New parameters of ultrafast dynamic contrast-enhanced breast MRI using compressed sensing

Abstract

Background: Ultrafast dynamic contrast-enhanced (UF-DCE) breast MRI is considered a promising method of accelerated breast MRI. However, the value of new kinetic parameters derived from UF-DCE need clinical evaluation.

Purpose: To evaluate the diagnostic performance of the maximum slope (MS), time to enhancement (TTE), and time interval between arterial and venous visualization (AVI) derived from UF-DCE MRI using compressed sensing (CS).

Study type: Retrospective.

Population: Seventy-five patients with histologically proven breast lesions. The total number of analyzed lesions was 90 (61 malignant and 29 benign).

Field strength/sequence: 3T MRI with UF-DCE MRI based on the 3D gradient-echo volumetric interpolated breath-hold examination (VIBE) sequence using incoherent k-space sampling combined with a CS reconstruction followed by conventional DCE MRI.

Assessment: The diagnostic performance of the MS, TTE, AVI, and conventional kinetic analysis was analyzed and compared with histology.

Statistical tests: Wilcoxon rank sum test, receiver operating characteristic analysis.

Results: The MS was larger and the TTE and AVI were smaller for malignant lesions compared with benign lesions: MS: 29.3%/s and 18.4%/s (P < 0.001), TTE: 7.0 and 12.0 seconds (P < 0.001), AVI: 2.7 and 4.4 frames (P = 0.006) for malignant and benign lesions. The discriminating power of the MS (area under the curve [AUC], 0.76) was slightly better than that of conventional kinetic analysis (AUC, 0.69) and comparable to that of the TTE and AVI (AUC, 0.78 and 0.76 for TTE and AVI, respectively). Invasive lobular carcinoma had smaller MS (21.8%/s) among malignant lesions (29.3%/s).

Data conclusion: The MS, TTE, and AVI can be used to evaluate breast lesions with clinical performance equivalent to that of conventional kinetic analysis. These parameters vary among histologies.

Level of evidence: 3 Technical Efficacy: Stage 2 J. Magn. Reson. Imaging 2020;51:164-174.

Keywords: breast diseases; compressed sensing; kinetic parameter; magnetic resonance imaging; ultrafast imaging.