Laurenti, Luca

Scoring systems designed to improve the accuracy of prognostication in chronic lymphocytic leukemia (CLL) usually rely on discretized/dichotomic values of the clinical and biological variables. Here we analyzed the immunophenotypic and (immuno)genetic profiles of 2,243 CLL patients with Rai stage 0-I-II, by applying unsupervised machine learning methods, elaborating prognostic factors as continuous variables, to identify novel relationships and interactions likely missed in conventional hierarchical models. The study included an internal training cohort (n=863) and two external validation cohorts provided from Cardiff University/Brighton Medical School (n=455) and the Mayo Clinic (Rochester, MN; n=925). Primary endpoint was Time to First Treatment (TTFT). Laboratory-based markers evaluated in the training cohort were: surface antigen expression by flow cytometry (as % of positive cells) of CD20, FMC7, CD49d, CD49c, CD38, CD23, CD43, CD22, ZAP-70; chromosomal aberrations (% of nuclei with abnormal signal) del13q, tris12, del11q and del17p,; mutational status of TP53 by NGS as % VAF; IGHV mutational status as % mutation. 420/863 cases received treatment (median 46 months). Univariate Cox regression identified FISH del11q, del17p and tris12, TP53 and IGHV mutations, CD38 and CD49d expression, as features associated with TTFT (p-value <0.001). An unsupervised k-means algorithm partitioned cases into 6 clusters (C1-C6), as selected by the elbow method, and centroid analysis evaluated which feature most contributed to each cluster (Panel A): C1 (n=275) was heavily IGHV-mutated (mutation range 4.5-22.0%) with low representation of all other features; C2 (n=208) was mostly IGHV-unmutated/intermediate (up to 4.5%) in the absence of other features; C3 (n=169) were CD49d-expressing cases (>50% expression), with low representation of other features, except CD38; C4 (n=127) were cases with trisomy 12 (>28%), also expressing CD49d and CD38; C5 (n=48) were cases with highly clonal del11q (44-98%) and IGHV-unmutated, mutually exclusive with TP53 disruption; C6 (n=34) contained TP53-disrupted cases with high mutation burden (VAF 36-97%) and/or del17p (range 40-96%), with low importance of all the other features. Of note, alternative use of either TP53 mutation or del17p as defining features of cluster 6 did not result in changes of the classification accuracy. Clinically, clusters 3-4-5-6 showed median TTFT of 26, 22, 9, 5 months respectively whereas median TTFT was not reached for clusters 1-2. Hierarchical agglomerative clustering aggregated the 6 clusters in 3 major risk tiers: high (C5-6), intermediate (C2-3-4) and low (C1), with median TTFT of 7, 45 and not reached, respectively (Panel B). Multivariate Cox analysis (MVA) with Rai staging demonstrated that the 3-tier risk score contributed significantly and independently to TTFT estimate (p<0.0001; Harrel's c-index=0.71). To classify novel patients, we designed a hierarchical algorithm, based on each clusters' estimated risk and according to calculated cut-off, i.e. C6, del17p/ TP53>36%; C5, del11q>43%; C4, tris12>28%; C3, CD49d>50%; C2, IGHV≤4.5%; C1, the remaining patients with IGHV>4.5%. This approach was highly robust by an internal training-validation bootstrap (cluster classification accuracy = 0.96±0.02). External samples from two independent cohorts were classified into the six clusters using the cut-offs defined by the hierarchical classifier, and then aggregated into the 3-risk tiers. The classifier performed robustly in both cohorts, again adding independent prognostic information to Rai stage in MVA, with p<0.0001 and c-index of 0.76 for Cardiff/Brighton cohort and p<0.0001 and c-index of 0.74 for Mayo cohort. In conclusion, our machine-learning-driven, laboratory-based classification identifies clusters of patients who at higher risk of requiring early treatment, independent of clinical staging, and may help to better identify patients who may benefit from early treatment or more frequent disease surveillance. For some features, TP53 or IGHV mutational status, our unsupervised approach selected cut-offs not recapitulating the canonical ones, envisioning different biological activities in the TTFT setting. For example, TP53 mutations were clinically relevant only if present in the vast majority of the clone , suggesting the involvement of specific activities of the TP53 mutant in this context.

Abstract Introduction: Observation is the standard of care for asymptomatic early stage chronic lymphocytic leukemia (CLL) however these cases follow a heterogenous course. Recent studies show novel biomarkers can delineate indolent from aggressive early stage disease and current clinical trials are exploring the role of early intervention in high risk cases. Although several scoring systems have been established in CLL, most are designed for overall survival, do not circumscribe early stage disease, and require cumbersome calculations relying on extensive laboratory and clinical information. Aim: We propose a novel laboratory-based prognostic calculator to risk stratify time to first treatment (TTFT) in early stage CLL and guide candidate selection for early intervention. Methods: We included 1574 cases of early stage CLL in an international cohort from Italy, the United Kingdom and the United States using a training-validation model. Patient information was obtained from participating centers in accordance with the Declaration of Helsinki. The training cohort included 478 Rai 0 cases from a multicenter Italian cohort, all referred to a single center (Clinical and Experimental Onco-Hematology Unit of the Centro Riferimento Oncologico in Aviano, IT) for immunocytogenetic lab analyses. Considering TTFT as an endpoint, we evaluated 8 variables (age>65, WBC>32K, 17p-, 11q-, +12, IGHV status, CD49d+, gender) with univariate and multivariate Cox regression internally validated using bootstrapping procedures. FISH thresholds were 5% for 11q-, and +12 and 10% for 17p-. Cases were categorized according to the hierarchical model proposed by Dohner. IGHV status was considered unmutated at ≥98%. CD49d+ was set at >30%. WBC cutoff of >32K was established by maximally selected log rank analysis. Variables were weighted based on the proportion of their normalized hazard ratios rounded to the nearest whole integer. We used recursive partitioning for risk-category determination and Kaplan-Meier analysis to generate survival curves. We compared the concordance index (C-index) of our model with the CLL international prognostic index (CLL-IPI) for 381/478 cases in the training cohort with available beta-2-microglobulin data and for all validation cohorts. We used 3 independent single-center cohorts for external validation. Results: The training cohort had 478 cases of Rai 0 CLL with a median (95% CI) TTFT of 124 months (m) (104-183m). Five prognostic variables emerged with respect to TTFT, and each assigned a point value of 1 or 2 according to their respective normalized HR values as follows: 17p-, and UM IGHV (2 pts); 11q-, +12, and WBC>32K (1 pt). We identified three risk groups, based on point cut-offs of 0, 1-2, and 3-5 established by recursive partitioning analysis with a median (95% CI) TTFT of 216m (216-216m), 104m (93-140m) and 58m (44-68m) (p<0.0001, C-index 0.75) for the low, intermediate, and high-risk groups, respectively (figure 1). A comparison with the CLL-IPI was possible in 381 cases with available beta-2-microglobulin data. In this subset, the C-index was 0.75 compared to 0.68 when patient risk groups were split according to the CLL-IPI. The scoring system was then validated in 3 independent cohorts of early stage CLL: i) Gemelli Hospital in Rome, IT provided 144 Rai 0 cases. Median (95% CI) TTFT was 86m (80-94m, 95% CI). Median (95% CI) TTFT for the low, intermediate and high-risk groups was 239m (239-239m), 98m (92-132m) and 85m (60-109m) respectively (p=0.002 between low and intermediate groups, p=0.09 between intermediate and high groups; C-index 0.64 v 0.60 for CLL-IPI). ii) Cardiff University Hospital in Wales, UK provided 395 Binet A cases. Median (95% CI) TTFT was 74 m (67-81m) overall and NR, 111m (97-146m) and 70m (29-114m) for the low, intermediate and high-risk groups respectively (p<0.001 between low and intermediate groups, p=0.009 between intermediate and high groups; C-index 0.63 v 0.63 for CLL-IPI). iii) Mayo Clinic in Rochester, MN provided 557 Rai 0 cases. Median (95% CI) TTFT was 127m (96m-NR) overall and NR, 76m (64m-NR) and 36m (31-59m) for the low, intermediate and high-risk groups respectively (p<0.0001; C-index 0.72 v 0.68 for CLL-IPI). Conclusion: We present a novel laboratory-based scoring system for Rai 0/Binet A CLL to aid case selection in risk-adapted treatment for early disease. Further comparison to existing indices is needed to verify its utility in the clinical setting. Disclosures Zaja: Novartis: Honoraria, Research Funding; Takeda: Honoraria; Abbvie: Honoraria; Celgene: Honoraria, Research Funding; Amgen: Honoraria; Janssen: Honoraria; Sandoz: Honoraria. Fegan:Roche: Honoraria; Napp: Honoraria; Janssen: Honoraria; Gilead Sciences, Inc.: Honoraria; Abbvie: Honoraria. Pepper:Cardiff University: Patents & Royalties: Telomere measurement patents. Parikh:AstraZeneca: Honoraria, Research Funding; Janssen: Research Funding; MorphoSys: Research Funding; Abbvie: Honoraria, Research Funding; Gilead: Honoraria; Pharmacyclics: Honoraria, Research Funding. Kay:Janssen: Membership on an entity's Board of Directors or advisory committees; Acerta: Research Funding; Celgene: Membership on an entity's Board of Directors or advisory committees; Pharmacyclics: Membership on an entity's Board of Directors or advisory committees, Research Funding; Infinity Pharm: Membership on an entity's Board of Directors or advisory committees; Cytomx Therapeutics: Membership on an entity's Board of Directors or advisory committees; Tolero Pharmaceuticals: Membership on an entity's Board of Directors or advisory committees, Research Funding; Morpho-sys: Membership on an entity's Board of Directors or advisory committees; Gilead: Membership on an entity's Board of Directors or advisory committees; Agios Pharm: Membership on an entity's Board of Directors or advisory committees.