Presenting Features Predictive of Long-Term Stability in Chronic Lymphocytic Leukemia.

Abstract

Prognostic factors to predict an aggressive clinical course in chronic lymphocytic leukemia (CLL) such as CD38, ZAP-70 and IgH mutation status have been well described, however, readily available presenting features predictive of long-term stability are less well defined. We performed a retrospective analysis of 335 consecutive patients with CLL seen at St. Paul’s Hospital from January 1969 to July 2005. Patients were identified from the practice data-base and clinical and pathological data abstracted by chart review. Long-term stability was defined as no requirement for treatment for at least 6 years (n=66, group A). Characteristics of group A were compared to all other pts (group B, n=269), and, to avoid follow-up bias, to pts followed for ≥6 y who required treatment (group C, n=62). 65 pts in group A were B-cell in phenotype and 1 was T-cell; group B 268 pts B-cell, 1 T-cell; group C all 62 B-cell. Median age at diagnosis for groups A, B and C were 59.5 (range 33–80) y; 68 (30–94) y; and 60 (30–82) y, respectively (p<0.0001 for A vs. B and p<0.42 for A vs. C). Male pts comprised 52%, 54.3% and 53% of groups A, B and C respectively (p<0.85). Rai stages were: stage 0, 1, 2, 3, and 4; 111, 48, 47, 3 and 6 respectively. ECOG Performance Status was 0 in 307 pts and was not analyzed further. Lymphocyte count at diagnosis (LCD) for groups A, B and C were 9 (4–50 X 109/L), 10 (3–394) and 12 (3–155; p<0.0001 for A vs. B and p< 0.01 for A vs. C); in group A, 25 (38%) of pts had a LCD ≤10, as did 81 (30%) in group B and 10 (16%) in group C. In groups A, B and C, 9 (13.6%), 21 (11.5%), and 7 (11.3%) pts, respectively, had a lymphocyte doubling time (LDT) ≤12mo (to an absolute value ≥50X109/L; p<0.65). Immunophenotyping was available in 202 pts; 173 were CD5+ CD19+, 2 were CD5- and 29 CD19-. Analysis for CD38 was available in 50 pts; in groups A, B and C, 5 of 36 (14.3%), 2 of 14 (16.7%) and 0 of 7, respectively, were CD38+ (defined as ≥30% of cells; p<0.84). Median follow-up for groups A, B and C was 110.5 (76–369) mo, 43 (0–309) mo, and 134.5 (72–309 mo; p<0.0001 for A vs. B and p<0.67 for A vs. C). Pts in group A did not require treatment by definition; time to treatment (TTT) for group B was median 35 (0–243) mo and group C 66 (0–243) mo (p<0.0001). Median OS for all pts was 191 (0–369) mo. Two deaths occurred in group A; 1 of Richter’s transformation at 77 mo and 1 of unknown causes at 78 mo; median OS for groups B and C was 189 (0–311) mo and 127 (69–311) mo respectively (p<0.025). In group A, 1 pt (1.5%) transformed, as did 15 (5.6%) in group B (8 Richter’s and 7 PLL) and 4 (6%) in group C (1 Richter’s, 3 PLL, p<0.46). At a cutoff in LCD of 10 and 20 the difference in OS between groups was maintained (see Table). Characteristic OS @ 120 mo (%) Lymphocyte count at Dx (X109/L) Group A (stable ≥ 72 mo, n=66) Group B (all others, n=269) Group C (not stable, ≥72 f/u mo, n=62) All pts 93 64 64 ≤10 100 100 100 >10 91 64 64 ≤20 93 38 38 >20 0 of 2 0 of 5 0 of 5 p (log rank) A vs. B p<0.023 A vs. C p<0.025 In conclusion, in this series of 335 patients with CLL, lower lymphocyte count at diagnosis predicted for long-term stability, decreased requirement for treatment and improved OS.