MR spectroscopy can improve specificity in breast cancer detection without losing sensitivity. And the acquisition and postprocessing of spectroscopic data are sufficiently user-friendly to become routine in a clinical setting, according to research presented Tuesday at the ECR.

Dr. Alfonso Fausto and colleagues at the Institute Polyclinic San Donato in Milan, Italy, evaluated 17 consecutive patients with a highly spatially resolved gadolinium-enhanced MR sequence and single-voxel proton MRS. The MRS protocol added an average 13 minutes to the 15-minute dynamic imaging.

A choline peak equal to or greater than 2 was considered malignant. In the eight malignancies, the median intensity choline peak was 4.6. Two benign cases had peaks greater than 6 (one papilloma, one dysplasia).

"The absence or presence of a choline peak strengthens the results of dynamic MR imaging," Fausto said.

Sensitivity was high at 90% for both dynamic imaging and MRS, while specificity was 75% for MRI and 88% for MRS.

In another study, Dr. Isil Topcu and colleagues at the School of Medicine Dokuz Eylul in Turkey found that the choline-fat ratio differentiates between malignant and benign breast lesions. The ratio also increases linearly with the tumor's malignancy grade.

The researchers prospectively evaluated 44 patients with single-voxel hydrogen-1 MRS and dynamic MRI. Mean lesion size was 2.4 cm. Elevated choline levels were detected in 17 of 19 carcinoma cases. In two carcinoma and 25 benign cases, choline levels were within normal range. Both false-negative lesions by MRS were smaller than 15 mm.

The sensitivity of MRS to detect breast cancer was 89%, and the specificity was 100%. Positive predictive value was 100% and accuracy 95%.

Several audience members questioned why the researchers chose to evaluate the choline-fat ratio. Topcu said that choline-fat ratios might prove useful to follow response to treatment. The group is currently evaluating the theory.