Mediastinal Lymphadenopathy and Interstitial Lung Disease in a Cancer Patient
Case Editor - Kamyar Afshar
Reviewed By Allergy, Immunology & Inflammation Assembly
Aaron S. Bruns, M.D.
The Ohio State University
Ken Knox, M.D.
John Mastronarde, M.D.
The Ohio State University
A 59-year old white woman presented to the clinic for evaluation of dyspnea and an abnormal chest computed tomography (CT). Three years ago, she was diagnosed with breast cancer. One of twenty-nine axillary lymph nodes were positive for tumor, which was estrogen receptor positive. She was treated with surgery, radiation, and chemotherapy that included doxirubicin, cyclophosphamide,paclitaxel, and gemcitabine. She was then given anastrozole. She was followed closely over the next two years and had no evidence of recurrence.
Before developing breast cancer, the patient was physically active, walking 3 to 4 miles daily. However, over the last two years she had increasing dyspnea on exertion and was able to walk only 1 mile. Just prior to the visit a chest radiograph revealed hilar and mediastinal lymphadenopathy. A chest CT confirmed bilateral hilar and mediastinal lymphadenopathy as well as perilymphatic nodules less than a centimeter in diameter throughout the lungs. In addition, a positron emission tomography (PET) scan revealed uptake in several mediastinal lymph nodes with a standard uptake value of 8. The patient denied any cough, weight loss, night sweats, fevers, or occupational or travel-related exposures. A repeat mammogram and breast ultrasound were both negative.
FEV1 2.77 (81%)
FVC 2.29 (86%)
TLC 4.18 (81%)
DLCO 20.3 (91%)
Chest CT: Multiple nodules smaller than a centimeter in diameter were present in both lungs with a perilymphatic distribution. A right paratracheal lymph node measured 1.9 cm. There was a 1.3 cm right precarinal lymph node, a 1.4 cm left anterior mediastinal lymph node, and a 1.4 cm subcarinal lymph node as well. There was no pleural effusion or pneumothorax (Figures 1 and 2).
PET scan: Uptake in several mediastinal lymph nodes with a standard uptake value of 8 (Images not available).
Figure 1. Axial view of high-resolution chest CT using lung windows showing multiple nodules less than one centimeter in diameter in a perilymphatic distribution.
Figure 2. Axial view of high-resolution chest CT using mediastinal windows showing bilateral hilar and subcarinal lymphadenopathy.
Figure 3. Slide of transbronchial biopsy showing multinucleated giant cells and epithelioid histiocytes in a granulomatous formation (Magnified 20x)
Multinodular lung parenchymal disease can be characterized by lobar distribution and by diffuse versus focal disease. However, in order to develop a useful differential diagnosis, characterizing nodular disease by its relationship to secondary lobar anatomy is the most useful (1). The three most commonly used distributions are: random, centrilobular, and perilymphatic.
Randomly distributed nodules are by definition diffuse in nature with no obvious pattern. Diseases that are spread via a hematogenous route commonly cause a random distribution of nodules. Although diffuse, these nodules may show a predominance in the lung bases due to the higher perfusion there (2). Metastatic disease, such as carcinomatosis, is the most common cause of random nodules. Miliary infections, such as tuberculosis, Mycobacterium avium complex, or fungal disease can also cause this pattern, but are usually seen in a centrilobular distribution (see below).
Nodular disease can also be found clustered around the fissures, along the interlobular septae, and along the pleura. These regions are where the lymphatics are extensive, and this pattern is termed a perilymphatic distribution (2). Sarcoidosis is the classic disease that fits this distribution. Sarcoidosis can also occur in the lymphatics of the bronchovascular bundle. Silicosis and coal-workers’ pneumoconiosis can also appear in this distribution. Importantly, lymphangitic carcinomatosis occasionally fits a perilymphatic distribution, but it is usually associated with lymphadenopathy, pleural effusions, and septal thickening.
The third category involves nodules that do not touch the pleura and are not along the fissures. These nodules are found around the centrilobular bronchioles and their pulmonary artery branches, hence the label centrilobular nodules (3). The differential of this category is broader than the other two, and breaking it down further into nodules with a “tree-in-bud” appearance versus those with a more “ground-glass” appearance can be helpful. Tree-in-bud refers to branches that can come off the nodules, giving them this appearance (4). This appearance usually fits with an infection such as Mycobacterium avium complex, Mycobacterium tuberculosis, fungal, or other bacterial infection.
Hypersensitivity pneumonitis and respiratory bronchiolitis are the most common diseases that present as ground-glass centrilobular nodules (2). Other, less common diseases such as lymphocytic interstitial pnuemonitis and Langerhans’ histiocytosis can present this way as well.
In our case, the nodules were in a perilymphatic pattern. Given the patient’s history of cancer, recurrence was high on the differential and needed to be thoroughly evaluated. However, no malignant cells were seen in the mediastinoscopy samples or the trans-bronchial biopsy samples. Also, little septal thickening and no effusions were seen. Pathology instead revealed granulomatous lymphadenitis in the lymph node samples and multiple non-necrotizing granulomas in the trans-bronchial samples. While clubbing is rare in sarcoid, given these findings and the perilymphatic distribution, sarcoidosis is the most likely diagnosis. However, the diagnosis of a sarcoid-like reaction to tumor antigens released from a recurrence cannot be completely ruled out (see below). Given the large number of biopsies with no evidence of recurrence, this is somewhat less likely.
Given the patient’s lack of significant symptoms and relatively normal pulmonary function testing, therapy should initially be withheld until the course of the disease is established. This is especially true in light of her malignancy history, as any treatment has the potential to increase the risk of recurrence or development of a secondary malignancy.
One of the first attempted links between sarcoidosis and malignancy was a paper by Brincker in 1974 (5). He compared the incidence of lymphoma in the general population with the incidence in 2544 patients with respiratory sarcoidosis and found an 11-fold increase in the occurrence of lymphoma in this population. Others have noted patients with sarcoidosis having a higher incidence of other malignancies, mostly lung and breast, although links have been reported with many others (6).The above evidence citing a relationship between sarcoid and malignancy has been refuted by several authors. Israel (7) reviewed Brincker’s original article and noted that in half of the cases in which patients with sarcoid developed cancer, the cancer was diagnosed within one year of the diagnosis of sarcoid. Using a large national sarcoid registry of 8541 patients, Askling et al (8) compared the incidence ratios of a variety of malignancies in this group to that of the general population. Skin cancer, lung cancer, melanoma, and lymphoma rates were all increased in the sarcoid group. Again, most of the increase was found in the first 1-4 years after the diagnosis of sarcoid. Additional studies have shown the development of non-caseating granulomas in the organ where the tumor originated, along with the spleen and bone marrow of patients with cancer (9) without having the systemic signs of sarcoidosis. These investigations have led to the theory that patients with malignancies can develop areas of non-caseating granulomas as a response to their tumors without having overt sarcoidosis. This has been termed the sarcoid reaction. The most common occurrence of this is in lymph nodes which drain the area of the malignancy (10). Further controversy has come from other observations finding no link between sarcoidosis and malignancy (11). It is not currently clear what the relationship between sarcoid and malignancy is at this point.
While MRI is useful in evaluating for neurosarcoidosis (especially with gadolinium enhancement) and for solid organ involvement, its use in the evaluation of lymphadenopathy is limited (12), and therefore would not be helpful in this case.
CT is widely used in the evaluation of diseases of the lung and mediastinum. It is useful in detecting enlarged lymph nodes and parenchymal disease that may not be seen on conventional radiography (13).
PET has been widely used for the evaluation of malignancies, but it is only recently that its use in inflammatory and infectious disorders has been investigated (14). As sarcoidosis is an inflammatory disease, false-positives on PET are often seen and a misdiagnosis of cancer made (15,16). While malignant disease usually has higher metabolic activity than benign disease, this is not the case in sarcoid, which often shows high levels of metabolic activity (17). In a study out of Japan (18), 24 patients with known sarcoidosis were compared to a control group of known lung cancer patients, using two different methods of PET scanning. Standard 18F-FDG-PET scanning was unable to differentiate between lung cancer and sarcoid. However, using a new tracer (18F-fluorine-18-α-methyltyrosine)-PET scanning was able to differentiate between the two, as lung cancer had a significantly higher uptake of this tracer than sarcoidosis. However, this tracer is not currently widely available. While some evidence exists that 18F-FDG-PET may be useful in diagnosing extrapulmonary sarcoid or in following treatment response (12), further research is needed in this area to better determine its role.
In our case, MRI would not be useful. The CT scan was suggestive of sarcoid, but as stated above, lymphangitic carcinomatosis can present with perilymphatic nodules and lymphadenopathy, so in this case CT was not helpful in differentiating between the two. Also, standard PET would not be helpful as all involved areas would likely by hypermetabolic with either sarcoid or malignancy.
Bronchoalveolar lavage fluid is not diagnostic of sarcoidosis, but can help narrow the differential if other testing is not available. 90% of sarcoid patients will have an increased number of lymphocytes in their bronchoalveolar lavage fluid (21), giving it a rather high sensitivity. A CD4:CD8 ratio of 4:1 or higher has a positive predictive value of 94%, but a sensitivity of only 59% in separating sarcoid from other interstitial lung diseases (22). The relatively low specificity of bronchoalveolar lavage fluid makes it less useful than other tests in achieving a diagnosis.
Transbronchial biopsy is currently the test recommended by the American Thoracic Society for the diagnosis of sarcoidosis (23). However, the yield of this test depends on operator experience, the stage of disease, and the number of biopsies performed. Roethe et al (24) performed one of the earliest studies evaluating the yield of transbronchial biopsy in sarcoid patients. They found that in previous studies, where only 2-4 biopsy specimens were taken, the yield of this test was only 66% in stage I disease, and 80-85% in stage II and III disease. When the number of biopsies was increased to ten, the yield increased to >90% in all stages of disease. Therefore, it is recommended that 8-10 biospies be obtained to increase the likelihood of achieving diagnoses.
Transbronchial needle aspirations can be used to obtain tissue from a mediastinal lymph node to obtain a diagnosis. This procedure is dependent on operator technique and the availability of on-site cytology to give the highest yield. In a study evaluating its use in sarcoid, transbronchial needle aspiration successfully allowed a diagnosis of sarcoid in 72% of 32 patients presenting with stage I disease. The yield was further increased (85%) if transbronchial needle aspiration was combined with transbronchial biopsy (25). With the more recent use of endobronchial ultrasound guided aspirations, yields similar to that of transbronchial biopsy have been obtained, with less risk of pneumothorax (26). As use of this technique becomes more widespread, its role in sarcoid will be better defined.
- Bergin C, Roggli V, Coblentz C et al. Secondary pulmonary lobule: normal and abnormal CT appearance. Am J Roentgenol 1988;151:21-25.
- Raoof S, Amchentsev A, Vlahos I et al. Pictorial Essay: Multinodular disease. A high-resolution CT scan diagnostic algorithm. Chest 2006;129:805-815.
- Gruden JF, Webb WR, Warnock M. Centrilobular opacities on HRCT: diagnosis and differential diagnosis with pathologic correlation. Am J Roentgenol 1994;162:569-574.
- Collins J, Blankenbaker D, Stern EJ. CT patterns of bronchiolar disease: what is tree-in-bud? Am J Roentgenol 1998;171:365-370.
- Brincker H, Wilbek E. The incidence of malignant tumors in patients with respiratory sarcoidosis. Br J Cancer 1974;29:247-251.
- Brincker H. Solid tumors preceding or following sarcoidosis. Med Pediatr Oncol 1987;15:82-88.
- Israel HL. Sarcoidosis, malignancy, and immunosuppressive therapy. Arch Int Med 1978;138:907-908.
- Askling J, Grunewald J, Eklund A et al. Increased risk for cancer following sarcoidosis. Am J Respir Crit Care Med 1999;160:1668-1672.
- Kurata A, Terado Y, Schulz A et al. Inflammatory cells in the formation of tumor-related sarcoid reactions. Hum Pathol 2005;36:546-554.
- Cohen PR, Kurzrock R. Sarcoidosis and malignancy. Clinics Derm 2007;25:326-333.
- Romer FK. Sarcoidosis and cancerÃ¢â‚¬â€a critical view. Proceedings of the VIIIth International Conference on Sarcoidosis. Cardiff: Alpha Omega Press; 1980:567-571.
- Yu JQ, Zhuang H, Mavi A, Alavi A. Evaluating the role of flurodeoxyglucose PET imaging in the management of patients with sarcoidosis. PET Clin 2006;1:141-152.
- Hamper UM, Fishman EK, Khouri NF et al. Typical and atypical CT manifestations of pulmonary sarcoidosis. J Comput Assist Tomogr 1986;10:928-936.
- Alavi A, Gupta N, Alberini JL et al. Positron emission tomography imaging in nonmalignant thoracic disorders. Semin Nucl Med 2002;32:293-321.
- Yamada Y, Uchida Y, Tatsumi K et al. Fluorine-18-fluorodeoxyglucose and carbon-11-methionine evaluation of lymphadenopathy in sarcoidosis. J Nuc Med 1998;39:1160-1166.
- Lewis PJ, Salama A. Uptake of fluorine-18-fluorodeoxyglucose in sarcoidosis. J Nuc Med 1994;35:1647-1649.
- Cook GJ, Fogelman I, Maisey MN. Normal physiological and benign pathological variants of 18-fluro-2-deoxyglucose positron emission tomography scanning: potential for error in interpretation. Semin Nucl Med 1996;26:308-314.
- Kaira K, Oriuchi N, Otani Y et al. Diagnostic usefulness of fluorine-18-ÃŽÂ±-methyltyrosine positron emission tomography in combination with 18F-fluorodeoxyglucose in sarcoidosis patients. Chest 2007;131:1019-1027.
- Reich JM. Neoplasia in the etiology of sarcoidosis. Eur J Int Med 2006;17:81-87.
- OÃ¢â‚¬â„¢Connell MJ, Schimpff SC, Krischner RH et al. Epithelioid granulmoas in HodgkinÃ¢â‚¬â„¢s disease. JAMA 1975;233:886-889.
- Yeager H Jr, Williams MC, Beekman JF et al. Sarcoidosis: analysis of cells obtained by bronchial lavage. Am Rev Respir Dis 1977;116:951-954.
- Winterbauer RH, Lammert J, Selland M et al. Bronchoalveolar lavage cell populations in the diagnosis of sarcoidosis. Chest 1993;104:352-361.
- American Thoracic Society. Statement on sarcoidosis. Am J Respir Crit Care Med 1999;160:736-755.
- Roethe RA, Fuller PB, Byrd RB et al. Transbronchoscopic lung biopsy in sarcoidosis. Optimal number and sites for diagnosis. Chest 1980;77:400-402.
- Trisolini R, Agli LL, Cancellieri A et al. The value of flexible transbronchial needle aspiration in the diagnosis of stage I sarcoidosis. Chest 2003;124:2126-2130.
- Garwood S, Judson MA, Silvestri G et al. Endobronchial ultrasound for the diagnosis of pulmonary sarcoidosis. Chest 2007;132:1298-1304.