Chintana P. Paramagul MD, Presenter: Nothing to Disclose

Michelle Foster MD, Abstract Co-Author: Nothing to Disclose

Gisela Christa Mueller MD, Abstract Co-Author: Nothing to Disclose

Mark Alan Helvie MD, Abstract Co-Author: Nothing to Disclose

Lubomir Hadjiiski PhD, Abstract Co-Author: Nothing to Disclose

Thomas Leonard Chenevert PhD, Abstract Co-Author: Nothing to Disclose

The shape of the time-signal intensity (SI) curves is an important criterion in differentiating benign and malignant enhancing lesions in dynamic breast MRI. Various quantitative methods are being used to obtain the region-of-interest (ROI. Our purpose was to assess the divergence of SI curves of primary breast cancers, in relation to four different types of ROI.

IRB approval was obtained prior to the commencement of the study.41 patients with palpable breast cancer had dynamic gadolinium-enhanced (0.1 mg/kg) breast MRI. Dedicated breast coil (GE signal 1.5 T) was employed. MQSA certified radiologists with knowledge of the patient’s breast cancer retrospectively obtained 4 different types of ROIs: ROI A (means, 667 mm²) derived from hand-segmented outline of malignant margins. ROI B was created by manually fitted an ellipse to the inner margins of the mass (means, 442 mm²). ROI C (mean, 17 mm²), and ROI D (mean, 18 mm2) were automated to fit within in the area of the maximum and sub maximal enhancement of the tumor on the same image, respectively. Kinetic measurement was determined at 75 seconds time point after contrast injection. SI curves of the cancer were obtained and classified according to their shapes as pattern 1, steady enhancement; pattern 2, plateau; and pattern 3, washout of signal intensity. The pattern of enhancement curve of each cancer was compared at 4 different types of ROI. Threshold for cancer determination was SI curve pattern 2 and pattern 3, obtained at 75 seconds after contrast injection.

The accuracy in predicting cancer was significantly different between the 4 (p < 0.05). Using pattern 2 and 3 as a threshold for determination of malignancy, the true positive value for the four different geographic regions measures were 66% (27/41), 71% (29/41), 88% (36/41), and 49% (20/41) respectively.

Time-signal intensity patterns of primary breast cancers by MRI vary significantly within the same cancer depending upon the region-of-interest and the size of ROI chosen. Small ROI at maximum enhancement area displayed the most frequent time-signal intensity curves fitted with malignant characters, but would likely be the least specific.

Paramagul, C, Foster, M, Mueller, G, Helvie, M, Hadjiiski, L, Chenevert, T, Variable Patterns in Time-Signal Intensity Curve of Primary Breast Carcinoma in Magnetic Resonance (MR) Imaging.  Radiological Society of North America 2004 Scientific Assembly and Annual Meeting, November 28 - December 3, 2004 ,Chicago IL. http://archive.rsna.org/2004/4414427.html