New research suggests that an updated approach to Kaiser scoring may bolster the detection of breast cancer in enhancing lesions on breast magnetic resonance imaging (MRI)
For the study, recently published in Academic Radiology, researchers compared the traditional Kaiser score with a new breast lesion diagnostic model that adds the parameters of age, clinical breast exam (CBE) results, maximum intensity projection (MIP) sign and associated imaging features to the Kaiser score. The study authors defined positive findings for associated imaging features as one or more of the following findings: axillary lymph node enlargement, architectural distortion, nipple retraction, nipple invasion, skin thickening, skin retraction, chest wall invasion, skin invasion and/or pectoralis muscle invasion.
The cohort was comprised of 284 patients in the training set (average age of 48) who underwent preoperative breast dynamic contrast-enhanced MRI (DCE-MRI) exams. There were 107 patients in the validation set, according to the study.
The researchers found that the breast lesion diagnostic model with the added parameters had a 94.8 percent AUC in the training set in comparison to 86.9 percent for the traditional Kaiser score. Similar differences were noted in the validation set with a 95.6 percent AUC for the enhanced Kaiser scoring model vs. 87.9 for traditional Kaiser scoring.
“This model can be used for preoperative assessment of the malignant probability of enhancing lesions on breast MRI and shows better diagnostic performance compared to the classic Kaiser score,” wrote Xi Yi, M.D., who is affiliated with the Department of Radiology at the Hunan Provincial People’s Hospital and the First Affiliated Hospital of Hunan Normal University in Changsha, China, and colleagues.
In the validation set, the study authors found that the enhanced model with the enhanced Kaiser model achieved a sensitivity rate of 91.18 percent, a specificity rate of 89.04 percent, an accuracy rate of 89.72 percent and a negative predictive value (NPV) of 95.59 percent.
In regard to the aforementioned added parameters to Kaiser scoring, the researchers pointed out that palpable CBE findings were found in 88.37 percent of mass presentations while associated imaging features were absent in 85.71 percent of non-mass presentations.
“These four new parameters are easily accessible in clinical practice and do not require additional examination costs, while they can optimize the diagnostic performance of the Kaiser score, making them highly usable,” maintained Yi and colleagues.
Three Key Takeaways
1. Enhanced diagnostic accuracy. The new breast lesion diagnostic model, which integrates additional parameters such as age, clinical breast exam results, and specific imaging features, significantly improves the detection of malignant lesions on breast MRI, showing a higher AUC (area under curve) compared to the traditional Kaiser score.
2. Increased application for clinical use. The added parameters in the enhanced model are accessible in routine clinical practice without extra costs. This may make the updated Kaiser scoring model more practical and potentially beneficial for preoperative assessments.
3. Implications for treatment planning. This improved diagnostic model may provide valuable insights into the risk of upgrade in cases diagnosed as ductal carcinoma in situ (DCIS) via core needle biopsy, perhaps facilitating improved preoperative diagnosis and treatment strategies.
Emphasizing that the additional parameters to Kaiser scoring were derived from a multivariate logistic regression analysis, the researchers suggested that this new model may offer additional data beyond the use of core needle biopsy (CNB) in treatment decision-making.
“The predictive model constructed in this study can provide insights into the risk of upgrade for cases diagnosed as (ductal carcinoma in situ) via preoperative CNB, thereby assisting in the formulation of preoperative diagnosis and treatment schemes,” noted Yi and colleagues.
(Editor’s note: For related content, see “Breast MRI and Dense Breasts: A Closer Look at Early Findings from a New Prospective Trial,” “Can Multimodal AI Enhance Prediction of Axillary Lymph Node Metastasis Beyond MRI or Ultrasound-Based Models?,” and “Use of Preoperative MRI 46 Percent Less Likely for Black Women with Breast Cancer.”)
In regard to study limitations, the authors noted the cohort was entirely comprised of Chinese women and accordingly cautioned against extrapolation of the results to a broader population. They also suggested possible bias in the results due to validation set samples drawn from one facility.
