Gene expression signature predicts induction treatment response and clinical outcome in adult Colombian patients with B-acute lymphoblastic leukemia. Homo sapiens

Background. B-Acute lymphoblastic leukemia (B-ALL) represents a hematologic malignancy with poor clinical outcome and low survival rates in adult patients. Only 61% of Colombian adult patients with ALL achieve complete remission and a median overall survival of less than 11.3 months and event-free survival of 7.34 months. Identification of prognostic factors is crucial for proper treatment strategies and optimal results of therapy. The goal of our study was to determine if gene expression signatures correlate with response to therapy and to evaluate the utility of these as prognostic tools for survival to better predict patient risk prior to therapy. Methods and findings. This study included 43 adult patients newly diagnosed with B-cell precursor or common B-ALL. Through microarray gene expression profiles, we identified 442 genes differentially expressed between 27 leukemia patients who responded or not to induction chemotherapeutic treatment. Hierarchical analysis with the 99 most differentially expressed genes between the two groups revealed 3 sets of patients that differed in their clinical characteristics, giving these genes high prognostic clinical outcome impact. We validated the expression of 7 genes by RT-PCR in 43 patients, and in addition to finding a correlation with gene expression profiles, we established correlations with good and poor prognosis from the time of diagnosis. Conclusions. Our study suggests that the response to induction treatment and clinical outcome of patients can be predicted from the onset of the disease and that gene expression profiles can be used to stratify patient risk adequately and accurately. The present study represents the first that shows that gene expression profiling could become a clinically relevant tool for stratification in the early course of disease of Colombian adults B-ALL. Overall design: Stem cells from bone marrow samples from B-acute lymphoblastic leukemia (B-ALL) patients responding or not to the inductive therapy were isolated by sorting, and the gene expression was compared between the two groups. This submission includes samples from 27 B-ALL patients and 3 healthy individuals.

背景。B-急性淋巴细胞白血病（B-Acute lymphoblastic leukemia, B-ALL）是一类成人患者临床预后不佳、生存率偏低的血液系统恶性肿瘤。哥伦比亚地区仅61%的成人ALL患者可获得完全缓解（complete remission），中位总生存期（median overall survival）不足11.3个月，无事件生存期（event-free survival）仅7.34个月。明确预后相关因素对于制定合理治疗策略、获取最优治疗效果至关重要。本研究旨在探究基因表达特征与治疗应答的相关性，并评估其作为生存预后工具的应用价值，以便在治疗前更精准地预测患者风险。

方法与结果。本研究共纳入43例初诊B细胞前体型或普通型B-ALL成人患者。通过基因芯片（microarray）表达谱分析，我们在27例对诱导化疗（induction chemotherapeutic treatment）存在应答与无应答的白血病患者中，筛选出442个差异表达基因（differentially expressed genes）。基于两组间差异最显著的99个基因开展层级聚类分析（Hierarchical analysis），可将患者分为3个具有不同临床特征的亚组，提示这些基因具备较高的临床预后价值。我们通过逆转录聚合酶链反应（Reverse Transcription Polymerase Chain Reaction, RT-PCR）验证了43例患者中7个基因的表达水平；除证实其与基因表达谱存在相关性外，我们还建立了自诊断起即可区分良好与不良预后的关联模型。

结论。本研究表明，可在疾病初发阶段预测患者对诱导治疗的应答情况及临床预后，且基因表达谱可用于精准、可靠地对患者风险进行分层（stratify patient risk）。本研究首次证实，基因表达谱分析可作为哥伦比亚成人B-ALL患者疾病早期阶段风险分层的临床实用工具。

整体实验设计：从对诱导治疗存在应答与无应答的B-ALL患者骨髓样本中分离干细胞，随后对两组样本的基因表达水平进行比较。本提交数据集包含27例B-ALL患者与3例健康个体的样本。