The utility of ultrafast MRI and conventional DCE-MRI for predicting histologic aggressiveness in patients with breast cancer
Seong Gwang Kim,Ah Young Park,Hae Kyoung Jung,Kyung Hee Ko,Yunju Kim
DOI: https://doi.org/10.1177/02841851241276422
Abstract:Background: Prediction of histologic prognostic markers is important for determining management strategy and predicting prognosis. Purpose: To identify important features of ultrafast and conventional dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) that can predict histopathologic prognostic markers in patients with breast cancer. Material and methods: Preoperative MRI scans of 158 consecutive women (mean age = 54.0 years; age range = 29-86 years) with 163 breast cancers between February 2021 and August 2022 were retrospectively reviewed. Inter-observer agreements for ultrafast MRI parameters were analyzed by two radiologists. The qualitative and quantitative MRI parameters were correlated with histopathologic prognostic markers including molecular subtypes and histologic invasiveness. Results: Inter-observer agreements for ultrafast MRI parameters were excellent (intraclass correlation coefficients of area under the kinetic curve [AUC], maximum slope [MS], maximum enhancement [ME], and slope = 0.987, 0.844, 0.822, and 0.760, respectively). Triple-negative breast cancers (TNBC) were significantly associated with rim enhancement (odds ratio [OR] = 9.4, P = 0.003) and peritumoral edema (OR = 17.9, P = 0.002), compared to luminal cancers. Invasive cancers were associated with lesion type-mass, increased delayed washout, angiovolume, ME, slope, MS, and AUC, compared to in situ cancers. In regression analysis, the combination of MS (>46.2%/s) (OR = 5.7, P = 0.046) and delayed washout (>17.5%) (OR = 17.6, P = 0.01), and that of AUC (>27,410.3) (OR = 9.6, P = 0.04), delayed washout (>17.5%) (OR = 8.9, P = 0.009), and lesion-type mass (OR = 4.6, P = 0.04) were predictive of histologic invasiveness. Conclusion: Conventional DCE-MRI with ultrafast imaging can provide useful information for predicting histologic underestimation and aggressive molecular subtype. MS and AUC on ultrafast MRI can be potential imaging markers for predicting histologic upgrade from DCIS to invasive cancer with high reliability.