Discipline of Genetics, Faculty of Medicine, Memorial University, St. John's, Newfoundland and Labrador, Canada.
Department of Biostatistics, Princess Margaret Hospital, University of Toronto, Toronto, Ontario, Canada.
Cancer Med. 2017 Jun;6(6):1220-1232. doi: 10.1002/cam4.1074. Epub 2017 May 23.
INDELs and CNVs are structural variations that may play roles in cancer susceptibility and patient outcomes. Our objectives were a) to computationally detect and examine the genome-wide INDEL/CNV profiles in a cohort of colorectal cancer patients, and b) to examine the associations of frequent INDELs/CNVs with relapse-free survival time. We also identified unique variants in 13 Familial Colorectal Cancer Type X (FCCX) cases. The study cohort consisted of 495 colorectal cancer patients. QuantiSNP and PennCNV algorithms were utilized to predict the INDELs/CNVs using genome-wide signal intensity data. Duplex PCR was used to validate predictions for 10 variants. Multivariable Cox regression models were used to test the associations of 106 common variants with relapse-free survival time. Score test and the multivariable Cox proportional hazards models with time-varying coefficients were applied to identify the variants with time-varying effects on the relapse-free survival time. A total of 3486 distinct INDELs/CNVs were identified in the patient cohort. The majority of these variants were rare (83%) and deletion variants (81%). The results of the computational predictions and duplex PCR results were highly concordant (93-100%). We identified four promising variants significantly associated with relapse-free survival time (P < 0.05) in the multivariable Cox proportional hazards regression models after adjustment for clinical factors. More importantly, two additional variants were identified to have time-varying effects on the risk of relapse. Finally, 58 rare variants were identified unique to the FCCX cases; none of them were detected in more than one patient. This is one of the first genome-wide analyses that identified the germline INDEL/CNV profiles in colorectal cancer patients. Our analyses identified novel variants and genes that can biologically affect the risk of relapse in colorectal cancer patients. Additionally, for the first time, we identified germline variants that can potentially be early-relapse markers in colorectal cancer.
INDELs 和 CNVs 是结构变异，可能在癌症易感性和患者预后中发挥作用。我们的目标是：a）计算并检测一组结直肠癌患者的全基因组 INDEL/CNV 谱，b）检测高频 INDEL/CNV 与无复发生存时间的关联。我们还鉴定了 13 例家族性结直肠癌 X 型（FCCX）病例中的独特变异。研究队列包括 495 例结直肠癌患者。利用 QuantiSNP 和 PennCNV 算法，根据全基因组信号强度数据预测 INDEL/CNVs。采用双 PCR 验证 10 个变异的预测。采用多变量 Cox 回归模型，检验 106 个常见变异与无复发生存时间的关系。应用 Score 检验和时变系数的多变量 Cox 比例风险模型，鉴定对无复发生存时间具有时变影响的变异。在患者队列中，共鉴定出 3486 个独特的 INDEL/CNVs。这些变异多数为稀有（83%）和缺失变异（81%）。计算预测和双 PCR 结果高度一致（93-100%）。在多变量 Cox 比例风险回归模型中，经过临床因素调整后，我们鉴定出与无复发生存时间显著相关的四个有前景的变异（P<0.05）。更重要的是，鉴定出另外两个对复发风险具有时变影响的变异。最后，鉴定出 FCCX 病例特有的 58 个稀有变异，它们均未在 1 个以上患者中检测到。这是首次在全基因组范围内分析结直肠癌患者的胚系 INDEL/CNV 谱。我们的分析鉴定出了可从生物学上影响结直肠癌患者复发风险的新型变异和基因。此外，我们首次鉴定出了可能成为结直肠癌早期复发标志物的胚系变异。
