Diagnostic Performance of DCE-MRI Breast Using Kaiser Score in Characterization of Mass and Nonmass Lesions and Its Comparison with ACR BI-RADS

Abstract

Introduction

Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) is a highly sensitive modality for the detection and characterization of breast lesions, yet its limited specificity and interpretative variability pose diagnostic challenges. The American College of Radiology Breast Imaging Reporting and Data System (ACR BI-RADS) provides a structured lexicon but lacks definitive guidance for certain lesion categories, particularly nonmass enhancements. The Kaiser score (KS), a semiquantitative decision-support tool, has emerged as a potential adjunct to standard interpretation and offers a structured approach to improve diagnostic accuracy.

Objectives

The study was aimed to evaluate the diagnostic performance of DCE-MRI of the breast by applying the KS in the characterization of mass and nonmass enhancement and its comparison with ACR BI-RADS.

Materials and Methods

Two radiologists assessed the KS and ACR BI-RADS on 103 sequential patients on 3-T DCE-MRI with 142 histopathologically verified lesions. The diagnostic performance of the KS was recognized through receiver operating characteristic (ROC) by the area under the ROC curve (AUROC). Cohen's kappa coefficient was used to evaluate the inter-reader agreement. These findings were compared and correlated with ACR BI-RADS.

Results

The KS has sufficiently high AUROC for all the lesions including mass and nonmass lesions (0.895, 0.955, and 0.622, respectively). The sensitivity of the KS was similar to that of ACR BI-RADS for both readers (93.6–91.5%) with a higher specificity of 85.4% compared with 62.5% for ACR BI-RADS. The improvement in specificity was also seen for mass as well as nonmass lesions. Excellent inter-reader agreement was observed with kappa values of greater than 0.9.

Conclusion

DCE-MRI using the KS showed high diagnostic accuracy as compared with ACR BI-RADS with an excellent inter-reader agreement. Thus, the KS in conjugation with ACR BI-RADS can enhance diagnostic accuracy and decrease experience-related variability.

 

MissingLinkText

Keywords

dynamic contrast-enhanced magnetic resonance imaging - breast cancer - ACR BI-RADS - Kaiser score

Introduction

Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) of the breast is an extremely sensitive modality for diagnosis of potential malignancies with low specificity.[1] DCE-MRI of the breast has been in use for a long time and its applications in clinical practice have evolved considerably.[2] [3] [4] [5] Apart from the well-established indications that include staging of breast cancer, occult primary detection, and high-risk screening, DCE-MRI is now being used in diagnostic settings when conventional radiological investigations like mammography and ultrasound are inconclusive.

One such important clinical setting that we routinely encounter is the presence of multiple lesions in both breasts that are indeterminate on mammography and ultrasound. Now we need to identify the most suspicious lesion for tissue diagnosis or diagnostically prove that all lesions are benign. In such a setting, high sensitivity of DCE-MRI may result in unnecessary biopsies or surgeries.[3] [4]

Interpretation DCE-MRI of the breast is complex as many sequences have to be assessed simultaneously and without a structured approach. Hence, it may lose most of its advantage as a highly sensitive modality to rule out malignancy due to experience-related inter-reader variability and may further lead to an increase in false positives on imaging.[3] [4] [5]

The American College of Radiology Breast Imaging Reporting and Data System (ACR-BIRADS) for MRI is a main step for standardized and structured reporting, but it does not offer an elaborate and succinct approach to guide diagnostic decision-making.[6] BI-RADS makes it simple to characterize a lesion using descriptors; however, unlike mammogram there are no exact guidelines for category assignment on DCE-MRI, especially for category III and IV lesions.[7] [8] This variability is especially notable while evaluating nonmass lesions. Attempts have been made to develop quantifiable parameters to reduce the subjectivity of DCE-MRI interpretation. Furthermore, to increase the diagnostic accuracy and specificity of DCE-MRI, attempts have been made to evaluate multiparametric scoring and evaluation systems with varying results. Many of these studies have variably incorporated kinetic curves, apparent diffusion coefficient (ADC) values, morphological characteristics with T2 signal, and T1 relaxivity to build up a quantitative model that can be translated to BI-RADS categories.[4] [7] [8] [9] [10]

A relatively semiquantitative sequential approach is given by the Kaiser score (KS), including five independent diagnostic BI-RADS lexicon criteria (root sign, margins, signal intensity–time [SI-time] curve type, internal enhancement, and presence of edema) in a flowchart.[11] [12] Based on these characteristics, a lesion is given a score that can be used for making clinical decisions. A KS of greater than 4 requires image-guided tissue sampling/biopsy of the breast lesion.[11] [12] [13] [14] Two other optional moderators that are microcalcifications on mammogram and ADC values of greater than 1.4 can be used to upgrade or downgrade the KS by 2 and 4 points, respectively.[12] [14]

With this background, we aimed to assess diagnostic performance of DCE-MRI of the breast by applying the KS in the characterization of mass and nonmass enhancement and its comparison with ACR BI-RADS.

 

MissingLinkText

Materials and Methods

Study Design

A retrospective observational analytical study was conducted at the Rajiv Gandhi Cancer Institute and Research Centre, a tertiary care hospital in New Delhi. The study was approved by institutional review board.

 

MissingLinkText

 

MissingLinkText

Study Population

Inclusion Criteria

This study used a retrospective Picture Archiving and Communication System (PACS) database of sequential DCE-MRI breast scans at our hospital between January and December 2023. A total of 214 sequential DCE-MRI breast scans were performed.

 

MissingLinkText

Exclusion Criteria

Screening scans (BI-RADS 1), scans with BI-RADS 2 findings, repetitive scans without new suspicious findings, cases lacking histopathology, and post neoadjuvant chemotherapy/radiotherapy (NACT) cases were excluded from the study. After applying these criteria, a total of 142 histologically verified lesions in 103 patients were included in the study. BI-RADS 3 to 5 lesions were independently evaluated and histologically verified based on suspicion.

 

MissingLinkText

Objectives

The aim of the study was to evaluate the diagnostic performance of DCE-MRI of the breast by applying the KS the characterization of mass and nonmass lesions and its comparison with ACR BI-RADS.

 

MissingLinkText

Primary and Secondary Outcomes

Primary Outcome

The primary outcome was diagnostic accuracy of KS in differentiating benign from malignant breast lesions (mass and nonmass) on DCE-MRI, as assessed by area under the receiver operating characteristic curve (AUROC), and comparison with the ACR BI-RADS categorization.

 

MissingLinkText

Secondary Outcomes

• Sensitivity and specificity of the KS compared with ACR BI-RADS for all lesions, mass lesions, and nonmass lesions

• Positive predictive value (PPV) and negative predictive value (NPV) of the KS versus ACR BI-RADS.

• Inter-reader agreement for both the KS and ACR BI-RADS using Cohen's kappa coefficient.

• Univariate analysis of individual morphologic and kinetic features used in the KS (e.g., margins, root sign, enhancement pattern) to determine their predictive value for malignancy.

 

MissingLinkText

 

MissingLinkText

DCE-MRI Protocol

All MRI studies were performed on a 3-T Siemens Magnetom Vida scanner, using an 8-channel dedicated breast coil. The standard protocol included T2-weighted (T2w)/T1-weighted (T1w) imaging (without fat saturation), Short tau inversion recovery (STIR), diffusion-weighted imaging (DWI) at 50/800, T1w precontrast and T1 weighted flash 3d Spectral Attenuated Inversion Recovery (T1 fl3D SPAIR)-dynamic images. Postprocessing included maximum intensity projection (MIP), peak enhancement images for region of interest (ROI) placement, and SI-time curves.

 

MissingLinkText

Image Analysis

Two readers with 14 (R1) and 3 years (R2) of experience in breast imaging blinded to the radiological and histopathology reports studied all the DCE-MRI images retrospectively. The BI-RADS category according to the fifth version of the ACR-BI-RADS lexicon was labeled to lesions on the basis of malignancy suspicion. Furthermore, the KS using the KS flowchart was assigned to each lesion and then the KS was translated into Kaiser BI-RADS (KB) category designations shown in [Fig. 1].

Fig. 1 Illustration of the Kaiser score flowchart.

It includes five morphologic and kinetic criteria (root sign, lesion margin, SI-time curve type, pattern of internal enhancement, and edema). The first parameter that was evaluated was the presence of the root sign. If it was present, then the SI-time curve was evaluated. If the curve showed plateau or washout kinetics, then the next parameter assessed was perifocal edema. If the root sign was absent, then the curves were assessed again followed by margins and internal enhancement pattern in a sequential manner.

The allotted KS can be extrapolated to BI-RADS categories, with KSs of 1 to 4 considered BI-RADS 2/3, 5 to 7 considered BI-RADS 4, and 8 to 11 considered BI-RADS 5. Examples of the KS calculation and ACR-BI-RADS assignment are depicted in [Figs. 2] [3] [4].

Fig. 2 (a) Early and (b) delayed postcontrast T1-weighted sequences, (c) T2w sequence and (d) signal intensity–time curve and flowchart revealed a lesion in the left breast of a 34-year-old woman with an absent root sign, a persistent enhancement curve type with well-circumscribed margins that corresponds to a Kaiser score of 1 and BIRADS 3. The patient asked for surgery, which revealed fibroadenoma on histopathology.

Fig. 3 (a) Early and (b) delayed postcontrast T1-weighted sequences, (c) Short tau inversion recovery (STIR) sequence, and (d) signal intensity–time curve and flowchart revealed a lesion in the right breast of a 55-year-old woman showing a root sign, a washout enhancement curve type with perifocal edema that corresponds to a Kaiser score of 11 and BIRADS 5. Histopathological diagnosis revealed invasive ductal carcinoma.

Fig. 4 (a) Early and (b) delayed postcontrast T1-weighted sequences, (c) Short tau inversion recovery (STIR) sequence, and (d) signal intensity–time curve and flowchart revealed a lesion in the left breast of a 43-year-old woman showing nonmass enhancement (NME) with a root sign, a plateau enhancement curve type with perifocal edema that corresponds to a Kaiser score of 10 and BIRADS 4. Histopathological diagnosis revealed invasive breast carcinoma.

 

MissingLinkText

Statistical Analysis

Statistical analysis was done using SPSS version 21.0. Descriptive statistics of the data were given as mean, standard deviation, median, minimum, maximum, frequency, and percentage. The ROC analysis and the AUROC were obtained to evaluate the diagnostic ability. The KSs were divided into positive (cutoff score > 4) and negative scores (cutoff score ≤4). BI-RADS was similarly dichotomized at cutoff category greater than 3. In addition, sensitivity, specificity, PPV, NPV, and positive and negative likelihood ratio were calculated.

Interobserver agreement using Cohen's kappa coefficient was calculated with values (κ) interpreted as poor (<0.20), fair (0.210.40), moderate (0.410.60), good (0.610.80), and excellent (>0.81). Univariate analysis was used for each of the individual characteristics. A p-value of ≤0.05 was regarded as significant.

 

MissingLinkText

Ethical Approval

This study was conducted in accordance with the Declaration of Helsinki. The institutional review board of our hospital (Nos. RGCIRC/IRB-BHR/27/2024) approved the study protocol on March 27, 2024, and all regulations were followed.

 

MissingLinkText

 

MissingLinkText

Results

Patient and Lesion Characteristics

The mean age for all the lesions was 37.3 ± 8.1 years with a mean size of 24.8 ± 12.7 mm. Out of a total of 142 lesions, 48 were benign (33.8%) and 94 were malignant (66.20%). Of all the benign lesions, 39 (81.2%) were mass lesions and 9 were nonmass lesions (18.2%). The most common benign lesion was fibroadenoma (n = 19/48, 41%; [Fig. 2]); the remaining included fibrocystic disease, infection/inflammation, papillomas, and one case of adenomyoepithelioma. Of all malignant lesions (n = 94, 66.2%), mass lesions were more common (n = 78/94, 82.9%; [Fig. 3]). The most common mass as well as nonmass lesions in the malignant category were infiltrating ductal carcinomas (n = 85/94, 90.4%); the remaining included ductal carcinoma in situ (n = 4/94), lobular carcinoma in situ (n = 2/94), atypical ductal hyperplasia (n = 1/94), lobular carcinoma (n = 1/94), and invasive papillary carcinoma (n = 1/94; [Fig. 4]).

 

MissingLinkText

Diagnostic Accuracy

The AUROC of the KS for all the lesions, mass lesions, and nonmass lesions were 0.895, 0.955, and 0.622, respectively, implying that it has the diagnostic ability to differentiate between benign and malignant lesions ([Table 1] and [Fig. 5]). The AUC of the KS is more than that of ACR-BI-RADS for mass as well as nonmass lesions ([Table 1]).The overall sensitivity and specificity of the KS at the cutoff of greater than 4 were 93.6 and 85.4%, respectively, while that of ACR BI-RADS (>3) were 97.9 and 62.5%, respectively ([Table 2]). Sensitivity of the KS is similar to that BI-RADS for masses (98.7%), while specificity of the KS was higher (92.3%) for both readers as compared with that of BI-RADS (74.4%).For nonmass lesions, sensitivity of the KS was lower (68.8%) than that of BI-RADS (100%); however, specificity was higher (55.6% for KS vs. 11.1% for BI-RADS).

Table 1 Area under ROC curve of Kaiser score and ACR BI-RADS for all the lesions, mass (M) and nonmass (NM) lesions

	All lesions		Mass		Nonmass
	AUC (95% CI)		AUC (95% CI)		AUC (95% CI)
Kaiser score	0.895 (0.830–0.960)	0.001* Significant	0.955 (0.904–1.000)	0.001* Significant	0.622 (0.386–0.857)	0.322 Nonsignificant
ACR BI-RADS	0.802 (0.713–0.890)	0.001* Significant	0.859 (0.773–0.945)	0.001* Significant	0.556 (0.310–0.801)	0.651 Nonsignificant

Abbreviations: ACR BI-RADS, American College of Radiology Breast Imaging Reporting and Data System; AUC, area under ROC; CI, confidence interval; ROC, receiver operating characteristic.

Fig. 5 Area under ROC curve of the Kaiser score and ACR BI-RADS for all the lesions, mass (M) lesions, and nonmass lesions (NM). AUC, area under ROC; ROC, receiver operating characteristic.

Table 2 Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) for all the lesions, mass (M) lesions, and nonmass (NM) lesions with inter-reader agreement using K coefficient

	Readers		Cohen's kappa (k)	Sensitivity	Specificity	PPV (%)	NPV (%)
All lesions	R1	ACR BI-RADS	0.980	98.9 (93/94)	62.5 (30/48)	83.8	96.8
	R2		0.001* Significant	97.9 (92/94)	62.5 (30/48)	83.6	93.8
	R1	Kaiser score	0.969	93.6 (88/94)	85.4 (41/48)	92.6	87.2
	R2		0.001* Significant	91.5 (86/94)	85.4 (41/48)	92.5	83.7
Mass lesions	R1	ACR BI-RADS	0.978	98.7 (77/78)	74.4 (29/39)	88.5	96.7
	R2		0.001* Significant	97.4 (76/78)	74.4 (29/39)	88.4	93.5
	R1	Kaiser score	0.961	98.7 (77/78)	92.3 (36/39)	96.3	97.3
	R2		0.001* Significant	96.2 (75/78)	92.3 (36/39)	96.2	92.3
Nonmass lesion	R1	ACR BI-RADS	1.000	100 (16/16)	11.1 (1/9)	66.7	100
	R2		0.001* Significant	100 (16/16)	11.1 (1/9)	66.7	100
	R1	Kaiser score	1.000	68.8 (11/16)	55.6 (5/9)	73.3	50
	R2		0.001* Significant	68.8 (11/16)	55.6 (5/9)	73.3	50

Abbreviation: ACR BI-RADS, American College of Radiology Breast Imaging Reporting and Data System.

The overall PPV and NPV of the KS were 92.6 and 87.2%, respectively, while those of ACR BI-RADS were 83.8 and 96.8%, respectively ([Table 2]). The PPV of the KS was higher than of BI-RADS for masses (96.3%), while the NPV of the KS was almost similar (97.3%) to that of BI-RADS (96.7%). For nonmass lesions, the PPV of the KS was higher (73.3%) than that of BI-RADS (66.7%); however, the NPV was almost 50% less (50% for KS vs. 100% for BI-RADS).

Interobserver Agreement

There was strong interobserver agreement with kappa values of greater than 0.9 for mass lesions, nonmass lesions, and all lesions ([Table 2]).

 

MissingLinkText

Univariate Analysis

Univariate analysis for each of the five individual parameters revealed that irregular margin was the most important parameter with an odds ratio of 37, meaning an irregular lesion was 37 times more likely to be malignant. The rest of the parameters in isolation proved to be insignificant.

 

MissingLinkText

 

MissingLinkText

 

MissingLinkText

Discussion

DCE-MRI has been used for breast cancer detection since its approval by the Food and Drug Administration (FDA) in 1991.[15] It has a better sensitivity for detection of cancer as compared with mammography and other conventional imaging methods (>90%); however, its specificity is low and variable at 64.7% (range: 43.7–85.7%).[16] DCE-MRI aids in the detection of multifocal, multicentric, or synchronous disease in the breast.[5] [17] Plana et al[17] concluded that DCE-MRI showed high diagnostic accuracy but should be pathologically verified because of the high false-positive rate. Similarly, Houssami et al[5] stated that DCE-MRI leads to more extensive breast surgery due to false-positive observations as additional foci of cancers. So, to improve the diagnostic accuracy of DCE-MRI breast, a three-step tree flowchart was introduced by combining five of the BI-RADS descriptors in the sequential manner to objectively arrive at a clinical decision by Baltzer et al.[11] [12] [13] This flowchart was later called the KS. ACR BI-RADS gives us a standardized lexicon that can be used to describe each of the abnormalities, but there is a lacuna in category assignment, especially for those with limited experience.[7] [8] [18]

In our study, we found that the AUROC for the KS for mass and nonmass lesions was significant, indicating good diagnostic accuracy. It was also observed that the AUROC for the KS was greater than BI-RADS for mass as well as nonmass lesions, indicating that it performs better than ACR BI-RADS. These findings are in concordance with studies validating the KS.[13] [18] [19] [20] We also found a strong interobserver agreement for the KS in the diagnosis of all benign and malignant lesions, in mass or nonmass lesions, with a p-value of 0.001. These findings are in concordance to those of Marino et al..[18]

Multiparametric MRI assessment using ACR BI-RADS has been reported to have a sensitivity and specificity 100 and 12%, respectively, and the corresponding KS values were 98.5 and 34.8%, respectively, for R1 and 98.7 and 47.5%, respectively, for R2 in a study done by Istomin et al.[21] They concluded that the KS exhibited high diagnostic accuracy with outstanding interobserver reproducibility. We also discovered near similar findings in our study with sensitivity of 93.6 and 91.5% and specificity of 85.4% using the KS, which was better than sensitivity of 97.9 and 98.9% and specificity of 62.5% using ACR BI-RADS for R1 and R2, respectively.

There are only a few studies describing the application of the KS in characterization of mass as well as nonmass lesions in the diagnostic setting using DCE-MRI.[22] The AUROC for the KS in a study done by Makboul et al were 0.985, 0.997, and 0.813 for all lesions, mass lesions, and nonmass lesions, respectively, while in our study these were 0.895, 0.955, and 0.622 for all lesions, mass lesions, and nonmass lesions, respectively. In nonmass lesions, ACR-BI-RADS have shown a better sensitivity but poorer specificity as compared with the KS. However, the KS showed better specificity in comparison to ACR-BI-RADS in nonmass lesions, which is contradictory to findings in the study by Makboul et al.[22] Thus, the KS can be used to identify true negatives among nonmass lesions and further downgrade the lesions to escape unnecessary biopsy. However, it should be noted that the KS can miss some malignancies in nonmass lesions due to its poorer sensitivity as compared with ACR BI-RADS. Thus, the KS should not be used in clinical decision-making in symptomatic patients with nonmass lesions. The nonmass lesions in which the KS performed poorly in our study were ductal carcinoma in situ, lobular carcinoma in situ, and lobular carcinomas.

While evaluating mass lesions, the sensitivity of the KS and ACR BI-RADS was almost similar; however, the major advantage that the KS offered was an enhancement in specificity by approximately 18% for mass lesions. These findings are similar to those reported by Makboul et al,[22] who reported an increase in specificity by approximately 17.8% with the use of the KS. We found that approximately 22.9% patients, for all lesions, could be downgraded and treated conservatively. These findings are similar to those of Istomin et al who downgraded 22.8 and 35.4% lesions by two readers.[21]

This study had several limitations. The relatively small sample size limited the statistical power and generalizability of the findings. Additionally, there was significant variability in patient presentation, with symptoms ranging from nipple discharge and palpable breast or axillary masses to pain and heaviness, resulting in heterogeneity that made meaningful subgroup analysis challenging.

Future studies should aim to address these limitations through multicenter collaborations with larger, more diverse patient cohorts. Integration of artificial intelligence and machine learning tools could further enhance diagnostic accuracy and standardization. Longitudinal studies assessing long-term outcomes are also warranted to better inform clinical decision-making and patient management.

In conclusion, our study validates that KS is easy to apply, delivers high diagnostic accuracy with excellent interobserver correlation, and adds objectivity to assessment. Application of KS improves the specificity of DCE-MRI for both mass and nonmass lesions. For masses, sensitivity is the same as ACR BI-RADS; however, specificity is higher. Thus, it can potentially downgrade lesions. In nonmass lesions, the improvement in specificity comes with a reduction in sensitivity. We feel that one should refrain from applying the KS in nonmass lesions.

 

MissingLinkText

 

MissingLinkText

Conflict of Interest

None declared.

Acknowledgments

The authors would like to thank the patients and their families for their munificence in contributing to this study. We would also like to thank all members of the IRB committee who gave their approval for this study.

Patient's Consent

Patient consent is not required due to the retrospective nature of the study.

• References

• 1 Baltzer PA, Schelhorn J, Dietzel M, Kaiser WA. Breast screening programs using MRI: is there a role for computer-aided diagnosis?. Imaging Med 2010; 2 (06) 659-673
• 2 Mann RM, Kuhl CK, Kinkel K, Boetes C. Breast MRI: guidelines from the European Society of Breast Imaging. Eur Radiol 2008; 18 (07) 1307-1318
• 3 American College of Radiology. ACR Practice Parameter for the Performance of Contrast-Enhanced MRI of the Breast. Reston, VA: American College of Radiology; 2008
• 4 Mann RM, Cho N, Moy L. Breast MRI: state of the art. Radiology 2019; 292 (03) 520-536
• 5 Houssami N, Ciatto S, Macaskill P. et al. Accuracy and surgical impact of magnetic resonance imaging in breast cancer staging: systematic review and meta-analysis in detection of multifocal and multicentric cancer. J Clin Oncol 2008; 26 (19) 3248-3258
• 6 Sickles EA. ACR BI-RADS® Atlas, Breast Imaging Reporting and Data System. Reston, VA: American College of Radiology; 2013: 39
• 7 Strigel RM, Burnside ES, Elezaby M. et al. Utility of BI-RADS assessment category 4 subdivisions for screening breast MRI. AJR Am J Roentgenol 2017; 208 (06) 1392-1399
• 8 Dijkstra H, Dorrius MD, Wielema M, Pijnappel RM, Oudkerk M, Sijens PE. Quantitative DWI implemented after DCE-MRI yields increased specificity for BI-RADS 3 and 4 breast lesions. J Magn Reson Imaging 2016; 44 (06) 1642-1649
• 9 Marino MA, Helbich T, Baltzer P, Pinker-Domenig K. Multiparametric MRI of the breast: a review. J Magn Reson Imaging 2018; 47 (02) 301-315
• 10 Meng L, Zhao X, Guo J. et al. Evaluation of the differentiation of benign and malignant breast lesions using synthetic relaxometry and the Kaiser score. Front Oncol 2022; 12: 964078
• 11 Baltzer PA, Dietzel M, Kaiser WA. A simple and robust classification tree for differentiation between benign and malignant lesions in MR-mammography. Eur Radiol 2013; 23 (08) 2051-2060
• 12 Dietzel M, Baltzer PAT. How to use the Kaiser score as a clinical decision rule for diagnosis in multiparametric breast MRI: a pictorial essay. Insights Imaging 2018; 9 (03) 325-335
• 13 Woitek R, Spick C, Schernthaner M. et al. A simple classification system (the tree flowchart) for breast MRI can reduce the number of unnecessary biopsies in MRI-only lesions. Eur Radiol 2017; 27 (09) 3799-3809
• 14 Baltzer A, Dietzel M, Kaiser CG, Baltzer PA. Combined reading of contrast enhanced and diffusion weighted magnetic resonance imaging by using a simple sum score. Eur Radiol 2016; 26 (03) 884-891
• 15 Kaiser WA, Zeitler E. MR imaging of the breast: fast imaging sequences with and without Gd-DTPA. Preliminary observations. Radiology 1989; 170 (3, Pt 1): 681-686
• 16 Aristokli N, Polycarpou I, Themistocleous SC, Sophocleous D, Mamais I. Comparison of the diagnostic performance of magnetic resonance imaging (MRI), ultrasound and mammography for detection of breast cancer based on tumor type, breast density and patient's history: a review. Radiography (Lond) 2022; 28 (03) 848-856
• 17 Plana MN, Carreira C, Muriel A. et al. Magnetic resonance imaging in the preoperative assessment of patients with primary breast cancer: systematic review of diagnostic accuracy and meta-analysis. Eur Radiol 2012; 22 (01) 26-38
• 18 Marino MA, Clauser P, Woitek R. et al. A simple scoring system for breast MRI interpretation: does it compensate for reader experience?. Eur Radiol 2016; 26 (08) 2529-2537
• 19 Milos RI, Pipan F, Kalovidouri A. et al. The Kaiser score reliably excludes malignancy in benign contrast-enhancing lesions classified as BI-RADS 4 on breast MRI high-risk screening exams. Eur Radiol 2020; 30 (11) 6052-6061
• 20 Jajodia A, Sindhwani G, Pasricha S. et al. Application of the Kaiser score to increase diagnostic accuracy in equivocal lesions on diagnostic mammograms referred for MR mammography. Eur J Radiol 2021; 134: 109413
• 21 Istomin A, Masarwah A, Vanninen R, Okuma H, Sudah M. Diagnostic performance of the Kaiser score for characterizing lesions on breast MRI with comparison to a multiparametric classification system. Eur J Radiol 2021; 138: 109659-109659
• 22 Makboul M, Farghaly S. MRI diagnostic performance and inter-observer agreement of Kaiser score in the assessment of different breast lesions. Egypt J Hosp Med 2022; 88 (01) 4068-4075

Publikationsverlauf

Artikel online veröffentlicht:
10. Juli 2025