Sarcoidosis and maligancy: the chicken and the egg?

Hematologic and Oncologic Aspects of Sarcoidosis: Some of the Least Studied but Most Common Dilemmas

The hematologic system is frequently involved in sarcoidosis. Lymphopenia is the most common hematologic manifestation noted, although anemia and thrombocytopenia also occur. The etiology of these common manifestations can be direct granulomatous infiltration of bone marrow, lymph nodes, or spleen or related to immunologic dysfunction. Although not life threatening, these problems can lead to cytopenias requiring close monitoring in patients receiving a variety of disease treatments. The relationship between sarcoidosis and malignancy remains complex. However, some sarcoidosis patients are at increased risk for the development of malignancies, particularly lymphomas and gastrointestinal cancers. Conversely, cancer patients can experience an increase in the likelihood for the development of breast cancer and lymphomas.

Autoimmune diseases and risk of non-Hodgkin lymphoma: A Mendelian randomisation study

Objective

To examine whether genetically predicted susceptibility to ten autoimmune diseases (Behçet's disease, coeliac disease, dermatitis herpetiformis, lupus, psoriasis, rheumatoid arthritis, sarcoidosis, Sjögren's syndrome, systemic sclerosis, and type 1 diabetes) is associated with risk of non-Hodgkin lymphoma (NHL).

Design

Two sample Mendelian randomization (MR) study.

Setting

Genome wide association studies (GWASs) of ten autoimmune diseases, NHL, and four NHL subtypes (i.e., follicular lymphoma, mature T/natural killer-cell lymphomas, non-follicular lymphoma, and other and unspecified types of NHL).

Analysis

We used data from the largest publicly available GWASs of European ancestry for each autoimmune disease, NHL, and NHL subtypes. For each autoimmune disease, we extracted single nucleotide polymorphisms (SNPs) strongly associated (P < 5×10-8) with that disease and that were independent of one another (R2 < 1×10-3) as genetic instruments. SNPs within the human leukocyte antigen region were not considered due to potential pleiotropy. Our primary MR analysis was the inverse-variance weighted analysis. Additionally, we conducted MR-Egger, weighted mode, and weighted median regression to address potential bias due to pleiotropy, and robust adjusted profile scores to address weak instrument bias. We carried out sensitivity analysis limited to the non-immune pathway for nominally significant findings. To account for multiple testing, we set the thresholds for statistical significance at P < 5×10-3.

Participants

The number of cases and controls identified in the relevant GWASs were 437 and 3,325 for Behçet's disease, 4,918 and 5,684 for coeliac disease, 435 and 341,188 for dermatitis herpetiformis, 4,576 and 8,039 for lupus, 11,988 and 275,335 for psoriasis, 22,350 and 74,823 for rheumatoid arthritis, 3,597 and 337,121 for sarcoidosis, 2,735 and 332,115 for Sjögren's syndrome, 9,095 and 17,584 for systemic sclerosis, 18,942 and 501,638 for type 1 diabetes, 2,400 and 410,350 for NHL; and 296 to 2,340 cases and 271,463 controls for NHL subtypes.

Exposures

Genetic variants predicting ten autoimmune diseases: Behçet's disease, coeliac disease, dermatitis herpetiformis, lupus, psoriasis, rheumatoid arthritis, sarcoidosis, Sjögren's syndrome, systemic sclerosis, and type 1 diabetes.

Main outcome measures

Estimated associations between genetically predicted susceptibility to ten autoimmune diseases and the risk of NHL.

Results

The variance of each autoimmune disease explained by the SNPs ranged from 0.3% to 3.1%. Negative associations between type 1 diabetes and sarcoidosis and the risk of NHL were observed (odds ratio [OR] 0.95, 95% confidence interval [CI]: 0.92 to 0.98, P = 5×10-3, and OR 0.92, 95% CI: 0.85 to 0.99, P = 2.8×10-2, respectively). These findings were supported by the sensitivity analyses accounting for potential pleiotropy and weak instrument bias. No significant associations were found between the other eight autoimmune diseases and NHL risk. Of the NHL subtypes, type 1 diabetes was most strongly associated with follicular lymphoma (OR 0.91, 95% CI: 0.86 to 0.96, P = 1×10-3), while sarcoidosis was most strongly associated with other and unspecified NHL (OR 0.86, 95% CI: 0.75 to 0.97, P = 1.8×10-2).

Conclusions

These findings suggest that genetically predicted susceptibility to type 1 diabetes, and to some extent sarcoidosis, might reduce the risk of NHL. However, future studies with different datasets, approaches, and populations are warranted to further examine the potential associations between these autoimmune diseases and the risk of NHL.

Diagnostic accuracy and clinical impact of endoscopic ultrasound-guided fine-needle aspiration (EUS-FNA) in Positron Emission Tomography - Computed Tomography (PET-CT)-positive mediastinal lymphadenopathies in patients with thoracic or extra-thoracic malignancies

BACKGROUND

The high sensitivity of PET-CT can identify hypermetabolic mediastinal adenopathies during cancer management, but specificity is low and a biopsy is sometimes required to eliminate benign adenopathies.

METHODS

This prospective diagnostic accuracy study included patients with hypermetabolic mediastinal lymphadenopathies revealed on PET-CT during either the initial management of a cancer, treatment evaluation, or monitoring. All patients underwent EUS-FNA. Diagnoses of malignancy based on cytological analysis following EUS-FNA were compared with clinical and radiological follow-up information. The treatment strategy decided before the results of the EUS-FNA pathology reports (Multidisciplinary Team Meeting [MTM-1]) was recorded and compared to the treatment strategy decided once pathological data from EUS-FNA were available (MTM-2).

MAIN FINDINGS

Between 2013 and 2018, 75 patients were included with 47 eligible and evaluable patients. Sensitivity, specificity, and positive and negative predictive values of EUS-FNA were 93%, 100%, 100% and 90%, respectively. The concordance value between the therapeutic strategies determined for MTM-1 and MTM-2 was 44.7%. There were no significant differences in the intensity of fixation on PET-CT between malignant and benign lesions.

CONCLUSION

The diagnostic accuracy of the minimally invasive EUS-FNA procedure is sufficiently robust to avoid the need for diagnostic surgery. The combination of PET-CT and EUS-FNA may alter the therapeutic strategy that would be considered after PET-CT alone.

REGISTRATION

NCT01892501.