EBUS, Granulomas, and Cancer

Lung cancer is the leading cause of cancer-related mortality worldwide. Lung cancer prognosis is strongly related to the stage of the disease, and appropriate selection of treatment is dependent on accurate staging. Regional nodal staging relies on a variety of investigations to determine the node size, metabolic activity, and cytopathologic confirmation or exclusion of malignant involvement. Determination of nodal involvement with computed tomography (on the basis of size) or positron emission tomography (on the basis of metabolic activity) is an important first step in the assessment of this patient population. However, as useful as noninvasive imaging modalities can be, they are subject to false-positive and false-negative results,1 which if relied on can lead to erroneous staging and treatment decisions. Endobronchial ultrasound (EBUS) is now widely accepted as a lymph node staging modality in lung cancer,2 and has also been shown to be a useful diagnostic procedure in identifying mediastinal metastases from distant tumors.3 EBUS appears to perform as well as mediastinoscopy (the accepted gold standard for preoperative mediastinal staging) in recent comparative clinical trials.4,5 EBUS has also been demonstrated to be sensitive in the evaluation of patients with suspected sarcoidosis, with performance characteristics above those of standard transbronchial needle aspiration.6,7 Tissue diagnosis of sarcoidosis is characterized by evidence of noncaseating epithelioid cell granulomas with the absence of organisms or particles,8 which in an appropriate clinical context is supportive of this diagnosis. Nevertheless, in subjects with a low pretest probability of having sarcoidosis (eg, cancer patients), such findings on EBUS specimens must be interpreted carefully. The finding of granulomatous inflammation in lymph nodes detected in patients with a concurrent or preceding diagnosis of malignancy can present a clinical conundrum, as suggested in a manuscript in this issue of JOBIP, which describes a series of such cases.9 First described in 1911 by Wolbach, sarcoid reactions have been described with both hematologic malignancies and solid tumors and are thought to represent a T-cell–mediated response to tumor factors.10 First reviewed in 1962,11 and more recently,12 sarcoid reactions in lymph nodes have been reported in a small subset of patients with primary tumors located in the intestine (4.8%), breast (3.3%), stomach (5.0%), skin (13%), head and neck (4.9%), and lung (up to 3.3%).10,13 Concurrent malignancy and epithelioid granulomas can present the following scenarios: Coexistent sarcoidosis/granulomatous disease and malignancy14,15 Past sarcoidosis/granulomatous disease and development of malignancy16,17 Sarcoid reaction to malignancy10,18 Cancer therapy–induced sarcoid reaction19,20 The present article reports a retrospective study identifying the incidence of sarcoid reactions in patients with a history of cancer discovered by EBUS. Of the 154 patients with granulomatous inflammation found on EBUS-TBNA, 12 patients (7.8%) had a recent history of cancer. Twenty-three lymph nodes were sampled in this cohort and no nodal station had simultaneous granuloma formation and malignancy, although no excisional biopsies or clinical follow-ups are presented to exclude this possibility. Although rare, granulomas can coexist with neoplastic cells in the same nodal station,13,21 and identification of granulomas on an aspirate cannot conclusively rule out metastatic involvement of a lymph node. As such, rapid onsite evaluation of specimens demonstrating granulomas should not preclude the bronchoscopist from obtaining additional specimens from the nodal station in question and from all other accessible sites. Cases where no nodal metastases are found by EBUS should still be considered for additional sampling (by mediastinoscopy or at the time of resection of a lung cancer), or at the very least be subject to careful radiologic monitoring to ensure a benign clinical course. The separation of a sarcoid reaction from coexistent sarcoidosis and malignancy may be difficult. A true sarcoidosis may be supported by clinical findings (eg, uveitis, erythema nodosum), elevated ACE levels, typical lung parenchymal findings on imaging, an elevated CD4:CD8 ratio on bronchoalveolar lavage,15 and the finding of additional sites of granulomatous inflammation such as in endobronchial or transbronchial lung biopsies. Sarcoidosis is typically referred to as a diagnosis of exclusion. As granulomas can be the presentation of a sarcoid reaction to malignancy, the diagnosis of sarcoidosis should be made carefully in the setting of a concurrent malignancy.22,23 Clinical monitoring of individual patients over time will usually be required before a more definitive diagnosis is reached. This study also highlights the importance of cytopathologic confirmation of imaging findings with tools such as EBUS-TBNA in staging patients with a pulmonary malignancy. In almost 90% of cases where positron emission tomography was available, lymph nodes with granulomatous inflammation but without nodal metastases were FDG-avid (ie, false positive), which may have led to inappropriate upstaging of the disease and inappropriate treatment selection. As the use of EBUS in cancer diagnosis and staging continues to proliferate, the challenge of interpreting the significance of granulomatous inflammation in this setting will be more common. The clinician will have to remain mindful of the various scenarios that could explain these findings, and consider appropriate next steps accordingly. Additional study of this phenomenon is critical, in particular with regard to EBUS (and other nonexcisional specimens). Future studies should aim to determine the frequency of concurrent intrastation malignancy by mediastinoscopy/resection or clinical follow-up confirming the stability or the resolution of adenopathy.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.