New Risk Index Sharpens Outcome Prediction After Fixed-Duration CLL Therapy

A refined risk-prediction model outperformed the CLL International Prognostic Index (CLL-IPI) and other single-factor tools in forecasting progression-free survival (PFS) and overall survival (OS) among patients with chronic lymphocytic leukemia (CLL) treated with fixed-duration regimens, according to results published in the Journal of Clinical Oncology.¹

The model, called CIRI2-CLL, is an updated version of the continuous individualized risk index (CIRI), a Bayesian Cox proportional hazards framework that integrates baseline clinical and cytogenetic risk factors with longitudinal measurable residual disease (MRD) data collected during and after treatment. The refined model adds MRD status assessed 12 months after end of treatment (EoT) to earlier CIRI inputs—CLL-IPI risk category, treatment regimen, and MRD status at interim and EoT time points—to generate an individualized, updated survival probability at any point along a patient's treatment course.

Investigators developed CIRI2-CLL using pooled data from 913 patients across the CLL8 (NCT00281918), CLL10 (NCT00769522), CLL11 (NCT01010061), and MURANO (NCT02005471) trials, incorporating both chemoimmunotherapy- and venetoclax (Venclexta)-based regimens. The model was then validated in an independent cohort from the CLL14 trial (NCT02242942) including 198 patients treated with chlorambucil plus obinutuzumab (Gazyva) and 183 patients treated with venetoclax plus obinutuzumab, with a sensitivity analysis performed in the CLL13 (NCT02950051) trial cohort.

Key Findings

In the validation set, CIRI2-CLL showed a C-statistic between 0.82 and 0.98 across all assessed time points for PFS prediction, representing a 17% to 37% improvement over CLL-IPI and a 14% to 37% improvement from 2 to 5 years compared with individual prognostic indices, including MRD status alone. Calibration was strong, with less than a 2% difference between observed and predicted event probabilities from 2 to 4 years among venetoclax-obinutuzumab–treated patients.

When patients were stratified into low-, intermediate-, and high-risk CIRI2-CLL groups, 3-year PFS rates from the most recent landmark were 100.0%, 70.2%, and 10.8%, respectively, among venetoclax-obinutuzumab–treated patients. Using an alternative tertile-based stratification, 3-year PFS rates were 89.4%, 68.8%, and 37.1% in the venetoclax-obinutuzumab arm and 65.1%, 23.0%, and 5.9% in the chlorambucil-obinutuzumab arm.

By comparison, CLL-IPI alone showed a significant PFS difference only between its lowest- and highest-risk categories (3-year PFS, 100.0% vs 66.7%; HR, 8.97; 95% CI, 2.01-40.1; log-rank P <.001) among venetoclax-obinutuzumab–treated patients; differences between intermediate, high, and very-high-risk CLL-IPI groups were not statistically significant. Notably, CIRI2-CLL reclassified a meaningful proportion of patients relative to CLL-IPI: 19.6% of venetoclax-obinutuzumab–treated patients and 15.6% of chlorambucil-obinutuzumab–treated patients with intermediate CLL-IPI risk shifted to high CIRI2-CLL risk, while 18.8% and 17.5% of patients, respectively, with high CLL-IPI risk shifted to low CIRI2-CLL risk.

For OS, CIRI2-CLL again outperformed individual indices, with a C-statistic of 0.77 to 0.98 across time points and calibration within 5% from 2 to 5 years among venetoclax-obinutuzumab–treated patients. Three-year OS rates from the most recent landmark were 92.7%, 58.1%, and 0% in the low-, intermediate-, and high-risk groups, respectively; however, investigators cautioned that the high-risk OS group included only 2 patients at the most recent landmark, warranting careful interpretation.

Limitations and Implications

The study authors noted several limitations, including immature OS data in the validation cohort despite long follow-up, and a model developed and validated exclusively with venetoclax plus CD20 antibody regimens rather than venetoclax plus BTK inhibitor combinations, which authors stated will be explored as longer-term data become available. Patients with TP53 alterations were also underrepresented in the CLL13 validation cohort.

The investigators, a collaboration between the German CLL Study Group and Stanford University, have made CIRI2-CLL publicly accessible as a web-based tool intended to support clinicians and investigators in individualized posttreatment risk assessment and in designing future risk-adapted, MRD-informed CLL trials.²

“Our study not only validates CIRI-CLL in the context of limited-duration targeted treatment of CLL, suggesting the adaptable nature of the CIRI algorithm, but also offers a dynamic clinically applicable prognostic tool with robust risk stratification to investigators and clinicians,” study authors Al-Sawaf et al concluded.

REFERENCES

1. Al-Sawaf O, Zhang C, Esfahani MS, et al. Refined Continuous Risk Index for Accurately Predicting Outcomes of Patients With Chronic Lymphocytic Leukemia After Limited-Duration Therapy. J Clin Oncol. Published online August 27, 2026. doi:10.1200/jco-26-00161

2. CIRI2-CLL. Stanford University. Accessed September 3, 2026. https://ciri2.stanford.edu/