Risk scores for monoclonal B-cell lymphocytosis in patients with European ancestry and chronic lymphocytic leukemia in African Americans

Monoclonal B-cell lymphocytosis (MBL) is a precursor to chronic lymphocytic leukemia (CLL) and is defined as a pre-malignant condition with < 5 × 10⁹ clonal B-cells/L in the peripheral blood without lymphadenopathy, cytopenias or organomegaly. Information regarding the risk factors for developing MBL—aside from age, sex, and family history of CLL—remain sparse. Single nucleotide polymorphisms (SNPs) have been investigated in association with CLL risk, with 41 found to be associated with increased risk in people with European ancestry (EA); these 41 SNPs were used to create a polygenic risk score (PRS). Together these SNPs explain ~25% of additive heritable risk.

To explore this area further, Geffen Kleinstern and colleagues examined the risk of MBL in patients of EA, as well as the risk of CLL in patients of African American (AA) heritage.¹

Study design

MBL cohort and controls

Two cohorts of individuals of EA were selected: a screening cohort and a clinical cohort. The screening cohort used samples of peripheral blood mononuclear cells from 3,041 asymptomatic adults; these were screened for MBL using flow cytometry. Of these patients, 410 had CLL phenotype MBL, and the remaining 2,631 were used as controls. The second MBL cohort was made up of high-count (HC)-MBL patients.

CLL patients and controls

Four different studies were used to provide patients with CLL of AA or EA, and the characteristics of these patients are shown in Table 1.

Table 1. Baseline patient characteristics*

CLL, chronic lymphocytic leukemia; HC, high count; LC, low count; MBL, monoclonal B-cell lymphocytosis. *Data from Kleinstern et al.1
Characteristic	European ancestry					African Americans
	Controls n = 2,631	MBL overall n = 560	LC-MBL n = 396	HC-MBL n = 164	CLL n = 696	Controls n = 235	CLL n = 173
Male, %	39	57	55	62	69	68.1	64.9
Median age, years (range)	64 (29−101)	70 (43−97)	72 (44−95)	68 (43−97)	62 (30−94)	61 (40−90)	59 (26−94)

This study included 3,887 patients of EA and 408 of AA heritage. Out of the EA controls, only 39% were male whereas in the MBL and CLL groups the percentage of male patients was between 55−69%. Median age was similar across all groups.

Results

In controls and cases of CLL among patients with EA, the odds ratios (ORs) of 40 (98%) out of the 41 SNPs were consistent with those reported in previous studies, and 32 (78%) were statistically significant at p < 0.05. In the AA group, ORs of 22 (54%) of the SNPs were directionally consistent with those reported in previous studies, and only two SNPs, rs7690934 (OR = 1.41, CI: 1.03−1.95; p = 0.03) and rs1679013 (OR = 1.56, CI: 1.08−2.25; p = 0.02), were nominally significant.

Between the EA and AA populations, the median difference in the minor allele frequency (MAF) across the 41 SNPs was 7.2% (0.2−26%), with most of the MAF in the AA population being lower than in the EA population. Variability in MAF may, therefore, play a role in the lack of statistical significance of the observed ORs of the SNPs in the AA population.

The observed ORs for 39 (95%) of the 41 SNPs in the MBL cohort were consistent with previous reports in CLL patients; 21 (51%) of the 41 SNPs were nominally statistically significant (p < 0.05), while 15 of the 41 SNPs showed little evidence of an association (OR < 1.1).

CLL-PRS and risk of MBL

The median CLL-PRS was:

• 7.90 for 560 MBLs
• 7.46 for 2,631 controls of EA, with a consistent distribution

The continuous PRS had 1.86-fold increased risk for MBL (CI: 1.67−2.07; p = 1.9 × 10−29), with a c-statistic of 0.72 (Table 2).

Quintile (Q)5 had a 2.38-fold increased risk for MBL compared with Q3 (CI: 1.81−3.13; p = 5.5 × 10⁻¹⁰), and Q1 had a 54% reduced risk vs Q3 (OR = 0.46, CI: 0.32−0.66; p = 2.9 × 10⁻⁵). The 99th percentile (5.5% of MBL) compared to Q3 demonstrated a 4.83-fold increased risk of MBL (CI: 2.81−8.31; p = 1.3 × 10⁻⁸).

Table 2. PRS and monoclonal B-cell lymphocytosis risk in patients of European Ancestry*

CI, confidence interval; MBL, monoclonal B-cell lymphocytosis; OR, odds ratio; PRS, polygenic risk score; Q, quintile; Ref, reference. *Data from Kleinstern et al.1 †Quintiles based on 7,983 controls. ‡Controls were screened negative for MBL. §Adjusted for age and sex.
PRS†	Controls‡ n = 2,631	MBL n = 560	MBL vs controls
			OR§	95% CI	p value
Q1, % (4.32, 6.80)	22	9	0.46	0.32−0.66	2.9 × 10−5
Q2, % (6.80, 7.32)	20	16	0.88	0.64−1.2	0.42
Q3, % (6.80, 7.32)	21	18	1	Ref	—
Q4, % (6.80, 7.32)	18	20	1.32	0.97−1.79	0.07
Q5, % (6.80, 7.32)	19	37	2.38	1.81−3.13	5.5 × 10−10
Continuous	—	—	1.86	1.67−2.07	1.9 × 10−29
Continuous unweighted	—	—	1.15	1.13−1.18	2.4 × 10−31
PRS (median), n	7.46	7.90	—	—	—
c-statistic	—	—	0.72	0.69−0.73	—
c-statistic unweighted	—	—	0.72	0.70−0.74	—

Risk of MBL subtypes

Out of the 396 patients with low-count MBL (LC-MBL), 10% were in PRS-Q1 and 34% were in PRS-Q5. The PRS progressively increased from controls to LC-MBL, HC-MBL, and CLL (Table 3). For LC-MBL, the continuous PRS had a 1.75-fold increased risk, which was significant (CI: 1.55−1.98; p = 7.5 × 10⁻¹⁹).

Table 3. PRS and associations with monoclonal B-cell lymphocytosis subtypes and CLL in European Ancestry individuals*

CLL, chronic lymphocytic leukemia; HC, high count; LC, low count; MBL, monoclonal B-cell lymphocytosis; OR, odds ratio; PRS, polygenic risk score; Q, quintile. *Data from Kleinstern et al.1 †Quintiles based on 7,983 controls. ‡Controls were screened negative for MBL.
PRS†	Controls‡ n = 2,361	LC-MBL n = 396	HC-MBL n = 164	CLL n = 696
Q1, % (4.32, 6.80)	22	10	6	5
Q2, % (6.80, 7.32)	20	16	17	10
Q3, % (7.32, 7.77)	21	19	16	16
Q4, % (7.77, 8.28)	18	21	17	20
Q5, % (8.28, 11.31)	19	34	44	49
PRS, median	7.46	7.84	8.05	8.24

For HC-MBL, the continuous PRS showed a 2.14-fold increase in risk (CI: 1.80−2.56; p = 3.9 × 10⁻¹⁷), with a c-statistic of 0.73 (CI: 0.69−0.77) (Table 4). Comparing PRS between the four groups, a significant difference was demonstrated (p = 4.3 × 10⁻¹²).

Table 4. PRS and ORs of monoclonal B-cell lymphocytosis subtypes and CLL vs controls in European Ancestry individuals*

CI, confidence interval; CLL, chronic lymphocytic leukemia; HC, high-count; LC, low-count; MBL, monoclonal B-cell lymphocytosis; OR, odds ratio; PRS, polygenic risk score; Q, quintile; Ref, reference. *Data from Kleinstern et al.1 †Quintiles based on 7,983 controls. ‡Adjusted for age and sex.
PRS†	LC-MBL vs control			HC-MBL vs control			CLL vs control
	OR‡	95% CI	p value	OR‡	95% CI	p value	OR‡	95% CI	p value
Q1 (4.32, 6.80)	0.51	0.34−0.76	0.001	0.33	0.15−0.70	0.004	0.31	0.21−0.46	1.0 × 10−8
Q2 (6.80, 7.32)	0.81	0.57−1.17	0.27	1.05	0.61−1.82	0.85	0.65	0.47−0.90	0.01
Q3 (7.32, 7.77)	1	Ref	—	1	Ref	—	1	Ref	—
Q4 (7.77, 8.28)	1.33	0.95−1.88	0.1	1.31	0.76−2.26	0.34	1.56	1.17−2.08	0.002
Q5 (8.28, 11.31)	2.10	1.53−2.88	4.0 × 10−6	3.13	1.97−4.98	1.0 × 10−6	3.49	2.70−4.51	1.2 × 10−21
Continuous	1.75	1.55−1.98	7.5 × 10−19	2.14	1.80−2.56	3.9 × 10−17	2.53	2.27−2.81	4.0 × 10−63
Continuous unweighted	1.14	1.11−1.17	1.3 × 10−20	1.19	1.14−1.23	1.9 × 10−17	1.23	1.20−1.26	4.0 × 10−63
c-statistic	0.72	0.70−0.75	—	0.73	0.69–0.77	—	0.77	0.75−0.79	—
c-statistic unweighted	0.72	0.70–0.75	—	0.725	0.69–0.77	—	0.775	0.76−0.79	—

CLL-PRS was calculated for the smaller group of AA individuals, with a median PRS of 7.25 for the controls and 7.53 for the CLL group. There was a 1.76-fold increased risk of CLL (Table 5).

Table 5. PRS and associations with different CLL subtypes in African American individuals*

CLL, chronic lymphocytic leukemia; MBL, monoclonal B-cell lymphocytosis; OR, odds ratio; PRS, polygenic risk score; Q, quintile; Ref, reference. *Data from Kleinstern et al.1 †Quintiles based on 7,983 controls. ‡Adjusted for age and sex.
PRS†	Controls (n = 235)	CLL (n = 173)	CLL vs control
			OR‡	95% CI	p value
Q1, % (4.32, 6.80)	20.0	8.1	0.47	0.22−0.99	0.048
Q2, % (6.80, 7.32)	20.0	15.6	0.88	0.45−1.70	0.70
Q3, % (6.80, 7.32)	19.6	17.3	1.00	Ref	—
Q4, % (6.80, 7.32)	20.0	28.9	1.60	0.87−2.95	0.13
Q5, % (6.80, 7.32)	20.4	30.1	1.61	0.88−2.96	0.12
Continuous	—	—	1.76	1.34−2.31	5.1 × 10−5
Continuous unweighted	—	—	1.07	1.01−1.13	0.03
PRS (median), n	7.25	7.53	—	—	—
c-statistic	—	—	0.62	0.57−0.68	—
c-statistic unweighted	—	—	0.57	0.53−0.64	—

A self-reported questionnaire was also filled out for 2,512 controls and 365 patients with MBL, almost all of whom had LC-MBL. As previously reported, age (in 10-year increments; OR = 1.83, CI: 1.64−2.04; p < 0.0001) and male sex (OR = 1.73, CI: 1.38−2.15; p < 0.0001) were significantly associated with increased risk of MBL. Compared with the controls, patients with MBL had a higher likelihood of familial leukemia/lymphoma, though the difference was not significant (OR = 1.44, CI: 0.99−2.09; p = 0.06). No other exposures were found to be significant after adjusting for age and sex.

Conclusion

This study demonstrates that a heritable factor, rather than environmental factors, exists for MBL development. Of the 41 CLL-susceptibility SNPs, 50% were found to be associated with MBL risk among individuals of EA. CLL-PRS was shown to be a significant predictor of CLL risk in people of EA, and to a lesser extent, AAs. The results of the study may help to predict which patients are at greater risk of developing MBL and CLL in different populations.