Immunoarchitectural Patterns of Germinal Center Antigens Including LMO2 Assist in the Differential Diagnosis of Marginal Zone Lymphoma and Follicular Lymphoma
Kathryn S Dyhdalo, Christopher Lanigan, Raymond R Tubbs, James R Cook. Cleveland Clinic, Cleveland, OH
Background: The distinction between marginal zone lymphoma (MZL) and follicular lymphoma (FL) may be challenging, especially in MZL with numerous germinal centers (GCs) or in FL lacking BCL2 expression. The neoplastic cells of FL typically express the GC antigens CD10 and BCL6, although a minority of cases may lack either of these markers. Recently, LMO2 has been described as another GC-associated marker that is positive in most FL. In MZL, the neoplastic cells lack expression of GC antigens, but altered GC architecture may be seen secondary to follicular colonization. In this study, we examined the immunoarchitectural patterns of CD10, BCL6 and LMO2 in MZL and FL and their clinical utility in addressing this differential diagnosis.
Design: 42 cases of lymph nodes involved by MZL were identified, including 25 primary nodal, 5 splenic MZL, and 12 MALT. In order to assess GC architectectural patterns, whole slides of MZL were analyzed. For comparison, 88 cases of FL were analyzed using a TMA containing duplicate 2 mm cores. Immunohistochemical stains for CD10, BCL6, BCL2, and LMO2 were performed (Ventana Medical Systems, Tuscon, AZ). Positive staining in interfollicular areas was defined by >20% positive cells. Germinal center staining was qualitatively categorized as confluent, disrupted, scattered, or negative.
Results: GC were present in 34/42 (81%) of MZL. Interfollicular staining for CD10, BCL6, LMO2, or any GC marker was identified in 64/88 (73%), 6/88 (7%), 57/88 (65%), and 73/88 (83%) in FL versus 0/34 (0%), 1/34 (3%), 5/34 (15%), and 5/34 (15%) in MZL containing GC (p<0.001, 0.67, <0.001, and <0.001, respectively). Interfollicular staining for any GC marker was more frequent in BCL2 negative FL (9/15, 60%) than in MZL (5/34, 15%, p=0.004) but less frequent than in BCL2 positive FL (64/73, 88%, p=0.018). Of the 34 MZL containing GC, 13 displayed no more than scattered CD10 positive cells with confluent or disrupted BCL6 and/or LMO2 (38%), and 13 showed no more than scattered positive cells with CD10, BCL6, and LMO2 (38%). These two patterns were seen in 7/88 (8%, p<0.001) and 2/88 (2%, p<0.001) of FL, respectively.
Conclusions: Staining for LMO2 in addition to CD10 and BCL6 facilitates the detection of a GC phenotype in FL. Interfollicular staining for LMO2, however, may also be seen in a minority of MZL. Lymph nodes involved by MZL frequently show characteristic alterations of GC immunoarchitecture, and recognition of these altered patterns assists in the distinction between MZL and FL.
Monday, March 19, 2012 1:00 PM
Poster Session II # 217, Monday Afternoon