Correlation of Immunohistochemistry, FISH and Clinical Outcomes in Primary Mediastinal (Thymic) Large B-Cell Lymphoma

Introduction

Primary Mediastinal (thymic) B-cell lymphoma (PMBCL) is a distinct B-cell lymphoma which accounts for 2-4% of Non-Hodgkin's lymphoma. PMBCL is unique as although B-cell markers are readily expressed, morphologically and on gene expression profiling, PMBCL resembles classical Hodgkin's lymphoma. The poor prognostic impact of BCL2/BCL6 and MYC immunohistochemistry and FISH have been well published in Diffuse large B-cell lymphoma patients from other sites, however less is known about their significance in PMBCL. We retrospectively analysed 36 patients with PMBCL and correlated immunohistochemistry and FISH with clinical outcome.

Methods

Patient selection

We performed a search for all PMBCL cases in our pathology database from 2008-2016. 36 cases who had sufficient tissue and fulfilled the clinicopathological criteria were included.

Histologic and Immunohistochemical Methods

Immunohistochemical studies were performed using formalin-fixed, paraffin-embedded tissue sections. Antibodies against CD3, CD20, CD79A, CD10, BCL6, MUM1, BCL2, C-MYC , Ki-67 and PDL1 were used for immunohistochemical studies. These studies were performed using an automated immunostainer (Benchmark XT) and a streptavidin-biotin peroxidase detection system. Tissue specimens were reviewed for percent tumour positivity and nuclear intensity. The cutoff scores for positive antibody staining were: 30% for CD10, 30% for BCL6, 60% for MUM1, 40% for CMYC and 70% for BCL2. PDL1 was graded positive if 3+/2+ (complete membranous staining in at least 10-30% of cells) and negative if 1+/0.

In-situ hybridization

For detection of Epstein-Barr virus encoded RNA (EBER), in situ hybridization was performed on formalin-fixed, paraffin-embedded tissue sections using the Ventana ready-to-use kit as prescribed by the manufacturer.

Molecular fluorescence in situ hybridization (FISH) studies

For detection of MYC/BCL2/BCL6 gene rearrangements, the Vysis LSI Dual Color, Break Apart Rearrangement Probes performed on formalin-fixed, paraffin-embedded tissue using the Vysis paraffin pre-treatment kit II system.

Results

Out of the 36 cases, 15 (42%) were of the germinal center phenotype while 21 (58%) were non germinal center phenotype by the Hans criteria. MYC was expressed in 26 cases (72%). 44% of cases were double expressors. In addition, there was also increased BCL6 (N=35, 97%), Ki67>60% (N=31, 86%), MUM1 (N=25, 69%) and PDL1 expression (N=24, 66%) noted. EBERISH was not present in any of the cases in this series. Table 1 summarizes the IHC findings. Of the 36 cases, only 1 case had unsatisfactory BCL2 FISH analysis. Out of 35 evaluable cases, there was 1 positive MYC FISH and 2 positive BCL6 FISH. There were no cases with positive BCL2 FISH.

We compared our results with published results from DLBCL cases^1,2 and found higher proportion of MYC/BCL6 positivity and Ki67% by IHC. BCL2 IHC expression was similar. FISH positivity for myc/bcl2/bcl6 translocations was lower in PMBCL compared to DLBCL. By Han's criteria, there was no significant difference in GC and non-GC, however a higher proportion of double expressors was noted. PDL1 expression was also higher than those published in DLBCL, similar to PDL1 expression in Hodgkin's lymphoma³. Table 2 summarizes the comparison between our data and results from other studies.

In our cohort of 36 patients, the median age was 29 years (range 15-61) with 19 female patients. The majority had early stage (n= 24 ,66.7%) and bulky disease (n=24, 66.7%). All 36 patients were treated with immunochemotherapy (R-CHOP n=11, R-EPOCH n=21). With a median follow up of 41 months, PFS and OS was 95% and 81% respectively. There was no correlation between IHC/FISH markers, clinical prognostic markers (bulk, stage, age, IPI) with outcomes on univariate and multivariate analysis.

Conclusion

Our results validate that PMBCL is a unique lymphoma which has similarities with DLBCL yet bears resemblance to Hodgkin's lymphoma on IHC. The usual prognostic markers for DLBCL have not correlated with clinical outcome in statistical analysis suggesting that due to the uniqueness of this lymphoma, novel markers should be explored to improve prognostic accuracy and thus improve patient outcomes.

References

1. Hu S et al. Blood 2013; 121:4021-4031

2. Meyer PN et al. J Clin Oncol 2011; 29: 200-207

3. Menter T et al. Human Pathology 2016;54:17-24

View largeDownload slide

View largeDownload slide

Close modal