Quantitative evaluation of Kaiser score in diagnosing breast dynamic contrast-enhanced magnetic resonance imaging for patients with high-grade background parenchymal enhancement

Abstract

Background: High-grade background parenchymal enhancement (BPE), including moderate and marked, poses a considerable challenge for the diagnosis of breast disease due to its tendency to increase the rate of false positives and false negatives. The purpose of our study was to explore whether the Kaiser score can be used for more accurate assessment of benign and malignant lesions in high-grade BPE compared with the Breast Imaging Reporting and Data System (BI-RADS).

Methods: A retrospective review was conducted on consecutive breast dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) scans from 2 medical centers. Included were patients who underwent DCE-MRI demonstrating high-grade BPE and who had a pathology-confirmed diagnosis. Excluded were patients who had received neoadjuvant chemotherapy or who had undergone biopsy prior to MRI examination. Two physicians with more than 7 years of experience specializing in breast imaging diagnosis jointly reviewed breast magnetic resonance (MR) images. The Kaiser score was used to determine the sensitivity, specificity, and positive predictive value (PPV), and negative predictive value (NPV) of the BI-RADS from different BPE groups and different enhancement types. The performance of the Kaiser score and BI-RADS were compared according to diagnostic accuracy.

Results: A total of 126 cases of high-grade BPE from 2 medical centers were included in this study. The Kaiser score had a higher specificity and PPV than did the BI-RADS (87.5% vs. 46.3%) as well as a higher PPV (94.3% vs. 79.8%). The value of diagnostic accuracy and 95% confidence interval (CI) for the Kaiser score (accuracy 0.928; 95% CI: 0.883-0.973) was larger than that for BI-RADS (accuracy 0.810; 95% CI: 0.741-0.879). Moreover, the Kaiser score had a significantly higher value of diagnostic accuracy for both mass and non-mass enhancement, especially mass lesions (Kaiser score: accuracy 0.947, 95% CI: 0.902-0.992; BI-RADS: accuracy 0.821, 95% CI: 0.782-0.860), with a P value of 0.006.

Conclusions: The Kaiser score is a useful diagnostic tool for the evaluation of high-grade BPE lesions, with a higher specificity, PPV, and diagnostic accuracy as compared to the BI-RADS.

Keywords: Breast Imaging Reporting and Data System (BI-RADS); Kaiser score; dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI); positive predictive value (PPV); specificity.

Figure 1

Flow diagram for the inclusion and exclusion of patients. BPE, background parenchymal enhancement; MRI, magnetic resonance imaging.

Figure 2

The Kaiser score diagnostic process. A 39-year-old woman with a mass found during physical examination. DCE-MRI showed a moderate BPE. The lesion was isointense on T1WI, with a smooth edge and no root sign observed on either early and late subtraction. The TIC was persistent type, and the mean ADC value obtained was 1.03×10⁻³ mm²/s. The Kaiser score was 1 point, falling within the 1–4-point category and corresponding to BI-RADS 2/3. The final pathological result was adenosis. (A) T1WI; (B) the early subtraction of DCE-MRI; (C) the late subtraction of DCE-MRI; (D) TIC; (E) ADC map. DCE-MRI, dynamic contrast-enhanced magnetic resonance imaging; BPE, background parenchymal enhancement; T1WI, T1-weighted imaging; TIC, time-intensity curve; ADC, apparent diffusion coefficient; BI-RADS, Breast Imaging Reporting and Data System.

Figure 3

The total statistics of lesions grouped by the results of BI-RADS and Kaiser score. BI-RADS, Breast Imaging Reporting and Data System.

Figure 4

A false-positive case. The patient was a 25-year-old woman with a mass found on ultrasonography 1 week earlier. DCE-MRI showed marked BPE. The lesion was isointense on T1WI, the root sign was found around the lesion, the TIC curve was flat, and surrounding edema was present. The Kaiser score was 10, falling within the 8–11 category, the result was intraductal papilloma, and the BI-RADS grade was 4a. Both assessment methods yielded false-positive results. The images showed that the lesion was small, at approximately 10 mm × 9 mm in size; root sign could be seen around the lesion, with marked BPE and blurred lesion edges; and the TIC curve did not indicate a high possibility of a benign lesion. (A) T1WI; (B) the early subtraction of DCE-MRI, with the arrow indicating the root sign of the lesion; (C) the late subtraction of DCE-MRI. DCE-MRI, dynamic contrast-enhanced magnetic resonance imaging; BPE, background parenchymal enhancement; BI-RADS, Breast Imaging Reporting and Data System; T1WI, T1-weighted imaging; TIC, time-intensity curve.