Use of mammography could lead to overestimation of tumor size and missing of lesions compared to magnetic resonance imaging (MRI), according to research presented here at the annual meeting of the Society of Surgical Oncology (SSO).

These differences in accuracy between mammography and MRI “might be particularly important for T1 and T2 lesions, where size is the most important criterion for the T-stage classification,” said presenting author Shalini Chitneni, MD, internal medicine resident, McLaren Regional Medical Center, Flint, Michigan, United States.

For their retrospective study, Dr. Chitneni and colleagues compared MRI and mammography using pathology as the gold standard. She presented the findings of this study on March 24^th.

They enrolled 100 patients with pathologically proven breast cancer who underwent both MRI and mammogram. Women who underwent excisional biopsy or neoadjuvant chemotherapy prior to MRI were excluded.

Results show that MRI found 138 lesions, of which 14 were benign based on pathology. Of the 110 lesions described by size, 104 were detected in the same breast and 6 in the opposite breast.

In 23% of patients, MRI detected additional lesions compared to mammography. Six of these patients had 1 additional lesion located by MRI on the opposite breast.

Mammogram detected 64 lesions, with 34% of these overestimated in size by at least 0.1 cm compared to pathology, and another 57 underestimated in size by the same margin. The mean overestimation was 1.10 cm and the mean underestimation was 0.78 cm.

For the 64 lesions detected by both MRI and mammogram, 55% were overestimated in size by at least 0.5 cm with MRI and 39% were underestimated by the same margin with MRI compared to pathological size.

MRI detected enlarged lymph nodes in the armpit and breast in 15% of patients, although 40% were true positives according to pathological findings.

Of the 78 patients with breast conserving surgery planned, 23 underwent additional lumpectomy or mastectomy due to additional lesions found by MRI; 6% had a change of surgery in the opposite breast due to new findings of contralateral breast cancer found with MRI.

The researchers concluded that mammography underestimated the breast lesions size significantly more often than MRI, and that when mammography overestimated size it was typically by more than 1 cm while MRI overestimated by less than 1 cm on average.

MRI is much better than mammography in detecting and estimating size of breast lesions, Dr. Chitneni said.