Abstract 2437

Clinical course in individual CLL patients (pts) is highly heterogeneous. Several pts show an indolent disease and other experience an aggressive course rapidly succumbing to disease-related events. Different biomarkers have been developed to identify pts with evolving disease but their application is mostly limited to clinical trials. It has been shown that measurement of clonal serum free light chains (sFLC) levels, through a straightforward assay validated for plasma cell disorders, could independently identify CLL pts with progressive disease. By analyzing the largest cohort of CLL pts ever reported, we wished to: i) determine the independent predictive value of sFLC abnormalities in the context of established biomarkers; ii) define the predictive hierarchy of these markers; iii) incorporate sFLC into a novel prognostic system. Cryopreserved sera at diagnosis from 449 untreated pts (282 males and 167 females, median age 65 yrs; r 33–89) were collected for analysis at two reference laboratories. After quantization (Freelite, The Binding Site, Ltd., UK) of sFLC k and l levels (r, k: 3.3–19.4 mg/mL; l: 5.71–26.3 mg/mL), their ratio was calculated as sFLC k/l (r: 0.26–1.65). An abnormal sFLCk/l was found in 150/449 cases (132 k/18 l). A significant correlation emerged between sFLCk/l and CD38, ZAP70, unmutated IgHV and unfavourable FISH but not with Binet stage (Table). At a median follow up of 3 yrs (r 1–20), 149 of 449 pts were treated due to progressive disease (NCI Guidelines). Treatment-free survival (TFS) was significantly shorter (p<0.0001) in pts with abnormal sFLCk/l (median TFS 14.2 yrs) as compared to cases with normal ratios (median TFS 4.4 yrs). Abnormal sFLCk/l remained independently associated with TFS after adjustment for CD38, ZAP70, IgVH and stage (HR=2.2, 95% C.I. 1.5–3.3, P<0.0001). In this model, ZAP70 (HR=3.2, 95% C.I. 2.0–5.1, P<0.0001), IgVH (HR=2.1, 95% C.I. 1.2–3.5, P=0.006) and stage (HR=2.7, 95% C.I. 1.7–4.2, P<0.0001) but no longer CD38 retained a significant association with risk of being treated. To further dissect the predictive value of sFLC testing, absolute levels of k and lFLCs were summed in a new variable named sFLCk+l (median value 39.0 mg/mL; r 8.2–1430.4). A higher sFLCk+l value was observed in pts who required treatment and ROC analysis disclosed that 60.6 mg/mL was the optimal cut-off value for distinguishing pts who required treatment from those who did not (AUC=0.62, P<0.0001). At 3 yrs, the probability of being therapy-free was 84.1% and 51.8% for pts with sFLCk+l below or above the threshold, respectively (Chi Square of log-rank 85.6, P<0.0001). Cox multivariate analysis demonstrated that a sFLCk+l value >60.6 mg/mL (HR=2.5, C.I. 2.6–3.9), remained the strongest independent predictor of TFS together with ZAP70 (HR=2.8, C.I. 1.8–4.6) and stage (HR=2.5, C.I. 1.6–3.8), while CD38 and IgVH status lost their predictive power. Most intriguingly, it emerged that sFLCk/l is irrelevant in predicting TFS if sFLCk+l is above cut-off and that cytogenetic risk remains significantly associated with sFLCk+l (P=0.047) but not FLCk/l. sFLCk+l retained an independent association with TFS together with cytogenetics, ZAP70 and stage. A novel prognostic model was then constructed with 1 point for each of these unfavorable marker and a score given by their sum. On 260 pts scores were: 104=0, 79=1, 56=2, 19=3, 4=4. The 3 year probability of avoiding treatment was 94.8%, 84.5%, 61.6% and 21.1% for pts scoring 0, 1, 2 and 3+4, respectively (P<0.0001). While sFLCk/l predicts TFS by mirroring the size/activity of the dominant CLL clone, the higher predictive power of sFLCk+l may reflect activity of microenvironmental oligo/polyclonal B-cells, yet clonotypic or not to the main CLL population, undergoing continuous antigen-driven clonal evolution. A score system integrating sFLCk+l levels is proposed to Identify CLL pts likely to require early treatment.

sFLC k/l in 449 untreated CLL pts

Table 1.

Death in complete remission by year

sFLC k/lP
Normal (%)Abnormal (%)
CD38    
− 234 (70.9) 96 (29.1) <0.0001* 
37 (46.3) 43 (53.7)  
ZAP70    
− 188 (72.5) 73 (28.0) 0.001* 
78 (55.7) 62 (44.3)  
IgVH    
Mut 177 (71.1) 72 (28.9) 0.038 
Unmut 78 (60.0) 52 (40.0)  
Binet    
227 (68.6) 104 (31.4) 0.2* 
B+C 72 (61.0) 46 (39.0)  
FISH    
13q14.3, normal (low risk) 186 (85.3) 32 (14.7)  
Trisomy 12 (int risk) 23 (71.0) 9 (28.1) 0.013** 
11q22.3, 17p13.1 (hi risk) 37 (69.8) 16 (30.2)  
sFLC k/lP
Normal (%)Abnormal (%)
CD38    
− 234 (70.9) 96 (29.1) <0.0001* 
37 (46.3) 43 (53.7)  
ZAP70    
− 188 (72.5) 73 (28.0) 0.001* 
78 (55.7) 62 (44.3)  
IgVH    
Mut 177 (71.1) 72 (28.9) 0.038 
Unmut 78 (60.0) 52 (40.0)  
Binet    
227 (68.6) 104 (31.4) 0.2* 
B+C 72 (61.0) 46 (39.0)  
FISH    
13q14.3, normal (low risk) 186 (85.3) 32 (14.7)  
Trisomy 12 (int risk) 23 (71.0) 9 (28.1) 0.013** 
11q22.3, 17p13.1 (hi risk) 37 (69.8) 16 (30.2)  

* Fischer's ET, **Pearson CS.

* Fisher's ET, **Pearson CS.

Disclosures:

Amoroso: The Binding Site, LtD: Consultancy.

Author notes

*

Asterisk with author names denotes non-ASH members.

Sign in via your Institution