Contributions
Abstract: PB1777
Type: Publication Only
Background
Traditional International Prognostic Index before (tIPI) and post-rituximab era[Revised IPI (R-IPI) and enhance National Cancer Comprehensive Network-IPI (NCCN-IPI)] has been used to obtained risk groups for overall survival (OS) and progression free survival (PFS). Serum albumin (SA) has proven to be a prognostic marker in many hematological malignancies, including diffuse large B-cell lymphoma (DLBCL), but it was not analized in any IPI. SA could be a surrogate for age, poor nutritional status, comorbid status, and aggressive disease.
Aims
Analyze in DLBCL patients the predictive factor of albumin in relation to OS.
Methods
From 2010 to 2016, 5 databases of public and private institutions from 4 different states were analyzed retrospectively. Where de novo DLBCL patients were identified and obtained 113 from 160 patients with next inclusion criteria: 18 years old and older, and R-CHOP treatment. Minimum the following clinical data for calculating tIPI,R-IPI and NCCN-IPI : Age, lactic dehydrogenase (LDH), performance status (PS, ECOG scale), number of extranodal disease (EN), Ann Arbor stage (localized vs. Advance), and others like SA. Diagnosis of DLBCL was confirmed through review the initial pathology report. 47 patients were excluded because they did not have some variable before mentioned. Primary clinical outcome was defined as date of diagnosis to date of death or date of last contact for those censored. For the descriptive statistics the square Chi was used. Time-to event data (OS) were estimated using the KaplanMeier method. Hazard Ratios (HR) were used to identify potential risk factors in different risk groups conformed. To obtain a correct cut-off point for albumin, we used the ROC curve and later we evaluated the best positive and negative likehood ratio (LHR+ and LHR-, respectively)
Results
From 113 DLBCL-patients, median age was 56.5 (range: 19-89); 52% were women; 58% were advanced stage (III/IV Ann Arbor); age >60 y/o,47%, but 15% >75 y/o; ECOG 2, 28.3%; EN disease (≥1), 35.5%; LDH elevated, 57.2%; Albumin <4g/dL, 46%. The median follow-up of whole series was 19 months; the median follow-up depending on the response after treatment was for complete response (CR): 32 months, partial response (PR): 13.6 months and Progression/failure (P): 8 months. The OS curves creation for tIPI, R-IPI and NCCN IPI was in accordance with already reported in other series.
the cut-off point for albumin using the ROC curve was ≤ 3.7g/dL (Sensitivity 63.4%, Specificity 80.6, AUC 0.77; p<0.001). Interestingly, the LHR+ in 19.32 and LHR- in 0.74 were found in accordance to 2.7 g / dL cut-off point of albumin, thus identifying a value lower than the original point of 3.7 of albumin that could be associated with higher mortality. Then, we can build two risk groups: one called low (3.7 to 2.8g/dL) and another ultra low-albumin (≤2.7g/dL). When we capture these groups in the OS curves, 3 groups can be identified perfectly (Figure 1)
Conclusion
The low and ultra low levels of serum albumin are a strong predictive factor for the early mortality of patients with DLBCL.
These findings guarantee to evaluate this biomarker more closely
Session topic: 21. Aggressive Non-Hodgkin lymphoma - Clinical
Keyword(s): DLBCL, Survival prediction
Abstract: PB1777
Type: Publication Only
Background
Traditional International Prognostic Index before (tIPI) and post-rituximab era[Revised IPI (R-IPI) and enhance National Cancer Comprehensive Network-IPI (NCCN-IPI)] has been used to obtained risk groups for overall survival (OS) and progression free survival (PFS). Serum albumin (SA) has proven to be a prognostic marker in many hematological malignancies, including diffuse large B-cell lymphoma (DLBCL), but it was not analized in any IPI. SA could be a surrogate for age, poor nutritional status, comorbid status, and aggressive disease.
Aims
Analyze in DLBCL patients the predictive factor of albumin in relation to OS.
Methods
From 2010 to 2016, 5 databases of public and private institutions from 4 different states were analyzed retrospectively. Where de novo DLBCL patients were identified and obtained 113 from 160 patients with next inclusion criteria: 18 years old and older, and R-CHOP treatment. Minimum the following clinical data for calculating tIPI,R-IPI and NCCN-IPI : Age, lactic dehydrogenase (LDH), performance status (PS, ECOG scale), number of extranodal disease (EN), Ann Arbor stage (localized vs. Advance), and others like SA. Diagnosis of DLBCL was confirmed through review the initial pathology report. 47 patients were excluded because they did not have some variable before mentioned. Primary clinical outcome was defined as date of diagnosis to date of death or date of last contact for those censored. For the descriptive statistics the square Chi was used. Time-to event data (OS) were estimated using the KaplanMeier method. Hazard Ratios (HR) were used to identify potential risk factors in different risk groups conformed. To obtain a correct cut-off point for albumin, we used the ROC curve and later we evaluated the best positive and negative likehood ratio (LHR+ and LHR-, respectively)
Results
From 113 DLBCL-patients, median age was 56.5 (range: 19-89); 52% were women; 58% were advanced stage (III/IV Ann Arbor); age >60 y/o,47%, but 15% >75 y/o; ECOG 2, 28.3%; EN disease (≥1), 35.5%; LDH elevated, 57.2%; Albumin <4g/dL, 46%. The median follow-up of whole series was 19 months; the median follow-up depending on the response after treatment was for complete response (CR): 32 months, partial response (PR): 13.6 months and Progression/failure (P): 8 months. The OS curves creation for tIPI, R-IPI and NCCN IPI was in accordance with already reported in other series.
the cut-off point for albumin using the ROC curve was ≤ 3.7g/dL (Sensitivity 63.4%, Specificity 80.6, AUC 0.77; p<0.001). Interestingly, the LHR+ in 19.32 and LHR- in 0.74 were found in accordance to 2.7 g / dL cut-off point of albumin, thus identifying a value lower than the original point of 3.7 of albumin that could be associated with higher mortality. Then, we can build two risk groups: one called low (3.7 to 2.8g/dL) and another ultra low-albumin (≤2.7g/dL). When we capture these groups in the OS curves, 3 groups can be identified perfectly (Figure 1)
Conclusion
The low and ultra low levels of serum albumin are a strong predictive factor for the early mortality of patients with DLBCL.
These findings guarantee to evaluate this biomarker more closely
Session topic: 21. Aggressive Non-Hodgkin lymphoma - Clinical
Keyword(s): DLBCL, Survival prediction