound or override any impact of diagnostic imaging on surgical outcome.“

Programmed Failure

Kuhl said the challenge was not to be impactful, but how to measure the impact.

Several large prospective randomized trials that were set up to demonstrate the impact of pre-operative MRI on re-excision rates suffer from that structural mistake. Kuhl referred to the COMICE (Turnbull 2010), and MONET (Peters 2011) trials, both of which showed no reduction of re-excisions. “Predictably, they failed,” commented Kuhl.

The currently ongoing MIPA study also looks at a lot of endpoints beyond radiologist’s control, noted Kuhl. “Surgeons are difficult to parameterize – let’s hope we have enough data,” she said.

These trials do have an impact on clinical practice. So far, none of the guidelines world wide recommend breast MRI for treatment planning in women scheduled for breast conserving surgery. Re-excision rates remain high. Data from a UK study with more than 55,000 women with breast conserving therapy shows an overall reoperation rate of more than 20%. It goes as high as 30% in DCIS patients (Jeevan 2012). The study claims that a “lack of accurate imaging, especially of DCIS And DCIS components of invasive cancer is the main reason for positive margins and repetitive surgery”. MRI could provide the diagnostic accuracy. Radiology can also provide more than sufficient data on this diagnostic accuracy. But it provides information, not treatment. Surgeons and oncologists need to use this information. Or as Kuhl put it: “A fool with a tool is still a fool  -we as radiologists are the tool”. If radiologists engage with outcome measures they are likely to fail. “Don’t be fooled, just stick with your convention”, she advised.


During a vivid discussion Kuhl strongly urged her colleagues to get involved in clinical trials. Radiologists and clinicians should get together and discuss their study goals.


Jeevan R et al. Reoperation rates after breast conserving surgery for breast cancer among women in England: retrospective study of hospital episode statistics. BMJ 2012;345:e4505.

McCahill LE et al. Variability in reexcision following breast conservation surgery. JAMA 2012;307(5):467-75.

Peters NH et al. Preoperative MRI and surgical management in patients with nonpalpable breast cancer: the MONET - randomised controlled trial.  Eur J Cancer. 2011;47(6):879-86.

Turnbull L et al. Comparative effectiveness of MRI in breast cancer (COMICE) trial: a randomised controlled trial. Lancet. 2010;375(9714):563-71.

ECR 2019
AutorIn: kf/ktg,
Related content: Breast, MRI

Presentation Title: What evidence do we need to justify the use of breast MRI?
Speaker: Christiane Kuhl, Aachen University, Germany
Date: Thursday, 28th February 2019
Session Code:
RC 502