Contributions
Abstract: PB1779
Type: Publication Only
Background
Pleural effusion (PE) is reported in up to 20% of patients (pts) with non-Hodgkin lymphoma (NHL), and it is mainly due to pleural infiltration or lymphatic and thoracic duct obstruction. However, the risk factors for its development, as well as the evaluation of the diagnostic tools for lymphoma cells identification are not yet well established.
Aims
We aimed to evaluate the risk factors and diagnostic techniques for PE study in pts with NHL.
Methods
We analyzed all pts that performed a diagnostic thoracentesis and had a previous or simultaneous diagnosis of NHL, between January/2007 and December/2017. Inter-technique reliability was evaluated using the Cohen’s kappa test. The overall survival (OS) between the groups was compared using a log-rank test, followed by subset analysis for estimated survival among patients using univariate Cox proportional hazard regression.
Results
During this period, we identified 55 pts with a median age of 76 years (26-89), 49.1% were males. The main NHL subtypes were diffuse large cell lymphoma (DLCL) (39.6%; n=21), follicular lymphoma (28.3%; n=15), chronic lymphocytic leukemia/small lymphocytic lymphoma (9.4%; n=5), marginal zone lymphoma (7.6%; n=4) and T-cell NHLs (5.7%; n=3). All the pts had stages III-IV. At the time of PE identification, 79.3% of pts (n=42) had mediastinal lymphadenopathies (28.6% bulky mediastinal disease) and 11.3% (n=6) were in leukemic form.
51.5% (n=31) pts had PE at presentation of NHL. Among those that developed PE later, during the course of the disease, the median time to diagnosis was 18.4 months; 20 had been treated with chemotherapy and 3 pts received mediastinal radiotherapy. In 17.0% of the pts the PE was bilateral; 37.7% had a chest tube drainage and only 13.2% had performed a pleurodesis.
Considering PE biochemical characteristics, 84.9% pts (n=45) presented an exudate (79.2% with ≥2 Light's Criteria), 9.4% (n=5) a chylothorax and 5.7% (n=3) a transudate. Curiously, the presence of bulky mediastinal disease was not associated with the risk to develop chylothorax (OR 1.9; p=NS).
The flow cytometry study was performed in 75.5% (n=40) pts and was positive in 80.5% (median effusion cellularity 3200 cells/uL; median infiltration by NHL cells 33%). NHL subtype was identified in PE in 75% of the cases. Cytology recognized abnormal cells in 42.6% pts. The consistency between flow cytometry and cytology to evaluate the presence of lymphoma cells in PE was moderate (κ=0.42).
Effusion relapse occurred in 20.8% (n=11) pts, with a median time of 3.7 months. The identification of lymphoma cells in immunophenotyping (OR 2.6; p=NS) or morphological (OR 4.8; p=NS) studies did not modify the risk for relapse. Only 2 pts had bacterial growth in effusion cultures. However, although no pt had positive cultures for Mycobacterium spp., 49.1% had increased levels of pleural adenosine deaminase (ADA) (>40 IU/L).
The median OS for the cohort was 3.7 months. There was no difference in OS between pts with positive effusion flow cytometry (HR 1.5; p=NS) or cytology (HR 1.5; p=NS).
Conclusion
Our study shows a higher sensitivity in the identification of lymphoma cells in PEs by flow cytometry compared with cytology, with a moderate concordance between the two techniques.
Pleural ADA levels had shown low specificity for tuberculosis pleuritic infection in NHL pts, suggesting that cultures for Mycobacterium spp. should always be performed. The presence of lymphoma cells in PE of NHL pts does not seem to have a prognostic relevance in OS.
Session topic: 21. Aggressive Non-Hodgkin lymphoma - Clinical
Keyword(s): Non-Hodgkin's lymphoma
Abstract: PB1779
Type: Publication Only
Background
Pleural effusion (PE) is reported in up to 20% of patients (pts) with non-Hodgkin lymphoma (NHL), and it is mainly due to pleural infiltration or lymphatic and thoracic duct obstruction. However, the risk factors for its development, as well as the evaluation of the diagnostic tools for lymphoma cells identification are not yet well established.
Aims
We aimed to evaluate the risk factors and diagnostic techniques for PE study in pts with NHL.
Methods
We analyzed all pts that performed a diagnostic thoracentesis and had a previous or simultaneous diagnosis of NHL, between January/2007 and December/2017. Inter-technique reliability was evaluated using the Cohen’s kappa test. The overall survival (OS) between the groups was compared using a log-rank test, followed by subset analysis for estimated survival among patients using univariate Cox proportional hazard regression.
Results
During this period, we identified 55 pts with a median age of 76 years (26-89), 49.1% were males. The main NHL subtypes were diffuse large cell lymphoma (DLCL) (39.6%; n=21), follicular lymphoma (28.3%; n=15), chronic lymphocytic leukemia/small lymphocytic lymphoma (9.4%; n=5), marginal zone lymphoma (7.6%; n=4) and T-cell NHLs (5.7%; n=3). All the pts had stages III-IV. At the time of PE identification, 79.3% of pts (n=42) had mediastinal lymphadenopathies (28.6% bulky mediastinal disease) and 11.3% (n=6) were in leukemic form.
51.5% (n=31) pts had PE at presentation of NHL. Among those that developed PE later, during the course of the disease, the median time to diagnosis was 18.4 months; 20 had been treated with chemotherapy and 3 pts received mediastinal radiotherapy. In 17.0% of the pts the PE was bilateral; 37.7% had a chest tube drainage and only 13.2% had performed a pleurodesis.
Considering PE biochemical characteristics, 84.9% pts (n=45) presented an exudate (79.2% with ≥2 Light's Criteria), 9.4% (n=5) a chylothorax and 5.7% (n=3) a transudate. Curiously, the presence of bulky mediastinal disease was not associated with the risk to develop chylothorax (OR 1.9; p=NS).
The flow cytometry study was performed in 75.5% (n=40) pts and was positive in 80.5% (median effusion cellularity 3200 cells/uL; median infiltration by NHL cells 33%). NHL subtype was identified in PE in 75% of the cases. Cytology recognized abnormal cells in 42.6% pts. The consistency between flow cytometry and cytology to evaluate the presence of lymphoma cells in PE was moderate (κ=0.42).
Effusion relapse occurred in 20.8% (n=11) pts, with a median time of 3.7 months. The identification of lymphoma cells in immunophenotyping (OR 2.6; p=NS) or morphological (OR 4.8; p=NS) studies did not modify the risk for relapse. Only 2 pts had bacterial growth in effusion cultures. However, although no pt had positive cultures for Mycobacterium spp., 49.1% had increased levels of pleural adenosine deaminase (ADA) (>40 IU/L).
The median OS for the cohort was 3.7 months. There was no difference in OS between pts with positive effusion flow cytometry (HR 1.5; p=NS) or cytology (HR 1.5; p=NS).
Conclusion
Our study shows a higher sensitivity in the identification of lymphoma cells in PEs by flow cytometry compared with cytology, with a moderate concordance between the two techniques.
Pleural ADA levels had shown low specificity for tuberculosis pleuritic infection in NHL pts, suggesting that cultures for Mycobacterium spp. should always be performed. The presence of lymphoma cells in PE of NHL pts does not seem to have a prognostic relevance in OS.
Session topic: 21. Aggressive Non-Hodgkin lymphoma - Clinical
Keyword(s): Non-Hodgkin's lymphoma