Diagnostic performance of Kaiser score in patients with newly diagnosed breast cancer: Factors associated with false-negative results

Abstract

PurposeTo investigate the factors associated with false-negative results in the diagnosis of breast cancer via breast magnetic resonance imaging (MRI) using the Kaiser score (KS). MethodsThis institutional review board (IRB)-approved, single-center, retrospective study enrolled 219 consecutive histopathologically proven breast cancer lesions in 205 women who underwent preoperative breast MRI. Two breast radiologists evaluated each lesion according to the KS. The clinicopathological characteristics and imaging findings were also analyzed. Interobserver variability was assessed using the intraclass correlation coefficient (ICC). Multivariate regression analysis was used to investigate factors associated with false-negative KS results for breast cancer diagnosis. ResultsOf 219 breast cancers, KS yielded 200 (91.3%) true-positive and 19 (8.7%) false-negative results. The interobserver ICC for the KS between the two readers was good, with a value of 0.804 (95% CI 0.751–0.846). Multivariate regression analysis revealed that small lesion size (≤1 cm) (adjusted OR 6.86, 95% CI 2.14–21.94, p = 0.001) and personal breast cancer history (adjusted OR 7.59, 95% CI, 1.55–37.23, p = 0.012) were significantly associated with false-negative KS results. ConclusionSmall lesion size (≤1 cm) and presence of personal breast cancer history are factors significantly associated with false-negative KS results. Our results suggest that radiologists should consider these factors in clinical practice as potential pitfalls of KS, which may be compensated for by a multimodal approach combined with clinical evaluation.