Endoscopic-Assisted Excision of Parotid Castleman Disease in Pediatric Patients: A Case Series of Three Patients and a Literature Review

Introduction

Castleman disease (CD), first described by Benjamin Castleman in 1956, is a heterogeneous group of lymphoproliferative disorders characterized by non-neoplastic lymph node enlargement [1]. Although CD can develop in any lymph node region, the mediastinum is the most commonly affected site, representing approximately 60% of cases, followed by the head and neck region, involved in about 14% of cases [2]. Within the head and neck region, 89% of lesions occur in the neck [3,4], whereas parotid gland involvement is rare, particularly among pediatric patients. Current understanding of parotid CD is largely based on isolated case reports and small series, with a notable scarcity of consolidated clinical analyses, especially in children [2,5–15].

CD is classified into 3 clinical types: unicentric CD (UCD), human herpesvirus-8–associated multicentric CD, and idiopathic multicentric CD. Pathologically, CD can be divided into 2 predominant patterns – hyaline vascular (HV) and plasma cell; a mixed pattern is occasionally observed. The HV type is most common, representing 80% to 90% of cases, and usually presents as solitary lymphadenopathy [4,16,17]. The multicentric form typically manifests with systemic symptoms and requires multimodal treatment, including corticosteroids, chemotherapy, antiviral therapy, and immunotherapy [4,18]. In contrast, the HV type, which constitutes the majority of UCD cases, generally follows a benign clinical course. Complete surgical excision and subsequent histopathologic confirmation remains the definitive management strategy, providing both diagnosis and cure. However, diagnostic challenges and management dilemmas – particularly those related to cosmetic outcomes in young patients – persist.

This study retrospectively analyzed 3 pediatric cases of UCD localized to the parotid gland; we supplemented the findings with a systematic review of the literature. The specific objectives were: (1) to delineate the clinical, radiologic, and histopathologic characteristics of parotid UCD in children; (2) to evaluate the efficacy and cosmetic outcomes of endoscopic-assisted resection as a minimally invasive surgical option; and (3) to highlight the importance of long-term follow-up, given the potential for disease progression or recurrence. Through these aims, we seek to establish clearer diagnostic pathways and refined surgical strategies for this rare condition, thus enhancing clinical management and outcomes in pediatric parotid CD.

Case Reports

CASE 1:

A 2-year-old girl was referred to our department with a 6-month history of a right-sided neck mass. According to her parents, the child was asymptomatic and exhibited no systemic signs such as fever, cough, or weight loss. Although the mass had gradually increased in size, it remained painless and did not cause dysphagia or feeding difficulties. The patient had previously been evaluated at a local hospital, where ultrasonographic findings led to an initial diagnosis of lymphadenitis, and oral antibiotics had been prescribed. However, her parents reported no change in mass size after treatment. Physical examination revealed a firm, non-tender, mobile mass measuring 3.5×3.0 cm, located inferior to the right mandibular angle. Facial nerve function was intact. Routine laboratory tests, including complete blood count and chemistry panel, yielded results within normal limits. Contrast-enhanced computed tomography (CT) showed a circular mass at the inferior margin of the right parotid gland (Figure 1). Fine-needle aspiration was not performed due to the patient’s age. Given the preoperative impression of a benign lesion and the parents’ aesthetic concerns, endoscopic resection through a retroauricular hairline incision was performed (Figure 2). Under general anesthesia, a 3- to 4-cm retroauricular hairline incision was made and extended through the skin and subcutaneous tissue to the fascia of the sternocleidomastoid muscle, providing direct access to the mass. Due to the limited operative field in pediatric patients, a gasless endoscopic technique was used to enhance surgical access. A unilateral distractor elevated the vestibular flap to improve visualization. A 5-mm endoscope was inserted through the central port; noninvasive graspers and a 3-mm electrocautery hook were introduced on either side. Under endoscopic visualization, key anatomical landmarks – including the greater auricular nerve, external jugular vein, inferior pole of the parotid gland, and the marginal mandibular branch of the facial nerve – were identified. After the facial nerve position had been confirmed, anterograde dissection was carried out to remove the tumor along with adjacent superficial parotid tissue. The specimen was excised en bloc and extracted through the incision. Histopathologic examination confirmed the diagnosis of HV-type CD (Figure 3). The postoperative course was uneventful. No recurrence has been observed during follow-up (9 months to date), and the patient’s cosmetic outcome remains excellent (Figure 4A, 4B).

CASE 2:

A 7-year-old boy presented with a 2-month history of an asymptomatic mass located anterior to the left ear. Physical examination revealed a firm, non-tender, mobile mass measuring 3.0×3.0 cm in the superficial lobe of the left parotid gland. Other physical examination findings were unremarkable. Routine laboratory investigations, including serum chemistry and complete blood count, showed results within normal limits. Chest radiography demonstrated no abnormalities. Magnetic resonance imaging revealed a well-circumscribed mass within the left parotid gland. Considering the parents’ aesthetic concerns, endoscopic-assisted excision was performed through a postauricular sulcus incision. A V-shaped incision was made along the natural skin folds from the tragus to the earlobe and cranio-auditory sulcus. The flap was elevated under direct vision; a unilateral distractor was used to lift the vestibular flap, improving visibility and surgical access. The endoscopic technique was consistent with the approach used in Case 1. Because the lesion was confined to the superficial lobe, active facial nerve dissection was not required. The specimen was excised en bloc and removed through the incision. Histopathologic examination confirmed the diagnosis of HV-type CD. The postoperative course was uneventful, with no recurrence during follow-up (6 months to date) and satisfactory cosmetic outcomes (Figure 4C, 4D).

CASE 3:

A 10-year-old boy was evaluated for a slowly enlarging mass inferior to the left earlobe that had been present for 1 month. The patient was otherwise asymptomatic and in good general health. Physical examination revealed a soft, fixed mass measuring 2.0×1.5 cm adjacent to the superficial lobe of the left parotid gland. No additional abnormalities were identified. Given the parents’ concern about cosmetic appearance, endoscopic-assisted excision was performed through a postauricular sulcus incision using the technique described for Case 2. Histopathologic examination confirmed the diagnosis of HV-type CD. No recurrence has been observed during follow-up (18 months to date; Figure 4E, 4F).

Only the patient in Case 3 demonstrated a marginally elevated C-reactive protein level (5.40 mg/L; reference range, 0–5 mg/L) prior to surgery. White blood cell count, C-reactive protein, and albumin levels were within normal limits in Cases 1 and 2. Preoperative ultrasonography and contrast-enhanced CT or magnetic resonance imaging revealed no lymph node enlargement beyond the solitary parotid lesion in any patient. Quarterly follow-up was conducted for the first year postoperatively, with decreasing frequency thereafter, given the decreasing risk of recurrence over time. Three months after complete excision, abdominal ultrasonography confirmed the absence of hepatosplenomegaly in all patients; targeted ultrasonography of the parotid bed showed no residual or new lymphadenopathy.

LITERATURE REVIEW:

A systematic search of PubMed and Web of Science covering January 2000 to July 2025 was conducted for English-language reports using the keywords “Castleman’s disease”, “Castleman”, and “parotid gland”. The search was restricted to studies involving human participants. Detailed information from published cases of parotid gland CD was summarized for analysis.

DEMOGRAPHIC AND CLINICAL PRESENTATION:

In total, 32 published cases of parotid gland CD from 27 articles were identified through the systematic review [2,5–15,17,19–32]. Among these, 13 (40.6%) involved pediatric patients younger than 18 years (Table 1). Patient ages ranged from 6 to 17 years; the mean age was approximately 12.8 years. There was a clear male predominance: 10 male patients (77%) and 3 female patients (23%). Most patients (12/13, 92.3%) were asymptomatic and presented with a slowly enlarging mass; 1 patient (Case 3) reported local pain. The duration of swelling before diagnosis varied from 3 months to 6 years. All 13 cases (100%) represented the UCD form. Facial asymmetry and nerve dysfunction were absent in all patients at presentation.

LESION CHARACTERISTICS:

The disease involved the left parotid gland in 6 cases, right parotid gland in 6 cases, and both the parotid gland and neck in 1 case. Maximum tumor diameter ranged from 2.5 cm to 4.8 cm; the average diameter was approximately 3.7 cm. When available, imaging findings (magnetic resonance imaging or CT) typically described a well-circumscribed, homogeneous mass confined to the parotid gland.

TREATMENT AND HISTOPATHOLOGY:

Surgical excision served as the primary treatment in all cases. Reported surgical approaches included superficial parotidectomy (most common), complete excision, and total parotidectomy. None of the reviewed reports documented the use of endoscopic assistance during surgery. Additionally, no data were provided regarding patient satisfaction or cosmetic outcomes. All 13 pediatric patients (100%) were diagnosed with HV-type CD. Notably, 1 adult patient initially diagnosed with HV-type CD received a modified diagnosis of B-cell lymphoma after reoperation [31].

OUTCOMES AND FOLLOW-UP:

Postoperative outcomes were overwhelmingly favorable; most reports indicated that the patients were alive and disease-free after surgery. One patient was lost to follow-up, and 2 cases lacked follow-up information. One patient developed additional systemic symptoms – including fever, night sweats, and abdominal masses – during follow-up.

Discussion

CD is a unique form of lymph node hyperplasia with an unclear origin. Some investigators have proposed that it results from lymphoproliferation triggered by chronic antigenic stimulation from viral infection or persistent inflammation, whereas others regard it as a lymphoid hamartoma [34]. Interleukin-6, receptor polymorphisms, human immunodeficiency virus, and human herpesvirus 8 have all been implicated in its etiopathogenesis [10,30,35,36].

The most common sites of CD are the mediastinum, head and neck, cervical lymph nodes, axilla, and abdomen [3]. Although the head and neck region represents the second most common location, with approximately 225 cases reported [4], involvement of the parotid gland is exceedingly rare [5,9,21,37]. Most cases originate within parotid or periparotid lymph nodes; few reports have described simultaneous unilateral parotid and cervical involvement [11].

Our case series, together with the systematic review of 32 published cases, confirms that the unicentric HV type is the predominant form of CD in the parotid region. This type follows a benign clinical course, typically presenting as a solitary, slow-growing, and asymptomatic mass. The consistent absence of systemic manifestations and facial nerve dysfunction across reviewed cases further supports its localized nature. The HV type is well defined, and complete surgical excision is curative in the majority of cases.

Although fine-needle aspiration cytology (FNAC) is a rapid, accessible, and reliable method for differentiating benign from malignant lesions, initial FNAC findings may be inconclusive because the cytomorphologic features of CD often resemble those of reactive lymphadenopathy. Moreover, FNAC may be nondiagnostic in approximately 6% of cases due to inadequate sampling or deep-lobe localization of the lesion [36]. Consequently, repeat FNAC or definitive histopathologic evaluation is often required for confirmation [7]. Complete surgical excision followed by histopathologic examination remains the most effective diagnostic and therapeutic approach. Conventional surgical methods, including superficial parotidectomy, are widely utilized to manage UCD of the parotid gland.

However, clinical management remains challenging. In pediatric patients, the absence of prominent symptoms often leads patients and their guardians to decline surgical intervention due to concerns about potential complications, such as scarring. Pediatric surgeons must also consider children’s psychological needs when determining treatment strategies. Based on our experience with pediatric endoscopic procedures, we adopted a modified approach for UCD in the parotid region to achieve effective tumor removal with excellent cosmetic results. Endoscopy provides superior visualization, improved facial nerve protection, and more precise lesion excision. This technique offers better cosmetic outcomes through shorter and more concealed incisions relative to conventional methods. Endoscopic-assisted resection of parotid gland tumors has gained increasing attention in recent years [34,38]. Although various endoscopic techniques have been utilized in parotid surgery [39], few reports have described their use in children. Our experience confirms that endoscopic-assisted resection of parotid gland tumors can be safely and effectively performed in pediatric patients. This approach ensures adequate visualization and safe dissection while minimizing visible scarring, thus addressing cosmetic concerns without compromising surgical efficacy.

Despite the excellent prognosis typically associated with HV-type parotid UCD, long-term follow-up remains essential. One pediatric patient developed fever, night sweats, and abdominal masses during follow-up [13], findings that are highly atypical for HV-type UCD and may indicate misclassification or coexistence of another condition. An adult patient adult patient initially diagnosed with HV-type CD received a modified diagnosis of B-cell lymphoma after reoperation [31]. Three additional adult patients experienced recurrence and required repeat excision [21,22,31]. These findings highlight the need for thorough initial evaluation and ongoing surveillance to detect recurrence, multifocal progression, or malignant transformation. Postoperative monitoring should include assessments for local recurrence and systemic involvement, incorporating routine laboratory testing of inflammatory markers – such as complete blood count, C-reactive protein, albumin, and erythrocyte sedimentation rate – along with clinical evaluations for lymphadenopathy and splenomegaly.

Conclusions

Parotid CD in pediatric patients is almost exclusively unicentric and exhibits the HV type, which is associated with an excellent prognosis after complete surgical excision. Among pediatric patients, for whom cosmetic outcomes are particularly important, endoscopic-assisted resection through discreet incisions represents a viable and aesthetically superior option. Despite the generally favorable outcomes, continued clinical and radiologic follow-up is advised to detect potential rare complications, such as recurrence or malignant transformation. This approach ensures optimal functional and cosmetic results while maintaining the high curative potential of surgical treatment.