Percentage of Smudge Cells Predicting Prognosis In Chronic Lymphocytic Leukemia

Identifying prognostic markers is important for clinical and pathological course of chronic lymphocytic leukemia. Percentage of smudge cells in CLL patients has recently been reported as a novel prognostic factor. We present here the preliminary results of the retrospective cohort study of 90 patients diagnosed with CLL at a major referral center in the state of Alabama between 1997 and 2009.

Smudge cell percentage (ratio of smudge cells to combined smudge cells and intact cells) was calculated by microscopic evaluation of archived Wright-Giemsa stained blood smears. A total of 200 cell differential was counted for inter-observer consistency. For the first time, the concordance between observers was assessed by quantization of morphological parameters of lymphocytes using computerized image analysis. Medical records of above mentioned patients were reviewed for molecular, genetic and clinical parameters including CD38, ZAP 70 expression, immunocytopenias and response to first treatment. Primary endpoint of this analysis was time to initiation of chemotherapy for first treatment (TIFT). It was calculated from date of first diagnosis until first dose of chemotherapy. Untreated patients in the follow up were censored. Kaplan Meier survival analysis and log rank test were used for statistical analysis. Prognostic factors for TIFT were analyzed with Cox's proportional hazards regression. Wilcoxon rank sum test and Fishers exact test were used to evaluate differences between groups for continuous and categorical variables, respectively. We had more than 90% Power to detect an association between smudge cell percentage and TIFT.

Of the 160 treatment naïve CLL patients enrolled for the study, 90 had archived blood smears available for evaluation. Patient characteristics include Rai stage 0 or 1 disease (87%), CD 38+ (30%), Zap 70 + (35%), 13q Del + (26%). Percentage of smudge cells ranged from 2 to 90 % (median 34%). Patients with CD38+ and ZAP 70+ on flow cytometry had lower percentage of smudge cell than those with negative markers and are trending towards significance (p=0.08). Our cohort included 16% of African American population and had no significant difference in TIFT when compared to Caucasians. In univariate analysis smudge cell percentage (stratified as less than or more than 30) was significantly associated with TIFT (p=0.04). Rai stages 0, 1 and 2 were significantly associated with longer TIFT than stage 3 and 4 (p= 0.01). Genetic parameters- deletion 13 q and 11q were associated with significantly greater TIFT (p= 0.008 and P=0.001 respectively). Patients who developed immunocytopenias during the course of study had shorter TIFT (p=0.01). We performed Step wise multivariate Cox regression analysis and identified del11q (HR=0.02, p=0.001) and Rai stage (HR=0.008, p <0.0001) as independent predictors of TIFT. Patients with longer TIFT had better response to treatment when compared to shorter TIFT (68% vs 32%, p=0.05). Contrary to our expectation, we found that, in multivariate analysis, we found smudge cell percentage is not significantly associated with TIFT (p= 0.75).

Time for initiation of first treatment is significantly shorter in CLL patients with del11q and del13q and those who develop immunocytopenias. Smudge cell percentage did not significantly predict TIFT in this single center study. We speculate this could be due to subjective decision on initiation of chemotherapy by physicians, based on several other clinical parameters and co-morbidities. Larger studies are needed that concurrently assess any confounders between smudge cell percentage and TIFT. Alternatively, computerized image analysis is superior to observer measurements and can be considered for accuracy and reproducibility of cytological evaluation. It can ascertain the objectivity of concordance in differential cell counting in similar studies.

No relevant conflicts of interest to declare.