Kikuchi–Fujimoto Disease:

Look For First

Multiple small-to-moderately sized cervical nodes (mean 1.62 cm, mostly <2.5 cm) at levels II–V, unilateral or asymmetric, with homogeneous enhancement on CT/MRI
Disproportionate perinodal inflammatory infiltration and fat edema relative to the modest nodal size—this is the key recognition clue
Absence of coarse calcification, drainable abscess, or convincing primary malignancy, with preserved echogenic hila on ultrasound

Key Image Findings

On CT, the dominant phenotype is multiple homogeneously enhancing cervical nodes (83.3% in the largest series) with surrounding fat infiltration (81.3%), and nodes typically measure 0.5–3.5 cm with mean 1.62 cm. Unilateral or asymmetric involvement of levels II–V is characteristic, with unilateral disease in 76 of 96 biopsy-proven cases.
Perinodal infiltration ranges from subtle fat-plane obliteration to extensive reticulation and increased attenuation within the cervical space, and this finding is particularly important because it is disproportionately prominent relative to modest nodal size—a major pattern-recognition feature.
Macroscopic necrosis is visible on CT in only a minority (16.7% in the large series) and appears as multiple small, indistinct low-attenuation foci rather than the smooth, sharply defined central caseation and thick peripheral rim seen with tuberculosis; ringlike enhancement can occur but should raise concern for suppurative adenitis or metastatic disease.
Nodal calcification is not characteristic of KFD and should favor alternative diagnoses such as tuberculosis, treated disease, or thyroid carcinoma metastasis.
On ultrasound, multiple hypoechoic cervical nodes with relatively uniform size and preserved echogenic hilum are common, with one pediatric series showing preserved central hila in 98% and another finding echogenic hila in 90.4% of KFD versus only 24.7% of tuberculous nodes; perinodal fat swelling reflects inflammation.
On MRI, findings parallel CT with multiple cervical nodes showing variable homogeneous enhancement, focal nonenhancing necrosis in some nodes, and conspicuous edema or inflammatory signal in adjacent fat; nodes may restrict diffusion due to hypercellularity and necrotizing material, making diffusion restriction unreliable for distinguishing KFD from lymphoma.
On FDG PET/CT, KFD nodes can be intensely avid with reported SUVmax values of 8.3–22.5 that overlap with lymphoma, and one series found no significant separation between KFD and malignant lymphoma, making PET/CT unreliable for declaring KFD benign by uptake intensity alone.
Young age, fever, tender nodes, multiple relatively uniform smaller nodes, unilateral or asymmetric level II–V distribution, homogeneous enhancement, prominent perinodal fat infiltration, absence of mucosal or thyroid primary, and lack of gas or drainable collection all comprise a favorable constellation of features that should raise KFD in the differential.

Differential Diagnosis

Lymphoma: multiple homogeneous nodes with restricted diffusion and intense FDG uptake can mimic KFD, but lymphoma typically presents with larger bulky or conglomerate nodes, broader systemic distribution, less tenderness, organ or extranodal lesions, and clonal pathology; SUV intensity alone is unreliable for distinction.
Tuberculous lymphadenitis: shares posterior cervical distribution, necrosis, perinodal infiltration, and systemic symptoms, but tuberculosis typically shows smooth central caseation, thick rim enhancement, calcification, fistula or sinus formation, greater loss of echogenic hilum, and relevant exposure or positive microbiology; KFD more often has small indistinct necrotic foci without calcification.
Suppurative bacterial adenitis: presents with tender nodes, fever, perinodal edema, and low attenuation similar to KFD, but is distinguished by marked leukocytosis, adjacent infectious source, cellulitis, gas, sharply liquefied abscess, and drainable collection; neutrophils dominate pathologically rather than being absent.
Metastatic squamous cell carcinoma: can present with necrotic cervical nodes and surrounding fat infiltration, but favors older or high-risk patients, shows mucosal primary, demonstrates irregular thick-walled cystic nodes with extranodal extension, and contains metastatic epithelial cells; a cystic level II node in an adult still requires HPV-associated oropharyngeal SCC exclusion.
SLE lymphadenitis: shares fever, cytopenia, necrotizing lymphadenitis, karyorrhexis, and paucity of neutrophils with KFD, but is distinguished by ANA, anti-dsDNA, anti-Sm positivity, low complement, renal or cutaneous findings, hematoxylin bodies, Azzopardi phenomenon, and more abundant plasma cells; KFD can precede or follow SLE.
Cat-scratch disease and viral/reactive adenitis: can present with fever and necrotizing or small homogeneous tender nodes with preserved hila, but cat-scratch disease shows exposure history, inoculation site, stellate suppurative granulomas, and positive testing, while viral adenitis shows identifiable viral syndrome and resolution without characteristic necrotizing histology.

Discussion

KFD is histologically a necrotizing lymphadenitis characterized by patchy paracortical coagulative necrosis, abundant karyorrhectic nuclear debris ('nuclear dust'), crescentic histiocytes at necrosis edges, plasmacytoid dendritic cells, predominantly CD8-positive T-cell background, and conspicuous absence or marked paucity of neutrophils—a combination that distinguishes it from suppurative adenitis.

The cause of KFD remains uncertain, but proposed mechanisms include an exaggerated T-cell and histiocytic immune response to an infectious or other antigenic trigger in a genetically susceptible host; KFD can precede, coincide with, or follow SLE, warranting autoimmune assessment at diagnosis and clinical surveillance for lupus manifestations.

Imaging cannot establish KFD; diagnosis requires clinicopathologic correlation and usually excisional lymph-node biopsy to preserve architecture and reliably separate KFD from lymphoma and SLE lymphadenitis, whereas fine-needle aspiration can suggest KFD but is less definitive.

Most KFD resolves spontaneously within 1–6 months with supportive care (rest, hydration, antipyretics, NSAIDs); corticosteroids are reserved for severe, persistent, recurrent, generalized, or extranodal disease after infection exclusion, and recurrence occurs in approximately 12.6% in contemporary pediatric data.

The key pattern-recognition feature for radiologists is disproportionately prominent perinodal inflammatory infiltration relative to modest nodal size—the soft-tissue component often appears more significant than the lymph nodes themselves would predict, setting KFD apart from reactive nodes and reactive-appearing disease.

Rare severe complications including neurologic disease, myocarditis, pulmonary involvement, and hemophagocytic lymphohistiocytosis can occur, so persistent high fever, progressive cytopenias, organ dysfunction, or marked hyperferritinemia warrants urgent reassessment rather than attribution to uncomplicated KFD.

Reporting Pearls

When KFD is suspected, describe the nodal distribution precisely (laterality, levels, number, and largest short-axis dimension), characterize morphology (homogeneous vs. necrotic, enhancement pattern, hilar preservation, calcification), search explicitly for alternative sources (oropharynx, thyroid, primary malignancy), and use language such as "Kikuchi–Fujimoto disease is a consideration in the appropriate clinical setting" while retaining lymphoma, infection, SLE, and metastasis as differentials; recommend tissue sampling in adults, with atypical age, bulky or progressive disease, necrosis, or absent convincing clinical context, targeting solid enhancing nodal components rather than necrotic centers for biopsy.

Pitfalls

High FDG uptake is a major pitfall: KFD nodes can be intensely avid with SUVmax values of 8.3–22.5 that completely overlap with lymphoma, making PET/CT unreliable for excluding malignancy; tissue diagnosis remains necessary when lymphoma is plausible.
Focal necrosis or ringlike enhancement can strongly mimic suppurative adenitis, tuberculosis, or necrotic metastases, but KFD necrosis typically consists of multiple small, indistinct low-attenuation foci without the smooth central caseation, thick rim enhancement, and calcification characteristic of tuberculosis.
Calcification, coarse or punctate, is not characteristic of KFD and should shift the differential away from uncomplicated KFD toward tuberculosis, treated disease, or thyroid carcinoma metastasis.
Prominent immunoblasts, necrosis, and high proliferative appearance can lead to erroneous lymphoma diagnosis on histology unless the characteristic histiocytic and karyorrhectic pattern with paucity of neutrophils is recognized; excisional biopsy preserves architecture better than fine-needle aspiration for reliable distinction.

Kikuchi–Fujimoto Disease:

Key Images

Secondary Links