S Lee1, J H Yoo, S W Lee. 1. Department of Radiology, Ewha Womans University Mokdong Hospital, Ewha Womans University School of Medicine, Seoul, Korea.
Abstract
BACKGROUND AND PURPOSE: KD, also known as histiocytic necrotizing lymphadenitis, is often mimicked by TL, especially when KD is accompanied by gross necrosis on CT. We analyzed the nodal necrosis patterns on CT in patients with KD and TL to identify the predictive features that differentiate these 2 entities. MATERIALS AND METHODS: We included 24 patients with KD (14 females, 10 males; mean age, 25.1 years) and 45 with TL (36 women, 9 men; mean age, 39.9 years) having nodal necrosis on CT. The extent, number, location, and margins of nodal necrosis; the presence of perinodal infiltration; and nodal calcification were analyzed. Two attenuation indices for nodal necrosis were derived: CTN and CTN/M. The CT features of the 2 diseases were compared, and multivariate logistic regression analysis was performed to identify predictive features that would aid in differentiating KD from TL. RESULTS: Mild or moderate extent of nodal necrosis (P = .000), multiple necrotic foci (P = .002), indistinct margins of necrotic foci (P = .000), CTN of >44.5 (P = .000), CTN/M of >0.7 (P = .000), and absence of nodal calcification (P = .012) were associated with KD. Multivariate logistic regression analysis revealed that indistinct margins of necrotic foci were independent predictors, with 80% accuracy. CONCLUSIONS: The presence of indistinct margins of necrotic foci was an independent predictor in differentiating KD from TL with high diagnostic accuracy.
BACKGROUND AND PURPOSE: KD, also known as histiocytic necrotizing lymphadenitis, is often mimicked by TL, especially when KD is accompanied by gross necrosis on CT. We analyzed the nodal necrosis patterns on CT in patients with KD and TL to identify the predictive features that differentiate these 2 entities. MATERIALS AND METHODS: We included 24 patients with KD (14 females, 10 males; mean age, 25.1 years) and 45 with TL (36 women, 9 men; mean age, 39.9 years) having nodal necrosis on CT. The extent, number, location, and margins of nodal necrosis; the presence of perinodal infiltration; and nodal calcification were analyzed. Two attenuation indices for nodal necrosis were derived: CTN and CTN/M. The CT features of the 2 diseases were compared, and multivariate logistic regression analysis was performed to identify predictive features that would aid in differentiating KD from TL. RESULTS: Mild or moderate extent of nodal necrosis (P = .000), multiple necrotic foci (P = .002), indistinct margins of necrotic foci (P = .000), CTN of >44.5 (P = .000), CTN/M of >0.7 (P = .000), and absence of nodal calcification (P = .012) were associated with KD. Multivariate logistic regression analysis revealed that indistinct margins of necrotic foci were independent predictors, with 80% accuracy. CONCLUSIONS: The presence of indistinct margins of necrotic foci was an independent predictor in differentiating KD from TL with high diagnostic accuracy.