Imaging Characteristics and Postoperative Outcomes of Head and Neck Unicentric Castleman Disease: A 20-Year Institutional Experience

Abstract

Purpose

Head and neck unicentric Castleman disease (HN-UCD) often lacks specific clinical and laboratory markers, leading to frequent preoperative misdiagnoses. While surgical resection typically offers a favorable prognosis, the management of atypical postoperative outcomes, such as new regional lymphadenopathy or persistent systemic symptoms, remains poorly defined. This study aimed to identify the characteristic imaging features of HN-UCD and explore effective management strategies for these atypical postoperative scenarios.

Methods

A retrospective analysis was conducted on 24 patients diagnosed with HN-UCD at the Peking Union Medical College Hospital over the past 20 years. The preoperative imaging characteristics and postoperative clinical outcomes were systematically evaluated.

Results

Imaging confirmed significant hypervascularity as a hallmark of HN-UCD: 94.4% (17/18) of lesions exhibited abundant and disordered blood flow signals on ultrasound, and 71.4% (10/14) of lesions presented with prominent feeding vessels on CT. Postoperatively, 87.5% (21/24) of the patients achieved complete remission within the surgical field. For the two patients exhibiting atypical outcomes (new regional lymphadenopathy and persistent fever), a strategy of rigorous follow-up was adopted after multidisciplinary exclusion of malignancy and infection; both patients remained clinically stable.

Conclusion

Prominent feeding vessels and marked hypervascularity are characteristic preoperative imaging markers for HN-UCD. For patients with atypical postoperative outcomes, a management protocol based on multidisciplinary evaluation and close clinical observation is recommended.

Keywords

feeding vessels head and neck imaging features outcome unicentric castleman disease

Introduction

Unicentric Castleman Disease (UCD) is a rare and heterogeneous lymphoproliferative disorder characterized by the pathological enlargement of a single lymph node station. UCD can affect individuals of all ages (ranging from 2 to 84 years), with a median age of 34 years at diagnosis.¹ The estimated incidence of UCD in the United States is 16–19 per million.² Histologically, UCD is primarily classified into three variants: the hyaline-vascular (HV) variant (79.0%), the plasma cell (PC) variant (13.1%), and the mixed subtype (7.9%).³ Topographically, UCD can occur in various anatomic regions. The mediastinum is the most common site of involvement (29%), followed by the neck (23%), abdomen (21%), and retroperitoneum (17%).⁴ Although it predominantly affects lymph nodes, rare extranodal involvement of parenchymal organs and soft tissues can occur, most frequently in the pancreas, adrenal glands, and skeletal muscle.⁵ Clinically, most UCD cases present as a painless mass; however, as the lesion expands, it may not only compress vital adjacent anatomical structures but also lead to complications such as autoimmune hemolytic anemia and autoimmune thrombocytopenia, which can severely impact the patient’s quality of life or even prove life-threatening.⁶ Consequently, early and precise diagnosis and intervention are critical for patients with UCD. Nevertheless, misdiagnosis remains common due to its tumor-like radiological appearance and clinical symptoms that often mimic infection.⁷ In this context, identifying characteristic preoperative imaging features is crucial for differential diagnosis and surgical planning.

Surgical resection remains the preferred therapeutic strategy for UCD and is generally associated with a favorable prognosis. A large-scale retrospective study involving 1,634 patients demonstrated a 3-year overall survival rate of 98.0%.³ However, the favorable overall prognosis can obscure specific clinical challenges: a minority of patients may experience postoperative local recurrence, new-onset lesions, or even persistent systemic inflammatory symptoms resembling multicentric Castleman disease (MCD), such as fever, fatigue, and weight loss.^8,9 Although some studies suggest that these symptoms often resolve spontaneously after surgery, a few patients exhibit refractory symptoms that necessitate medical intervention.^9,10 Due to the scarcity of relevant case reports, there are currently no clear clinical guidelines for managing such complex postoperative outcomes.

As noted above, the head and neck is a common site for UCD, accounting for approximately 23% of all cases.⁴ Specifically within this region, lesions typically arise in regional cervical lymph nodes but can also manifest in mucosal locations and major salivary glands.^5,11 However, specialized research on head and neck UCD (HN-UCD) remains limited, particularly regarding in-depth discussions on imaging characteristics and atypical postoperative outcomes. This study retrospectively reviewed the clinical data of patients with HN-UCD at Peking Union Medical College Hospital (PUMCH) over the past 20 years, focusing on summarizing the imaging features and postoperative outcomes. We aim to provide clinical insights and the “Chinese experience” to support the preoperative diagnosis and postoperative management of HN-UCD.

Materials and Methods

Patient Selection

This study was approved by the Institutional Review Board of PUMCH (Ethics No. K25C3638), and informed consent was obtained from all participants. Clinical data of patients pathologically diagnosed with HN-UCD between 2005 and 2025 were retrospectively retrieved. The inclusion criteria were as follows: (1) histopathological confirmation of Castleman disease; (2) disease localization to a single anatomical region in the head and neck, with involvement of other lymph node stations ruled out by preoperative systemic imaging; (3) exclusion of lymphoma, IgG4-related disease, systemic lupus erythematosus, and other autoimmune or infectious lymphadenopathies; and (4) availability of complete clinical and imaging data.

Data Collection and Imaging Assessment

Patient demographics, clinical manifestations, laboratory findings, surgical records, and pathological subtypes were collected from the electronic medical record system. Preoperative imaging data, including ultrasound, CT, MRI, and PET/CT, were systematically re-evaluated for all patients. We recorded the lesion size, location, borders, density/signal characteristics, enhancement patterns, the presence of necrosis or calcification, and the relationship with adjacent vasculature. Follow-up data were obtained from outpatient review records or telephone interviews. The primary observation endpoints included postoperative symptom resolution, local recurrence, and survival status.

Statistical Analysis

Statistical analysis was performed using R software (version 4.4.3). The normality of continuous variables was determined using the Shapiro-Wilk test. Normally distributed continuous data were expressed as mean ± standard deviation (Mean ± SD). Non-normally distributed data were presented as median (interquartile range) [M (IQR)]. Categorical variables were reported as frequencies and percentages [n (%)].

Results

Baseline Characteristics and Clinical Features

This study included 24 patients with HN-UCD, consisting of 12 males and 12 females, with a mean age of 34.8 ± 14.5 years. All 24 cases presented as regional nodal disease, most frequently involving cervical lymph node Level II (41.7%). A painless neck mass was the initial symptom in 75% (18/24) of cases. Beyond the presence of a mass, 20.8% (5/24) of patients presented with compressive symptoms, including ipsilateral upper limb weakness, numbness of the ipsilateral thumb, and snoring, while 4.2% (1/24) experienced recurrent fever. The median follow-up duration was 43 months (IQR: 66 months), with one patient lost to follow-up. Details are presented in Table 1.

Table 1.

Baseline Characteristics and Postoperative Outcomes of 24 Patients With HN-UCD

Characteristic	n/N	%
Age (years, Mean ± SD)	34.8 ± 14.5	-
Follow-up [months, Median (IQR)]	43(66)	-
Gender
Male	12/24	50
Chief Complaint
Painless neck mass	18/24	75
Upper limb weakness	2/24	8.3
Thumb numbness	1/24	4.2
Headache	1/24	4.2
Snoring	1/24	4.2
Fever	1/24	4.2
Comorbidities
Henoch-Schönlein Purpura	1/24	4.2
Pulmonary Sarcoidosis	1/24	4.2
Histological Subtype
Hyaline vascular	18/24	75
Plasma cell	0/24	0
Mixed cellularity	3/24	12.5
Unclassified	3/24	12.5
Laboratory Abnormalities
Anemia	1/24	4.2
Thrombocytosis	5/24	20.8
Hypoalbuminemia	0/21	0
Elevated hs-CRP	2/15	13.3
HIV-positive	0/24	0
Postoperative Outcomes
Complete remission	21/24	87.5
Persistent systemic symptoms (fever)	1/24	4.2
New regional lymphadenopathy	1/24	4.2
Local recurrence	0/24	0
Lost to follow-up	1/24	4.2

Note. Due to the retrospective nature of this study, baseline laboratory data were incomplete for certain variables. Specifically, serum albumin levels were available for 21 of the 24 patients, and hs-CRP levels were available for 15 patients.

Imaging Findings

Preoperatively, neck lymph node ultrasonography was performed in 75% (18/24) of patients, while 58.4% (14/24) underwent CT, 41.7% (10/24) underwent neck soft tissue MRI, and 16.7% (4/24) underwent PET/CT. Imaging confirmed that among all enlarged lymph nodes, 83.3% (20/24) were distributed in neck levels I–IV, 8.3% (2/24) were located in the parapharyngeal space, and 8.3% (2/24) extended from the upper mediastinum into the neck, with a mean maximum diameter of 5.17 ± 1.96 cm; specific imaging characteristics are detailed in Table 2.

Table 2.

Imaging Features of Patients With HN-UCD

Modality	Characteristic	Frequency
Ultrasound (n=18)	Hypoechoic mass	94.4%(17/18)
	Obscured corticomedullary differentiation	50%(9/18)
	Abundant internal vascularity	94.4%(17/18)
	Internal reticular pattern/echoes	27.8%(5/18)
	Calcification	5.6%(1/18)
CT (n=14)	Homogeneous density	50%(7/14)
	Rich vascular supply	71.4%(10/14)
	Avid enhancement	57.1%(8/14)
	Calcification	7.1%(1/14)
MRI (n=10)	Isointense on T1WI	60%(6/10)
	Hyperintense on T2WI	90%(9/10)
	Avid enhancement	80%(8/10)
	Peripheral or internal flow voids	30%(3/10)
PET/CT (n=4)	Mildly increased radiotracer uptake	100%(4/4)

On ultrasonography, 94.4% (17/18) of the enlarged lymph nodes appeared as hypoechoic masses with regular shapes, well-defined margins, and poorly defined corticomedullary junctions (Figure 1A). Among these, 27.8% (5/18) showed heterogeneous internal echogenicity, 16.7% (3/18) exhibited iso-to-hyperechoic or hyperechoic signals, and 22.2% (4/18) displayed linear or reticular echoes within the nodes. Color Doppler flow imaging (CDFI) revealed abundant and disorganized blood flow signals within the masses in 94.4% (17/18) of patients (Figure 1B). On CT, all lymph nodes appeared as round or oval soft-tissue density masses with clear margins; they exhibited relatively homogeneous density similar to surrounding muscles on plain scans and showed homogeneous or heterogeneous enhancement on contrast-enhanced scans. Rich vascular supply around the tumors was observed in 71.4% (10/14) of patients, with some showing tortuous and enlarged intra- or extranodal vessels (Figure 1C-E), while distinct internal calcification was noted in 7.1% (1/14). Neck soft tissue MRI showed well-defined tumor margins, with isointense or slightly prolonged T1 signals and long T2 signals; contrast-enhanced MRI demonstrated significant homogeneous enhancement in the arterial phase and slightly decreased enhancement in the venous phase. Flow-void signals were visible within or around the tumors in 30% (3/10) of patients (Figure 1F-H), and all three cases involving DWI sequences showed high signals. In the limited subset of patients who underwent PET/CT (n=4), imaging revealed mildly increased radioactivity uptake in the enlarged neck lymph nodes (Figure 1I-J), with a mean SUVmax of 4.03 ± 1.04 g/mL.

Figure 1.

Representative imaging findings of HN-UCD. (a-b) Ultrasound imaging revealed a hypoechoic lesion with abundant internal blood flow signals. (c-e) CT shows a soft tissue density mass with visible peripheral and internal vasculature. (f-h) MRI demonstrates isointense to slightly prolonged T1 signals and prolonged T2 signals with internal flow-void phenomena. (i-j) PET/CT shows mild metabolic activity within the lesion. Red arrows indicate external (extranodal) feeding vessels; blue arrows indicate internal (intranodal) feeding vessels

Treatment and Postoperative Outcomes

The surgical strategy was tailored to the extent and location of the lesions. Most patients (19/24, 79.2%) underwent complete local mass excision, while a subset of patients (5/24, 20.8%) received local mass excision combined with selective neck dissection to ensure thorough regional clearance. No severe perioperative complications, such as hemorrhage or infection, were observed, and no adjuvant radiotherapy or chemotherapy was administered postoperatively. A comprehensive patient-by-patient breakdown of demographics, laterality, specific lesion localizations, and individualized surgical procedures is meticulously detailed in Table 3. During the follow-up period, 87.5% (21/24) of the patients achieved complete remission within the surgical field. The detailed follow-up outcomes are summarized in Table 3. However, two cases exhibited atypical clinical outcomes that warranted further clinical attention (Figure 2).

Table 3.

Detailed Anatomical Localizations and Surgical Procedures of the 24 HN-UCD Patients

Patient no.	Gender	Age (years)	Laterality	Lesion localization (node Level)¹	Surgical procedure²	Postoperative outcome
1	M	60	R	Level II	Local mass excision	Complete remission
2	M	54	L	Level II	Local mass excision with selective neck dissection (Levels II-IV)	Complete remission
3	M	27	L	Level VII	Local mass excision	Complete remission
4	F	23	R	Level V	Local mass excision	Complete remission
5	M	24	R	Level III	Local mass excision with selective neck dissection (Levels II-IV)	Complete remission
6	F	37	R	Level IV	Local mass excision	Complete remission
7	M	24	L	Level II	Local mass excision	Complete remission
8	M	16	L	Level II	Local mass excision with selective neck dissection (Levels II-IV)	Complete remission
9	F	39	R	Level IV-2R	Local mass excision	Complete remission
10	M	52	L	Level IV	Local mass excision	Complete remission
11	F	46	R	Level VII	Local mass excision	Complete remission
12	F	28	L	Level II	Local mass excision	Complete remission
13	M	24	R	Level II	Local mass excision	Complete remission
14	F	35	L	Level III	Local mass excision with selective neck dissection (Levels II-V)	Complete remission
15	F	36	R	Level IV-2R	Local mass excision	Complete remission
16	M	52	R	Level III	Local mass excision	Complete remission
17	M	60	R	Level II	Local mass excision	Complete remission
18	F	53	R	Level II	Local mass excision with selective neck dissection (Levels II-V)	Complete remission
19	F	24	R	Level V	Local mass excision	New regional lymphadenopathy
20	M	16	L	Level IV	Local mass excision	Persistent systemic symptoms
21	F	27	L	Level II	Local mass excision	Lost to follow-up
22	F	26	L	Level IV-2R	Local mass excision	Complete remission
23	F	40	R	Level IV	Local mass excision	Complete remission
24	M	12	L	Level II	Local mass excision	Complete remission

Notes.

¹ Lesion location refers to the primary site of the dominant disease focus identified preoperatively and intraoperatively.

² Surgical procedure was determined based on the complete excision of the primary lesion and was further guided by intraoperative exploration to address any suspicious regional lymphadenopathy.

Figure 2.

Postoperative outcomes and surveillance protocols for patients with HN-UCD

Case of New Regional Lymphadenopathy

One patient underwent resection of lesions in the right cervical Levels II and V. A follow-up examination seven months postoperatively revealed a new enlarged lymph node (2.2 × 1.2 cm) at the junction of the right cervical Levels III and IV. Laboratory investigations showed no abnormalities, and ultrasound and PET/CT revealed no evidence of lymphadenopathy in other regions. Given the absence of subjective discomfort and the exclusion of adverse high-risk prognostic factors, an active surveillance strategy was adopted following thorough patient-clinician communication, and the patient’s condition remained stable.

Case of Persistent Systemic Symptoms

One patient, who primarily presented with recurrent fever preoperatively, experienced a febrile relapse four months after surgery following a period of normal body temperature. The fever was periodic, with peak temperatures fluctuating between 37.0°C and 38.0°C, and resolved spontaneously. Following multidisciplinary consultations with the Departments of Hematology and Infectious Diseases, comprehensive evaluations, including inflammatory markers, a full immunological profile, and PET/CT, revealed no evidence of malignancy, active infection, or autoimmune disease. This patient is currently under close follow-up.

Additionally, among the patients presenting with preoperative local compression symptoms (n=5), symptoms resolved completely in four cases, while one patient was lost to follow-up.

Discussion

Although the exact etiology of HN-UCD remains elusive, it may involve clonal expansion of lymph node stromal cells.^12,13 While preoperative laboratory tests in patients with HN-UCD often lack specificity, our study and existing literature confirm that characteristic imaging features are the primary clues for preoperative diagnosis, with the most prominent hallmark being extreme hypervascularity.^14,15 Disorganized blood flow signals on Doppler ultrasound and coarse, prominent feeding vessels on CT or MRI are consistent with previous findings.¹⁶ Furthermore, internal linear or reticular echoes and heterogeneous enhancement correspond pathologically to fibrous deposition, cystic degeneration, or necrosis.¹⁵ Some studies suggest these features are more common in lesions with a diameter > 5 cm,^17,18 aiding in the differentiation from lymphoma. Although lymphoma can also be hypervascular, it typically lacks coarse feeding vessels and often involves multiple or bilateral lymph node regions. In contrast, benign reactive lymph nodes usually exhibit absent or regular hilar blood flow.^16,19 Regarding calcification, while some scholars consider it a characteristic of UCD,²⁰ we observed it in only one patient. This differs from the 22.4%–24.4% incidence reported in the literature,^21-23 suggesting that calcification may not be a sensitive marker for HN-UCD.

Surgical resection remains the therapeutic cornerstone for HN-UCD.¹² While local mass excision is the standard approach for typical well-circumscribed nodal disease, the surgical strategy should be tailored to the disease extent and specific anatomical location. As demonstrated in our cohort, selective neck dissection may be warranted when lesions involve multiple adjacent lymph nodes. Furthermore, for complex anatomical sites, a recent report demonstrated that transoral robotic surgery (TORS) is a safe and effective minimally invasive option for debulking mucosal UCD lesions, such as those situated at the base of the tongue.²⁴ However, there is limited guidance in the literature regarding the management of regional recurrence or new postoperative lesions. In this study, one patient developed a new enlarged lymph node in an adjacent area postoperatively; due to its slow growth and the absence of systemic symptoms, an active surveillance strategy was chosen. This scenario may represent an intermediate state between UCD and MCD, termed “Oligocentric Castleman Disease” (OligoCD). OligoCD is defined by the involvement of more than one but a limited number of lymph node regions, with systemic inflammation less severe than that of idiopathic MCD.^10,12,25 Current expert consensus tends to recommend surgery over systemic chemotherapy for these patients.¹² Talat⁴ also noted in a 10-year retrospective study that lesions in some patients under observation remained stable over the long term without affecting their survival. Therefore, after excluding adverse prognostic factors, such as the plasma cell subtype, anemia (Hb ≤ 100 g/L), hypoalbuminemia (albumin ≤ 30 g/L), or involvement of ≥ 3 lymph node regions, a “wait-and-watch” strategy for asymptomatic regional new lesions is a viable option.^26,27

While surgery is typically curative, a very small number of patients still exhibit persistent postoperative systemic symptoms, posing a clinical challenge. One patient in our cohort experienced recurrent low-grade fever post-surgery, which, despite resolving spontaneously, underscores the clinical heterogeneity of HN-UCD. Such cases with hyperinflammatory manifestations clinically mimic MCD. Recent reports have emphasized that while most patients experience symptom resolution within several years after resection, symptoms such as fatigue and night sweats may persist in a minority.^9,10 Rigorous follow-up is essential to exclude lymphoma transformation or occult infection.²⁸ If symptoms persist and affect the quality of life after such exclusions, immunomodulatory therapy may be considered.¹⁰ For patients with favorable postoperative recovery, we recommend annual imaging (CT) and laboratory tests (including complete blood count, LDH, liver and kidney function, CRP, and immunoglobulin quantification). If there is no evidence of recurrence for five consecutive years, routine CT scans may be discontinued, maintaining only laboratory monitoring.¹²

This study had several limitations. First, its retrospective single-center design and relatively small sample size may introduce selection bias; specifically, PET/CT data were available for only four patients, making our SUVmax findings preliminary. Second, the follow-up duration for some cases was short, which may be insufficient to capture long-term complications or late-onset recurrence. Future research should focus on multicenter collaborations and extended follow-up periods to establish more evidence-based diagnostic and therapeutic standards for HN-UCD.

Conclusion

In conclusion, HN-UCD is a rare hypervascular lymphoproliferative disorder. Its preoperative imaging features are highly suggestive, typically manifesting as well-circumscribed, intensely enhancing masses with characteristic coarse feeding vessels and internal fibrous septa, whereas necrosis or cystic degeneration is rare. While surgical resection remains the gold standard of treatment, this study underscores the necessity of long-term postoperative follow-up. Specifically, for atypical cases, such as the emergence of new regional lymphadenopathy or persistent systemic symptoms, individualized surveillance protocols established through multidisciplinary collaboration are essential to guide subsequent clinical decision-making.

Footnotes

Acknowledgments

The authors would like to express their sincere gratitude to the staff of the Medical Records Department at PUMCH for their invaluable assistance and professional support in retrieving the historical clinical data spanning the 20-year study period.

ORCID iDs

Chuan Liu

Jing-lei Li

Chuan Chen

Wei Gu

Ethical Considerations

The study was conducted with the approval of the Peking Union Medical College Hospital Ethics Committee (approval no: K25C3638).

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the National High Level Hospital Clinical Research Funding (2022-PUMCH-B-094).

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author.*

References

Talat

Schulte

. Castleman’s disease: systematic analysis of 416 patients from the literature. The Oncologist. 2011;16(9):1316-1324. doi:10.1634/theoncologist.2011-0075

Munshi

Mehra

van de Velde

Desai

Potluri

Vermeulen

. Use of a claims database to characterize and estimate the incidence rate for Castleman disease. Leuk Lymphoma. 2015;56(5):1252-1260. doi:10.3109/10428194.2014.953145

Zhang

Dong

Peng

, et al. A national, multicenter, retrospective study of Castleman disease in China implementing CDCN criteria. Lancet Reg Health West Pac. 2023;34:100720. doi:10.1016/j.lanwpc.2023.100720

Talat

Belgaumkar

Schulte

. Surgery in Castleman’s disease: a systematic review of 404 published cases. Ann Surg. 2012;255(4):677-684. doi:10.1097/SLA.0b013e318249dcdc

Shen

Zeng

Liu

. Retrospective analysis of primary extranodal unicentric Castleman disease: a systematic review. Front Med. 2026;13:1749931. doi:10.3389/fmed.2026.1749931

Boutboul

Fadlallah

Chawki

, et al. Treatment and outcome of Unicentric Castleman Disease: a retrospective analysis of 71 cases. Br J Haematol. 2019;186(2):269-273. doi:10.1111/bjh.15921

Jiang

Song

Liu

Cao

Zhang

. Castleman disease of the neck: CT and MR imaging findings. Eur J Radiol. 2014;83(11):2041-2050. doi:10.1016/j.ejrad.2014.08.013

Wang

, et al. Comprehensive analysis of 225 Castleman’s diseases in the oral maxillofacial and neck region: a rare disease revisited. Clin Oral Investig. 2018;22(3):1285-1295. doi:10.1007/s00784-017-2232-x

Bustamante

Shyamsundar

Coren

, et al. Ongoing symptoms following complete surgical excision in unicentric Castleman disease. Am J Hematol. 2023;98(11):E334-E337. doi:10.1002/ajh.27065

10.

Zhang

Jia

Chen

, et al. UCD with MCD-like inflammatory state: surgical excision is highly effective. Blood Adv. 2021;5(1):122-128. doi:10.1182/bloodadvances.2020003607

11.

Maruyama

Hao

Cheng

, et al. Castleman’s disease of the buccal mucosa: report of a case and review of the literature of head and neck cases. Oral Surg Oral Med Oral Pathol Oral Radiol Endod. 2002;93(3):305-310. doi:10.1067/moe.2002.120026

12.

van Rhee

Oksenhendler

Srkalovic

, et al. International evidence-based consensus diagnostic and treatment guidelines for unicentric Castleman disease. Blood Adv. 2020;4(23):6039-6050. doi:10.1182/bloodadvances.2020003334

13.

Chang

Wang

Hung

, et al. Monoclonality and cytogenetic abnormalities in hyaline vascular Castleman disease. Mod Pathol Off J U S Can Acad Pathol Inc. 2014;27(6):823-831. doi:10.1038/modpathol.2013.202

14.

Pitot

Tahboub Amawi

Alexander

, et al. Imaging of Castleman Disease. RadioGraphics. 2023;43(8):e220210. doi:10.1148/rg.220210

15.

Bonekamp

Horton

Hruban

Fishman

. Castleman Disease: The Great Mimic. RadioGraphics. 2011;31(6):1793-1807. doi:10.1148/rg.316115502

16.

Wang

Yang

Wang

Che

. Castleman disease versus lymphoma in neck lymph nodes: a comparative study using contrast-enhanced CT. Cancer Imaging. 2018;18(1):28. doi:10.1186/s40644-018-0163-7

17.

Moon

Kim

, et al. Mediastinal Castleman disease: CT findings. J Comput Assist Tomogr. 1994;18(1):43-46. doi:10.1097/00004728-199401000-00009

18.

Zhou

Zhang

Peng

Yang

Guan

Zhou

. Imaging findings of Castleman disease of the abdomen and pelvis. Abdom Imaging. 2008;33(4):482-488. doi:10.1007/s00261-007-9282-5

19.

Chang

Hsu

Sheen

Hsieh

. Usefulness of Doppler spectral analysis and power Doppler sonography in the differentiation of cervical lymphadenopathies. AJR Am J Roentgenol. 1998;171(2):503-509. doi:10.2214/ajr.171.2.9694484

20.

Wang

Dong

Wen

. A 10-year observational single-center study of retroperitoneal unicentric Castleman disease. Medicine (Baltimore). 2021;100(10):e25088. doi:10.1097/MD.0000000000025088

21.

Liu

Niu

Gao

, et al. The sonographic characteristics of unicentric castleman disease - a single-center retrospective study. Cancer Imaging. 2025;25:117. doi:10.1186/s40644-025-00937-2

22.

Pickhardt

Wong

Mellnick

Sugi

Aswani

. Abdominal CT findings characteristic of Castleman disease: multi-centre review of 76 adult cases with abdominopelvic nodal involvement. Br J Radiol. 2024;97(1160):1431-1436. doi:10.1093/bjr/tqae111

23.

Zhao

Wan

Huang

Song

. Imaging and clinical features of Castleman Disease. Cancer Imaging Off Publ Int Cancer Imaging Soc. 2019;19(1):53. doi:10.1186/s40644-019-0238-0

24.

Cantore

Cianfrone

Tauro

, et al. Transoral robotic surgery tongue base debulking in Castleman’s disease. Acta Otorhinolaryngol Ital. 2024;44(3):204-206. doi:10.14639/0392-100X-N2833

25.

Beckham

Yang

Chau

Noy

Yahalom

. Excellent Outcomes with Surgery or Radiotherapy in the Management of Castleman Disease Including a Case of Oligocentric Disease. Clin Lymphoma Myeloma Leuk. 2020;20(10):685-689. doi:10.1016/j.clml.2020.05.002

26.

Pierson

Brandstadter

Torigian

, et al. Characterizing the heterogeneity of Castleman disease and oligocentric subtype: findings from the ACCELERATE registry. Blood Adv. 2025;9(8):1952-1965. doi:10.1182/bloodadvances.2024014391

27.

Liu

Cai

, et al. Clinical characteristics and outcomes of Castleman disease: a multicenter Consortium study of 428 patients with 15-year follow-up. Am J Cancer Res. 2022;12(9):4227-4240.

28.

Haap

Wiefels

Horger

Hoyer

Müssig

. Clinical, laboratory and imaging findings in castleman’s disease - the subtype decides. Blood Rev. 2018;32(3):225-234. doi:10.1016/j.blre.2017.11.005