Immunohistochemical Characterization of Cell of Origin in Diffuse Large B-cell Lymphoma and its Correlation with the Double Expressor Phenotype: A Retrospective Study in a Tertiary Cancer Care Centre

Background – Diffuse Large B Cell Lymphoma (DLBCL) is the most common type of Non- Hodgkin Lymphoma (NHL) and is categorized into the Germinal Centre B-cell (GCB) and Activated B-cell (ABC/ Non-GCB) subtypes as per the Cell-of-origin (COO) model with the help of Gene Expression Proling/ Immunohistochemistry. The non -GCB subtype has been found to correlate with the Double Expressor (DE) phenotype (i.e. co-expression of bcl-2 and c-myc by IHC) and this correlation was substantiated and proven to be statistically signicant in the present study. Results- A retrospective study of 50 patients was carried out with the help of archival material led in the Department of Anatomic Pathology at Dharamshila Narayana Superspeciality Hospital from 1 st January 2019 to 30 th June 2020. The study cohort was divided into two Groups – Group-A with nodal presentation and Group -B with extranodal presentation. By using the Hans and Choi IHC algorithms, the cases were categorized into the GCB and non – GCB subtypes in both the groups. The DE phenotype was determined in each case by the co-expression of c-myc (>40%) and bcl-2 (>50%) by IHC. Conclusion – Due to the heterogeneity inherent in DLBCL, prediction of the Double Expressor phenotype and the COO by IHC is a sensitive tool and helps in the prognostication and therapeutic triage of patients. Hence in all cases diagnosed as DLBCL, a detailed morphological and IHC work up is mandatory to determine prognosis and possibly tailor therapy according to cell of origin and double expressor phenotype. The present study includes 50 patients, who presented to our institution during a period of one and a half years from the 1 st January 2019 to 30 th June 2020. The median age at presentation was 49 years (20-81 years) with a male to female ratio of 2:1. Group A with nodal disease includes 28 patients and Group B with Extranodal disease includes 22 patients, constituting 56% and 44% respectively of the study cohort. Histological sections examined from biopsy specimens showed diffuse inltration by predominantly large atypical lymphoid cells with prominent nucleoli and mitotic activity. All cases were positive for the pan B-Cell marker CD 20 and a few showed positivity for CD30.

Hence in all cases diagnosed as DLBCL, a detailed morphological and IHC work up is mandatory to determine prognosis and possibly tailor therapy according to cell of origin and double expressor phenotype.
There is a high frequency of extranodal disease, if there is generalized lymphadenopathy.
The "cell of origin" model divides DLBCL into the Germinal centre B-cell (GCB) type and Non Germinal centre B-cell type or Activated B-cell (ABC) type based on the pattern of gene expression. The gene expression of the Germinal Centre type ts with the normal germinal centre derived B cell and in the subtype non -GCB DLBCL or activated B-cell lymphoma, the gene expression pro ling more closely ts in with a normal activated B-cell.
"Cell of origin" (COO) model also determines the prognosis of the two biological subtypes of DLBCL, that should be treated differently.
C-myc is a proto-oncogene on chromosome 8q24 and encodes a transcription factor, which when dysregulated leads to cellular survival and proliferation. Bcl-2 is an oncogene with an anti-apoptotic property.
Myc rearrangement t(8;14) is prototypically associated with Burkitt lymphoma but is also associated with 12 to 15 % of DLBCL. BCL-2 rearrangement t(14;18) is also important as it leads to a drug resistant phenotype with increase in cancer cell survival.
Thus a concurrent rearrangement of Bcl-2 and c-myc, which are present in approximately 5 to 7 % of DLBCL, leads to a clinically resistant formof DLBCL, termed as a Double Hit Lymphoma and is associated with a poor prognosis. [3] The revised WHO classi cation recognizes the co-expression of myc (> 40%) and bcl-2 (> 50%) proteins as a new adverse prognostic marker within DLBCL, NOS as " Double Expressor Lymphoma". [4,5,6] Another variant of the double hit lymphoma presents with co-rearrangement of c-myc and bcl-6 genes.
All the 3 genes bcl-2, myc and bcl-6 are simultaneously rearranged in the phenotype termed as "Triple -Hit Lymphoma".
Patient who have the double hit rearrangement usually have protein over-expression and therefore have the double expressor phenotype. However the converse is not always true. [7] Double hit lymphoma occurs more commonly in GCB type of DLBCL, while DE Lymphoma is commoner in the Non-GCB type of DLBCL.
Double expressors are detected by IHC staining and Double Hit Lymphomas are detected by FISH.
The most common upfront treatment for Double Expressor Lymphomas is chemoimmunotherapy (CI) with R-CHOP (Rituximab, Cyclophosphamide, Doxorubicin, Vincristine and Prednisolone) which leads to a cure in 50-60% of patients. Patients who develop disease that is refractory to upfront treatment or relapse after achieving remission, are treated with autologous stem cell transplantation (ASCT) and they generally have a poor outcome.
The majority of patients with Double Expressor Lymphoma relapse after R-CHOP [8] , hence such patients should be treated with ASCT. For patients unable to undergo transplantation, the median survival is 6 months.
CART (Chimeric Antigen Receptor T-cell Therapy) is promising for patients with aggressive B-Cell Lymphomas,that do not respond to other treatment modalities.
Therefore all patients of DLBCL should be assessed for the DE phenotype, which aids in risk strati cation of the patient and further helps in optimizing the treatment strategy.

Methods
This was a retrospective study to validate a model to predict the DE phenotype based on COO subtype. All newly diagnosed DLBCL patients were divided into Group A including patients with Nodal involvement and Group B including patients with Extranodal involvement.
Based on "COO" both the groups were further divided into GCB type and Non GCB type of DLBCL. With the help of IHC using the Hans or Choi algorithm IHC staining for bcl-2 & c-myc was used to determine the DE phenotype and correlated with the COO subtypes of DLBCL.
Our study includes 50 patients, assessed over a period of one and half years from 1st January 2019 to 30th June 2020, who presented with generalized lymphadenopathy and/or Extranodal disease along with other associated constitutional symptoms.
A Biopsy was performed in each case and sent in 10% neutral buffered formalin to the Department of Surgical Pathology for further processing and evaluation. Formalin xed para n embedded tissue sections were cut and mounted on slides and stained with routine Haematoxylin and Eosin stain.
Further, IHC stains were applied to assess the 'COO' and DE phenotypes based on the percentage of staining of the cells and were correlated.

Results
The present study includes 50 patients, who presented to our institution during a period of one and a half years from the 1 st January 2019 to 30 th June 2020. The median age at presentation was 49 years (20-81 years) with a male to female ratio of 2:1. Group A with nodal disease includes 28 patients and Group B with Extranodal disease includes 22 patients, constituting 56% and 44% respectively of the study cohort.
Histological sections examined from biopsy specimens showed diffuse in ltration by predominantly large atypical lymphoid cells with prominent nucleoli and mitotic activity. All cases were positive for the pan B-Cell marker CD 20 and a few showed positivity for CD30.  In the Extranodal group though the non GCB subtype showed the DE phenotype more commonly than the NDE phenotype, the association was not found to be statistically signi cant (p-0.1434).  This distinction is important because better outcomes are seen in patients with the germinal-center subtype than with the non-germinal-center subtype. We therefore now have a "cell-of-origin" model showing that there are at least 2 genetic biologic types of DLBCL that might explain prognosis and could perhaps be targeted differently. This insight has led to recent trials that test new treatments in one genetic subtype vs the other. We are hopefully very close to understanding whether these 2 subtypes of DLBCL should be treated differently. [15] The most challenging pro le to detect in routine practice is "cell of origin." The gold standard for identifying germinal-centre DLBCL vs. non-germinal-centre DLBCL was rst de ned as gene-expression pro ling patterns in frozen tumor material. However, gene-expression pro ling is not routinely available nor is it considered a standard test. The most common approach is immunohistochemistry using different algorithms, such as the Hans, Choi or Tally algorithms, [9,10] to determine whether a lymphoma is germinal-centre DLBCL or non-germinal-centre DLBCL. Compared with the gold standard of gene expression pro ling in frozen material, there is an error rate of approximately 20% with immunohistochemical algorithms shown below.
The present study has used the Hans/Choi Algorithms to determine the Cell of Origin. [9,10] Double-expressor lymphoma The presence of both the MYC and BCL2 rearrangements de nes double-hit lymphoma (DHL). This phenotype is very proliferative and drug-resistant, and it is associated with a poor prognosis. Another variant of double-hit lymphoma is co-rearrangement of MYC and the BCL6 gene. Rarely, all 3 genes-BCL2, MYC, and BCL6-are simultaneously rearranged in a phenotype termed triple-hit lymphoma (THL).
Both double-hit and triple-hit lymphomas have a poor prognosis with standard treatment. [6,7] Immunohistochemical staining to identify protein expression of MYC also showed that there are lymphomas in which MYC and BCL2 genes are overexpressed at a protein level, without the genetic rearrangements. This pro le has been referred to as the "double-expressor" phenotype in DLBCL in the revised 2016 World Health Organization (WHO) classi cation of lymphoid neoplasms. The WHO classi cation de nes overexpression as greater than 40% c-MYC-expressing cells and greater than 50% BCL2-expressing cells by IHC. As shown in a study by Hu and colleagues, patients with double-expressor DLBCL have worse outcomes than patients in whom these proteins are not overexpressed; in general, only one-third of patients have long-term disease control with R-CHOP. [16] Double-hit lymphoma is relatively uncommon, occurring in approximately 5% to 7% of patients with DLBCL. However Dual-expressor lymphomas may be present in as many as one-third of patients with DLBCL and serve to identify a signi cant subset of cases with a worse prognosis.
Patients who have the double-hit rearrangement usually have protein overexpression, and therefore have the double-expressor phenotype. However, the converse is not always true: dual-expressor protein overexpression is not always associated with an underlying double-hit re arrangement. Complicating the picture is that most double-hit lymphomas occur in the setting of a germinal-centre DLBCL, whereas most double-expressor lymphomas occur in non-germinal-centre DLBCL as demonstrated in this study which shows a strong correlation between the Double Expressor phenotype and the Non-GCB Cell of Origin subtype by IHC.
The double-hit lymphomas can be detected with uorescence in situ hybridization (FISH) or standard cytogenetic analysis. The double-expressor lymphomas are diagnosed by immunohistochemistry.
The double-expresser phenotype was not given a unique category, but was recognized by the WHO as a poor prognostic sign within DLBCL.
When patients with aggressive B-cell lymphomas relapse or become refractory to therapy, standard options are limited. For patients unable to undergo transplant, or for those who relapse after a stem cell transplant, the median survival is approximately 6 months. Despite the many trials that have tried to improve upon this dismal statistic, there are no breakthroughs at this time. Chimeric antigen receptor (CAR) T-cell therapy is exciting. This treatment is still in the early phases of research and associated with toxicity, but it is promising for patients with aggressive B-cell lymphomas that do not respond to other therapies. There are also a number of new biologic and targeted agents that are promising, and nding which patients may respond to a particular treatment is a high priority. [17] Conclusion According to the 2016 WHO classi cation, the diagnosis of DLBCL, NOS requires the inclusion of the COO (GCB or ABC/non-GCB subtype) determined either with molecular techniques (GEP and mRNA based techniques) or immunohistochemistry, as an alternative solution. The distinction of GCB versus ABC-DLBCL has not yet led to differences in primary treatment. The current standard of care for most patients is R-CHOP, which has improved dramatically the outcome of DLBCL. However, for patients who fails to respond to R-CHOP, the choice of therapy is very likely to be in uenced by the COO and the molecular pathways used by the tumors for survival and proliferation. Emerging new targeted therapy will certainly in uence the diagnosis and treatment of DLBCL and HGBL in the near future. The routine use of FISH and IHC to detect MYC and BCL2 alterations/overexpression is recommended.
Patients with DHL and double expression of MYC and BCL2 protein, represent poor-risk subsets in which alternative strategies should be explored.
HGBL with MYC and BCL2 and/or BCL6 rearrangements (i.e., DHL or THL) should be separated from DLBCL, NOS, due to its clinicopathological features, molecular ndings, and dismal prognosis with standard R-CHOP therapy.
Although there are no strict recommendations on how to select cases for FISH analysis, a reasonable approach is to perform FISH analysis for MYC, BCL2 and/or BCL6 in cases with aggressive clinical presentation, GCB phenotype, and double expression of MYC and BCL2 which has been shown to correlate with the Non-GCB Subtype as demonstrated by this study ,thereby mandating a comprehensive and detailed morphologic and IHC work-up in all cases of High Grade B-cell Lymphoma in order to determine prognostic and therapeutically relevant and biologically distinct phenotypes .
Abbreviations COO-cell of origin