Minimal residual disease undetectable by next-generation sequencing predicts improved outcome in CLL after chemoimmunotherapy

Abstract

Patients with chronic lymphocytic leukemia (CLL) who achieve blood or bone marrow (BM) undetectable minimal residual disease (U-MRD) status after first-line fludarabine, cyclophosphamide, and rituximab (FCR) have prolonged progression-free survival (PFS), when assessed by an assay with sensitivity 10-4 (MRD4). Despite reaching U-MRD4, many patients, especially those with unmutated IGHV, subsequently relapse, suggesting residual disease <10-4 threshold and the need for more sensitive MRD evaluation. MRD evaluation by next-generation sequencing (NGS) has a sensitivity of 10-6 (MRD6). To better assess the depth of remission following first-line FCR treatment, we used NGS (Adaptive Biotechnologies Corporation) to assess MRD in 62 patients, all of whom had BM U-MRD by multicolor flow cytometry (sensitivity 10-4) at end-of-FCR treatment. Samples from these patients included 57 BM samples, 29 peripheral blood mononuclear cell (PBMC) samples, and 32 plasma samples. Only 27.4% of the 62 patients had U-MRD by NGS. Rate of U-MRD by NGS was lowest in BM (25%), compared with PBMC (55%) or plasma (75%). No patient with U-MRD by NGS in BM or PBMC was MRD+ in plasma. Patients with mutated IGHV were more likely to have U-MRD by NGS at the end of treatment (EOT; 41% vs 13%, P = .02) than those with unmutated IGHV. Median follow-up was 81.6 months. Patients with U-MRD at EOT had superior PFS vs MRD+ patients, regardless of sample type assessed (BM, P = .02, median not reached [NR] vs 67 months; PBMC, P = .02, median NR vs 74 months). More sensitive MRD6 testing increases prognostic discrimination over MRD4 testing.

Graphical abstract

Figure 1.

Sample types tested. 28 patients had only bone marrow specimens available, 22 patients had all 3 specimens (bone marrow, peripheral blood mononuclear cells, and plasma), and 12 patients had 2 available specimens.

Figure 2.

MRD concordance between sample types. The red line in each graph represents equivalent disease quantification for each sample type. In a given patient, quantities of malignant clone detected were generally higher in BM than PBMCs and higher in BM or PBMCs than plasma. The blue line is the linear regression line demonstrating the mathematical correlation between sample types.

Figure 3.

PFS according to posttreatment MRD status by NGS. (A) PFS from the end of treatment (EOT) according to MRD status in BM by NGS. Fifty-seven patients had BM specimens available. Results were dichotomized as detectable vs undetectable, regardless of sensitivity. (B) PFS from the EOT according to MRD status (detectable vs undetectable) in PBMC by NGS. Twenty-nine patients had PBMC specimens available. Results were dichotomized as detectable vs undetectable, regardless of sensitivity. (C) PFS from the EOT according to MRD6 status (<10⁻⁶ vs ≥10⁻⁶) in BM. Fifty-three of 57 BM specimens were useful for this analysis; 4 were not included as MRD was undetectable, but sensitivity did not reach 10⁻⁶. (D) PFS from the EOT according to MRD status (<10⁻⁶ vs ≥10⁻⁶) in PBMC. Twenty-one of 29 PBMC specimens were useful for this analysis; 8 were not included as MRD was undetectable, but sensitivity did not reach 10⁻⁶. (E) PFS from the EOT according to MRD status (detectable vs undetectable) by NGS in BM and IGHV mutation status. (F) PFS from the EOT according to absolute MRD level. Fifty-three of 57 BM specimens were useful for this analysis; 4 were not included as MRD was undetectable, but sensitivity did not reach 10⁻⁶. Prog. free, progression free.

Figure 4.

Time to MRD reemergence according to posttreatment MRD status by NGS. (A) Time to MRD reemergence, according to MRD status (detectable vs undetectable) in BM at the EOT, assessed by NGS. Testing for MRD reemergence was performed with 4-color FLC, in peripheral blood (PB). Fifty-one of 57 patients who had BM samples available for NGS MRD testing had serial PB FLC MRD results available in PB for this analysis. (B) Time to MRD reemergence according to NGS-MRD status (detectable vs undetectable) in PB. Twenty-six of 29 patients who had PBMC samples available had serial FLC MRD results available for this analysis. (C) Time to MRD reemergence according to NGS-MRD status (<10⁻⁶ vs ≥10⁻⁶) in BM. Of 51 patients with BM samples for NGS and serial FLC MRD in PB, 47 were included in this analysis. Four were excluded as they had U-MRD in BM by NGS, but sensitivity did not reach 10⁻⁶. (D) Time to MRD reemergence by NGS-MRD status (<10⁻⁶ vs ≥10⁻⁶) in PBMC. Of 26 patients with PBMC samples for NGS and serial FLC MRD performed in PB, 18 were included in this analysis. Eight were excluded as they had U-MRD in PBMC by NGS, but sensitivity did not reach 10⁻⁶. (E) Time to progression (TTP) and time to next treatment (TTNT) after reemergence of PB MRD. Thirty patients had reemergence of MRD by FLC during follow-up and were included in this analysis.