Geltamo-IPI Distinctly Identifies a Very High Risk Group in Difuse Large B-Cell Lymphoma (DLBCL) of Non B-Germinal-Center Origin Treated with Chemo-Immunotherapy

Backgound. GELTAMO-IPI has shown to have better prognostic effect than NCCN-IPI for DLBCL treated with immunochemotherapy (BJH;2017;176;918) with the additional advantages to identify a real high risk group and not being influenced by the primary extranodal presentation or the use of treatment more intense than R-CHOP. The variables in GELTAMO-IPI and its statistical weight are: age (<65 years: 0 points; ≥65-79:1 pt.; ≥80 2 pt.), PS (0-1: 0 p.; 2: 1pt. ≥2: 2 pt.), >LDH, > β₂mcg and stage IV, one point each. The combination of these factors separate four risk groups, Low risk (LR, 0 points), Low-Intermediate risk (LIR, 1-3 points), High-Intermediate risk (HIR, 4 points and High Risk (HR≥5 points), with 5-year overall (5y-OS) survival of 93%, 79%, 66% and 39%, respectively.

Aims. Present study explores the prognostic effect of GELTAMO-IPI in the histological subgroups depending of the cell of origin: B Germinal Center (BGC) or non B Germinal Center (Non-BGC).

Methods. the original series for the development of GELTAMO-IPI comprised 1848 patients. The BGC/non-BGC origin was calculated with the Hans algorithm by the Immunohistochemical expression of 3 markers (CD10, BCL6 and MUM-1) being available in 1097 patients. The statistical study consisted in the development of a Cox model which includes the three variables: GELTAMO-IPI groups (LR, LIR, HIR, HR), cell of origin (BGC/non-BGC) and its product in order to evaluate the presence of interaction between these variables. Kaplan-Meier and Log-rank were used for survival curves and comparison.

Results. The cell of origin was retrieved in 1097 patients: 380 BGC, 459 non-BGC and 258 non-evaluable. The series includes: LR 72 p. (9.2%), LIR 464 p. (59.5%), HIR 127 p. (16.3%) and HR 117 p. (15%) with 5y-OS of 98%, 79%, 67% and 41%, respectively. The Cox model showed interaction between the risk groups and the GBC/non-GBC origin (p=0.005), indicating that GELTAMO-IPI has a different effect in each group. In the BGC-DLBCL group with LR (34 p.), LIR (221 p.), HIR (53 p.) and HR (45 p.) the 5y-OS differences between LR and LIR are maintained (100%, 75%, respectively), but not between HIR and HR (53% and 51%); Nevertheless, GELTAMO-IPI stratified 3 distinct risk groups (LR, LIR and HIR+HR), although it is not able to identify a real high risk group of patients (Figure 1, left panel). In the Non-BGC-DLBCL the differences among the LR, LIR and HIR groups (38, 243, and 74 patients with 5y-OS of 97%, 82% and 77%) have lower significance than in the original study (p=0.01 and p=0,26, respectively), but a distinct HR group of 72 patients with a 5yOS of 35% is clearly identified (p<0.001) (Figure 1, right panel).

Conclusions. GELTAMO-IPI separates three risk groups in BGC-DLBCL, but fails to identify a genuine HR group, whereas GELTAMO-IPI in the non-BGC-DLBCL positively identifies a real high risk group. This different outcome in the GC-DLBCL and non-GC-DLBCL using simple clinical factors, as included in GELTAMO-IPI, represents an interesting practical prognostic tool in the clinical practice.

Figure 1

View largeDownload slide

Figure 1

View largeDownload slide

Close modal