Clinico-Pathologic Profile and Treatment Outcomes of Patients with Diffuse Large B Cell Lymphoma Based on Cell of Origin Classification

Abstract

Diffuse large B cell lymphoma (DLBCLs) constitute 40% of all non-Hodgkin lymphoma and it represent a heterogeneous group of neoplasms rather than a single clinicopathological entity. We analysed the outcomes and clinical features based on the cell of origin in a series of patients with DLBCL from our institute. Medical case records of all newly diagnosed DLBCL treated in our institute from January 2015 to July 2017 were analysed for this study. Cell of origin classification was based on immunohistochemistry using Hans algorithm. Kaplan–Meier curves were used to determine survival. Ninety-five patients were diagnosed to have DLBCL subtype. Immunophenotypic subtyping was available for 71 patients. The median age at diagnosis was 56 years with no difference between Germinal centre B cell (GCB) and non-Germinal centre B cell (non-GCB) subtypes. Approximately 44% of patients had extra-nodal disease, stomach being the commonest site. Forty percent of patients had stage III/IV disease. Bulky disease and extra-nodal presentation was predominantly seen with non-GCB subtype (46% vs 20% and 36% vs 29% respectively). Rituximab was used in 75% of the patients with DLBCL. The 2-year disease-free survival was 70% versus 53% (p = 0.38) in GCB versus non-GCB subtype. This is one of the few data on DLBCL patients reported from India which has described outcomes based on the cell of origin. The disease-free survival in our country appears to be superior in GCB subtype which needs to be confirmed in a larger subset of patients.

Introduction

Diffuse large B-cell lymphoma (DLBCL) is the most common subtype of non-Hodgkin lymphoma (NHL), constituting up to 40% of all NHL cases globally [1]. DLBCL although considered usually as a specific category, the diversity in the clinical presentation, morphology, genetic and molecular alterations strongly suggests that these tumors represent a heterogeneous group of neoplasms rather than a single clinicopathological entity [2]. Diffuse large B cell lymphoma can be cured in more than half of the cases by conventional chemotherapy regimens. Standard treatment for NHL since the 1970s was CHOP regimen. In the past 2 decades, integration of Rituximab improved survival significantly and R-CHOP became the standard of care for patients with DLBCL. There was improvement in complete response rates (CR) by 15–20%, 5-year event-free survival (EFS) from 29 to 47% in patients 60–80 years old and 3-year EFS or progression-free survival (PFS) from 59 to 79% in patients aged 18–60 years. More importantly, immuno-chemotherapysignificantly improved overall survival (OS) [3]. The use of rituximab has been reported to be low in Indian populations (42.7%) and other developing countries due to financial constraints and poor access to healthcare [4]. Prognostic indices aid in the prognosis of individual NHL subtypes. A clinical prognostic model, termed the International Prognostic Index (IPI), was developed, in the pre-rituximab era, with five factors age, stage, serum LDH, performance status (ECOG) and extra-nodal site, which stratifies patients into quartiles correlating with 5 year DFS (disease free survival) [5].

Sub-classification of DLBCL to identify high risk group for better treatment strategy and outcome have been attempted, based on cell of origin by immuno-histochemistry into Germinal centre (GCB) type and Activated B cell (ABC) type (non-Germinal type) and, into double hit and double expresser lymphomas based on molecular subtyping. There have been various immuno-histochemistry based algorithms like Hans, Muris, Choi and Tally to classify into subtypes. The commonly used algorithm is Hans algorithm which takes into consideration CD10, BCL-6 and MUM1 to classify into 2 groups namely GCB and Non GCB. Bcl-6 being a B-cell marker with varying reproducibility, Meyel et al. analysed modified-Hans algorithm with omission of Bcl-6 and still retained the concordance to about 86% with GEP [6]. This classification has prognostic and predictive value however treatment decisions are yet to be customized to achieve better outcomes. This classification will help in development of personalized therapy with intensification for high risk DLBCL. The aim of this study was to determine the clinicopathologic and treatment outcomes of patients with DLBCL based on cell of origin classification.

Materials and Methods

Medical case records for all newly diagnosed diffuse large B cell lymphoma (DLBCL) patients treated in our institute from January 2015 to July 2017 were analysed for this study. Institute Ethical approval for this study was obtained. Out of the 232 NHL patients registered during the study period, 111 (47.8%) DLBCL patients were identified. Newly diagnosed DLBCL patients with age greater than 18 years were included for the study. Patients previously treated and relapsed DLBCL were excluded. Based on the above criteria, ninety-five patients were included for the study while sixteen patients were excluded (10 patients were relapsed DLBCL and 6 patients were less than 18 years). Details about clinical presentation, cell of origin subtype, treatment outcomes in terms of response and survival were recorded. Paraffin-embedded tissue blocks were identified and 3 µm thick sections were taken for IHC staining. Staining for BCL-6, CD10 and MUM1 were performed and positivity was taken when more than 30% cells were stained with respective antibody. The morphology of tumour cells was evaluated. Hans Algorithm was used for classification into GCB and non-GCB subtype. Cases were allocated to the GCB group if CD10 alone was positive. CD 10 negative cases were tested for BCL-6. If BCL-6 was negative it was grouped under non-GCB group. Cases with BCL-6 positivity was further stained with MUM1, if MUM1 was negative, then it was grouped to GCB subtype and positive cases under non-GCB group [6]. Staging was done using CT scan or PET scan and Bone marrow biopsy was done in all the patients.

To analyze the data, we performed descriptive statistics with frequencies expressed as numbers, Median with interquartile range and inferential statistics, contrasting between groups using a Chi square or Fisher’s exact test for the categorical variables with a level of significance p < 0.05. Survival analysis was performed using the Kaplan–Meier method. Overall Survival (OS) was defined as time from diagnosis to death due to any cause. Disease free Survival (DFS) was defined as time from diagnosis to relapse or death due to any cause. All statistical analysis was performed by using SPSS 19 version software.

Results

A total of 95 newly diagnosed patients of DLBCL were included for the study. Baseline clinical and demographic characteristics of these patients are described in Table 1. Immuno-phenotypic subtype information was available in 71 patients (74.7%). Based on the cell of origin classification, GC type was seen in 24 (25%), non-GC type was seen in 47 (50%) and unknown in 24 patients (25%). Seventy five percent of patients received R-CHOP chemotherapy as shown in Table 2.

Table 1.

Clinical characteristics of DLBCL (n = 95)

Sl. no.	Variables	GCB n = 24	Non GCB n = 47	Unknown n = 24	p value
		N (%)	N (%)	N (%)
1	Age (years)				0.35
	Median (range)	56 (26-74)	54 (20-74)	48 (18-77)
	< 60	16 (66.7)	37 (78.7)	20 (83.3)
	≥ 60	8 (33.3)	10 (21.3)	4 (16.7)
2	Gender				0.36
	Male	12 (50)	31 (66)	16 (66.7)
	Female	12 (50)	16 (34)	8 (33.3)
3	ECOG PS				0.4
	1	19 (79.2)	30 (63.8)	17 (70.8)
	≥ 2	5 (20.8)	17 (36.2)	7 (29.2)
4	Stage (Ann arbor)				0.11
	Early	11 (45.8)	27 (57.4)	18 (75)
	Advanced	13 (54.2)	20 (42.6)	6 (25)
5	B-symptoms				0.36
	Present	11 (45.8)	24 (51.1)	8 (33.3)
	Absent	13 (54.2)	23 (48.9)	16 (66.7)
6	Primary nodal v/s extra-nodal				0.002*
	Nodal	17 (70.8)	30 (63.8)	6 (25.0)
	Extra-nodal	7 (29.2)	17 (36.2)	18 (75.0)
7	IPI				0.31
	Low (0, 1, 2)	13 (54.2)	31 (66)	19 (79.2)
	Intermediate [3]	7 (29.2)	13 (27.7)	3 (12.5)
	High [4, 5]	4(16.7)	3 (6.4)	2 (8.3)
8	Bulky disease				0.048*
	Present	5 (20.8)	22 (46.8)	6 (25)
	Absent	19 (79.2)	25 (53.2)	18 (75)
9	Bone marrow involvement				0.36
	Present	4 (16.7)	5 (10.6)	1 (4.2)
	Absent	20 (83.3)	42 (89.4)	23 (95.8)

*p < 0.05 was considered statistical significant

Table 2.

Distribution of DLBCL patients according to chemotherapy regimen received in overall population

S. no.	Chemotherapy regimen	N = 95	%
1.	R-CHOP	72	75.8
2.	CHOP	9	9.5
3.	Other regimena	10	10.4
4.	RT only	1	1.1
5.	Supportive care	3	3.2

^aOther regimen: COPAD-M, CVP, DA-R-EPOCH, pre phase, R-CVP

In our study primary extra-nodal involvement was seen in 42 patients. Most common site of extra nodal presentation was gastro-intestinal tract comprising 50% followed by head and neck in 26% of patients. The cell of origin (COO) subtype was unknown in majority of the extra-nodal presentation. Treatment related adverse events were assessed in 87 patients. Neuropathy of any grade during treatment was seen in 3 (3.4%) patients. Febrile neutropenia of any grade developed in 21 (24.1%) patients. In-patient admission was required in 19 (21.8%) for grade 3/grade 4 toxicity management or supportive care. Primary or secondary G-CSF prophylaxis was required in 39 (44.8%) patients.

Survival analysis was performed in 92 patients who received at least one cycle of chemotherapy. The median duration of follow up was 17.5 months (0.86–47). The 2-year overall survival (OS) and disease-free survival (DFS) was 74%, and 61% respectively. Based on the immunophenotypic subgroup, the 2-year DFS for GCB was superior (70%) when compared to non-GCB (53%) but was not statistically significant. The 2-year OS in GCB and non-GCB patients was 80% and 70% respectively (Figs. 1, 2). The 2 years OS in patients treated with rituximab based regimen (77%) showed a significantly better survival compared to non-rituximab based regimen (56%), p value-0.05.

Fig. 1.

Kaplan meier curve for overall survival

Fig. 2.

Kaplan meire curve for disease free survival

Discussion

Baseline demographic characteristics of our study population was similar to other studies that reported outcomes in lymphomas from India [4, 7–11]. The median age was 54 years with presence of bulky disease in 35%, extra nodal presentation in 44% in the overall DLBCL study group. Our patients were a decade younger as compared to western data which is possibly due to large proportion of Indian population is being of younger age group. The frequency of extra-nodal presentation and bulky disease was  two times higher as compared to data from other centers in India [8, 10]. The incidence of extra-nodal lymphomas among all NHL varies across different countries from 24 to 48%. Seventy-two (75.8%) patients in our series, received R-CHOP (rituximab based) chemotherapy, whereas use of rituximab was reported to be in the range of 18–38% in registry based collective data from private and public institutions in India [10]. Widespread use of the government provided insurance (especially to lower income group patients) and policy changes in drug pricing made rituximab more accessible to patients. Most of our patients tolerated chemotherapy well with infection rate of 31% and febrile neutropenia rate of 24.1%. One death due to chemotherapy toxicity was reported.

The frequency of GCB in our study was 25% which was significantly lower than studies from Gogia et al. [7] and Dwivedi et al. [12] which reported a frequency of 50% (Table 3). It is not clear if there is a regional variation or a referral bias. In about 25% of our patients the cell of origin status was unknown. There is data on variation in the frequency distribution of the subtypes from different parts of the world. In a study by Schiozawa et al. the frequency of GCB subtype was 30% in the Asian countries which was similar to our study [13]. The western population had a higher frequency of GCB subtype seen in 58% of patients [14]. The Hans algorithm uses CD10, Bcl-6 and MUM1 in that order (Bcl-6 may be omitted in modified-Hans) and has a sensitivity of 90% and specificity of 50–80%. When compared to gene expression profiling, Hans algorithm has a positive predictive value of 55–80% and negative predictive value of 80–90% [15]. However due to ease and availability of the IHC method, the minor reduction in predictive/prognostic ability of this algorithm is acceptable, as this may still be used in the clinics. The cell of origin identified by this technique can differentiate the clinical presentation and treatment outcomes between the groups reasonably well. The median age at presentation was similar among the GCB and non-GCB subgroups. The gender, stage at presentation, B symptoms, extra-nodal and bone-marrow involvement were comparable between the groups. This was in concordance with a study reported by Dwivedi et al. [12].There are very few studies from India looking at the survival based on subtypes. In our study, GCB had a favorable DFS when compared to non-GCB subtype which was similar to the western data [14]. In the study by Dwivedi et al. from India, the non-GCB had a better relapse free survival of 78% versus 56% in the GCB group which was contrasting to the results from our study [12].

Table 3.

Comparison to other studies of DLBCL from Indian subcontinent

Study	Nimmagadda et al. [10]	Prakash et al. [16]	Ganesan et al. [5]	Gogia et al. [7]	Current study
Period	2010–2012	2012	2000–2013	2013–2015	2015–2017
Number of patients	791	185	444	267	99
Median age	52	50	47	49	56
Rituximab uses	42% Ritux based	18%	27%	45%	75%
			(> 90% in 2014–2015)
ORR(objective response rate)	66%	73%	79% CR	84%	NA
Survival analysis	2 year PFS = 75%	4 year EFS = 54%	5 year EFS = 59%	2 year EFS = 70%	2 year DFS = 61%
	OS = 79%	OS = 64%	OS = 68%		OS = 74%
Status of cell of origin	Not analysed	Not analysed	Not analysed	GCB = 55%	GCB = 25%
				ABC = 45%	NonGCB = 50%
					Unknown 25%.
Survival based on GCB versus nonGCB	NA	NA	NA	NA	OS: 80% versus 70%
					DFS: 70% versus 53%

Our study had the limitation of being from a single centre and retrospective in nature with small sample size and shorter duration follow up. There were 25% of patients with unknown COO data. Although our study was not powered for survival, there was a trend towards poorer survival in non-GCB patients which needs to be further explored and confirmed from our country.

Conclusion

Our retrospective analysis reported the outcomes of DLBCL based on cell of origin. There was lower frequency of GCB subtype albeit with better survival than the non-GCB subtype. This is one of the few reports from Indian sub-continent studying the outcomes of DLBCL on cell of origin. We would suggest that IHC based cell of origin classification be used routinely. Further studies looking at intensification of therapy to improve outcomes in non-GCB subtype are suggested.

Abbreviations

GCB

Germinal centre B cell

Non-GCB

Non-Germinal centre B cell

NHL

Non-Hodgkin lymphoma

DLBCL

Diffuse large B cell lymphoma

OS

Overall survival

DFS

Disease free survival

Compliance with Ethical Standards

Conflict of interest

None.

Ethical Approval

Institute ethical approval for this study was obtained (Approval no: JIP/IEC/2017/0379).