Abstract

Introduction.

In the past years there were many attempts to replace or improve the IPI score for overall survival evaluation (OS) in diffuse large B cell lymphoma (DLBCL). We focused our attention on prognostic factors, studied in the past or in recent years, that could be used as surrogates of the disease status. We considered the serum albumin (SA) level as surrogate of age, comorbidity and disease severity; LDH as surrogate of cell activity and disease stage; absolute monocyte count (AMC) as surrogate of inflammatory status and absolute lymphocyte count (ALC) as surrogate of immune status of patients.

Patients and method

We collected data of 322 patients comprehensive of IPI factor, SA, AMC and ALC recorded in the "Gruppo Itliano Studio Linfomi" (GISL) database from 2003 to 2012. The score was obtained summing the factors SA <3.7 g/dL, LDH >UNL, AMC>630 /uL and ALC <840 /uL, any with weight one. The factors were not the results of statistical procedure but were chosen on the basis of clinical and pathophysiological considerations, about the role played by these factors in the DLBCL. The OS was estimated using the Kaplan-Meier method; the discriminate ability of the score was checked by means of log-rank test and using the ROC curve, sensibility and specificity at 3-years of follow-up.

Results

All the 322 patients included in the study were treated with R-CHOP and R-CHOP like regimens. The median age at diagnosis was 67 years (range 22-86) and 54% were males; SA <3.7 was identified in 44% of the patients, LDH>UNL in 51%, AMC >630 in 33% and ALC <840 in 23%. The sum of those four factors defined a score with three levels of risk: low (0-1, n=69, 21%), intermediate (2, n=94, 29%) and high (3-4, n=159, 49%). The OS at 5-yrs was 96%, 76% and 46% for score 0-1, 2 and 3-4, respectively The log-rank between score 0-1 and 2 was 11.2 (p=0.0008) and 16.9 between score 2 and 3-4 (p<0.0001). The ROC curve at 5-yrs was 72.1%, with 75.8% of sensitivity and 68.5% of specificity. The score showed a moderate accordance with IPI coded 0-1, 2 and 3/5 (k-statistic 0.37).

Conclusion

This approach showed that it is possible to define a simple prognostic score taking into account simple factors, easily available in every hospital, and related to the disease status. This selection was based on the clinical and pathophysiological knowledge accumulated over the years and, also, avoiding complicated statistical procedures for the selections of the prognostic factors. As it appears in the graph (Fig.1) , patients can be clearly stratified in three groups with deeply different survival outcome, based only on biochemical factors.

Disclosures

No relevant conflicts of interest to declare.

Author notes

*

Asterisk with author names denotes non-ASH members.