A 17-gene expression signature to distinguish patients who are likely to achieve long-term remissions following front-line FCR chemoimmunotherapy from those who might benefit from alternative regimens

Chemoimmunotherapy with fludarabine, cyclophosphamide, and rituximab (FCR) is the gold-standard regimen for physically fit patients with chronic lymphocytic leukemia (CLL). The regimen is brief and inexpensive and patients younger than 65 years, without comorbidities and with mutations in IGHV, usually achieve long-term durable remissions. On the contrary, patients older than 65 years, or with comorbidities, often cannot tolerate a full treatment course.

In this paper, Carmen D Herling, Department I of Internal Medicine, Center for Integrated Oncology, Aachen-Bonn-Cologne-Duesseldorf, Cologne, DE, and colleagues hypothesized that the duration of response to FCR chemoimmunotherapy depends on differences in the expression of protein-coding genes. Therefore, they developed and validated a 17-gene expression signature to identify patients that might achieve durable remissions following front-line FCR chemoimmunotherapy.

Study design and patients¹

• A retrospective cohort study was performed in two cohorts (a training cohort and a validation cohort) of treatment-naive patients:
• The training cohort consisted of peripheral blood samples collected at the MD Anderson Cancer Centre (MDACC cohort) from treatment-naive patients with CLL. Patients included (N= 101) were aged ≥ 18 years; received at least three cycles of standard FCR (n= 61) or rituximab intensified FCR (n= 40) therapy; had samples with adequate amount and high-quality RNA
  • On blood or bone marrow samples, the authors assessed the IGHV somatic mutation status, ZAP70 expression, and abnormalities commonly associated with CLL
  • Among the 101 patients, 79 (78%) completed six cycles of chemoimmunotherapy and 98 (97%) responded to treatment. A complete remission was observed in 84 (83%) patients, a partial remission in 14 (14%), no response in two (2%), and one (1%) had an unknown status
  • The median follow-up was 146.5 months (interquartile range [IQR] 137.7–162.3) after treatment
  • At the final follow-up, of the 101 patients:
    • Forty-nine (49%) had progressed, with a median time to progression of 92.4 months (IQR, 44.1–not reached [NR])
    • Forty-two (42%) had died
    • Sixty-three (62%) had either progressed or died
  • The median progression-free survival (PFS) was 66.6 months (40.5–NR)
• The validation cohort (N= 109), in comparison with the MDACC cohort, had more patients with advanced stage disease, and consisted of treatment-naive patients with unmutated IGHV and available gene expression profiling data who were enrolled on the CLL8 trial
  • The median follow-up was 73.4 months (IQR 68.2–82.9)
  • The median overall survival (OS) was 80.8 months (58.3–NR)
  • The median time to progression was 46.5 months (23.1–83)
  • The median PFS was 42.4 months (21.4–72)
  • At the final follow-up, of the 109 patients:
    • Seventy-seven (71%) had progressed
    • Forty-six (42%) had died
    • Eighty-three (76%) had either progressed or died

Results¹

After the gene expression data analysis for the MDACC cohort, the authors identified 1,136 probes associated with time to progression. Using these probes, patients with similar gene expression patterns were divided into favorable, intermediate, and unfavorable prognosis subsets. The intermediate prognosis and unfavorable prognosis subset had a shorter time to progression compared with patients in the favorable subset.

Genes highly expressed in unfavorable cases (n= 424) were associated with metabolic pathways, including oxidative phosphorylation and ribonucleoside metabolism. Genes highly expressed in favorable or intermediate cases (n= 401) encoded products involved in ATP binding, purine ribonucleoside triphosphate binding, nucleic acid binding, and DNA-template transcription.

The authors developed a prognostic model with 17 genes to distinguish IGHV-unmutated patients that had an intermediate outcome from those with an unfavorable outcome after front-line FCR therapy. The development process included:

• Selection of reliable samples, 24 in the intermediate subset and 30 in the unfavorable subset
• Identification of 726 differentially expressed genes, of which 177 (24%) were present in the subsets associated with time to progression
• Exclusion of duplicate probes, pseudogenes, and genes of uncertain function, leading to the identification of 139 genes, and of these, through a regression analysis (Lasso method), 17 genes were selected, separating the two prognostic groups

These 17 genes were validated in 109 patients with an IGHV-unmutated status from the CLL8 cohort. In this cohort, patients classified as high risk (unfavorable prognosis; median time to progression of 39 months [IQR, 22–69]) had a hazard ratio of 1.90 (95% CI, 1.18–3.06; p = 0.008) compared with low-risk (intermediate prognosis; median time to progression of 59 months [IQR 28–84]) patients. Of the 17 genes, 13 came from the cluster of genes highly expressed in unfavorable cases with shorter time to progression, and increased expression corresponds to increased risk of progression. Three of the 17 genes came from the cluster of genes highly expressed in favorable or intermediate cases with longer time to progression and increased expression corresponds to decreased risk.

Conclusions

• In order to distinguish patients with IGHV-unmutated CLL who might achieve durable remissions after front-line FCR chemoimmunotherapy from those who might benefit from alternative regimens, the authors developed and validated a 17-gene signature
• Genes highly expressed in unfavorable cases are involved in metabolism, oxidative phosphorylation, and in purine metabolism. Differences in the expression of these genes are related to therapy response
• As expected, the vast majority of patients with mutated IGHV had a favorable prognosis, but about a third of patients in the intermediate or poor prognosis groups also had mutated IGHV. Thus, mutated IGHV did not ensure a good prognosis
• The 17-gene signature needs to be tested in a study that compares FCR treatment with alternative therapies