Sunday, November 21, 2010

Preoperative MRI can be used without increasing mastectomy rates

RSNA (Radiologists' Society of North America) holds a gigantic radiologists conference every year in Chicago. This year's conference starts one week from today and apparently one of the presentations on the first day will be from Fabio Chiesa and colleagues from the University of Milan's School of Medicine. Apparently they will present data from 2003 to 2008 showing that preoperative MRI examinations of the breast do not necessarily cause increases in mastectomy rates.

This is a significant finding, particularly as another recent study demonstrated a 6% increase in the mastectomy rate when a preoperative MRI was performed (COMICE trial). This has been used as an argument that preoperative MRI is not good for patients (Kell, Magnetic resonance mammography, British Medical Journal, October 2010).

Thanks to this article, we have the following preview of their data, where the numbers and percentages of mastectomies are provided for each year, and then the number of preoperative MRIs performed is provided afterwards. Even though the number of MRIs has been increasing there hasn't been a corresponding increase in the mastectomy rate.

"25 of 121 (20.7%) in 2003
25 of 159 (15.7%) in 2004
22 of 140 (15.7%) in 2005
21 of 161 (13%) in 2006
26 of 182 (14.3%) in 2007
49 of 220 (21.4%) in 2008

There were three preoperative MRI examinations in 2003, two in 2004, 13 in 2005, 14 in 2006, 31 in 2007, and 42 in 2008,"

This would seem to indicate that MRI technology does not necessarily need to cause an increase in the mastectomy rate. This is good news for researchers like me who are trying to improve breast MRI based cancer detection technologies and great news for women at high risk of developing breast cancer as MRI has recently been shown to improve breast cancer survival among those at elevated risk for the disease. Ideally we can avoid unnecessary mastectomies for women who receive breast MRI examinations.

Jacob Levman, PhD
Imaging Research
Sunnybrook Research Institute
University of Toronto