弥漫性大B细胞淋巴瘤的细胞起源亚型分类及预后

郝婷; 高倩; 郗彦凤; 关红卫; 王彤

doi:10.16462/j.cnki.zhjbkz.2024.02.011

弥漫性大B细胞淋巴瘤的细胞起源亚型分类及预后

doi: 10.16462/j.cnki.zhjbkz.2024.02.011

郝婷¹,
高倩¹,
郗彦凤²,
关红卫³,
王彤^1, ,

1.
山西医科大学公共卫生学院卫生统计学教研室，煤炭环境致病与防治教育部重点实验室，太原 030001
2.
山西省肿瘤医院病理科，太原 030013
3.
伊萨卡学院健康科学与人类行为学院，纽约 14850

基金项目:

国家自然科学基金 82073674

国家自然科学基金 82204163

山西省基础研究计划 202203021212382

详细信息

通讯作者:
王彤，E-mail：tongwang@sxmu.edu.cn

中图分类号: R195.1;R733.4
计量
- 文章访问数: 571
- HTML全文浏览量: 125
- PDF下载量: 33
- 被引次数: 0
出版历程
- 收稿日期: 2023-08-07
- 修回日期: 2023-10-20
- 网络出版日期: 2024-03-30
- 刊出日期: 2024-02-10

Cell-of-origin subtype classification and prognosis of diffuse large B-cell lymphoma based on variable importance analysis

1.
Department of Health Statistics, School of Public Health, Shanxi Medical University, Key Laboratory of Coal Environmental Pathogenicity and Prevention, Ministry of Education, Taiyuan 030001, China
2.
Department of Pathology, Shanxi Provincial Cancer Hospital, Taiyuan 030013, China
3.
School of Health Sciences and Human Performance, Ithaca College, New York 14850, USA

Funds:

National Natural Science Foundation of China 82073674

National Natural Science Foundation of China 82204163

Shanxi Province Basic Research Programme 202203021212382

More Information

Corresponding author: WANG Tong, E-mail: tongwang@sxmu.edu.cn

摘要

摘要: 目的基因表达谱(gene expression profiling, GEP)是弥漫性大B细胞淋巴瘤(diffuse large B-cell lymphoma, DLBCL)细胞起源(cell-of-origin, COO)分类的金标准。本研究旨在建立一个基于GEP的简约模型来准确预测DLBCL的COO亚型并为其在临床上的应用提供参考。方法收集GEO数据库中6个DLBCL数据集中的基因和临床数据，将其中1个数据集作为训练集，其余5个作为验证集。构建基于惩罚回归分析的变量重要性分析策略，识别最优变量子集，并通过logistic回归分析确定最终用于COO分类的六基因模型，采用生存分析评估训练集和验证集预测的两种COO亚型与临床预后的关系。结果六基因模型在训练集预测效果较好[AUC(95% CI): 0.999(0.997~1.000)，判别斜率及其95% CI为0.944(0.920~0. 966)]，在验证集也表现出较好的效果[AUC及其95% CI波动范围从0.910(0.820~0.999)到1.000，判别斜率及其95% CI波动范围从0.506(0.350~0. 966)到0.927(0.841~0.987)]。预后模型显示，在训练集和验证集中6个基因预测的亚型均为风险预测因子(均P < 0.05)。结论六基因模型中的6个基因对DLBCL的分型和预后有重要的临床应用价值。基于变量重要性的基因排序为基因功能和靶向药物的进—步研究提供了参考依据。
- 弥漫性大B细胞淋巴瘤 /
- 惩罚回归 /
- 预后 /
- 亚型分类 /
- 变量重要性分析
Abstract: Objective Gene expression profiling (GEP) is the gold standard for cell-of-origin COO classification of diffuse large B-cell lymphoma (DLBCL). The aim of this study was to establish a GEP-based parsimony model to accurately predict the COO subtypes of DLBCL and provide a reference for its clinical application. Methods Genetic and clinical data from 6 DLBCL datasets in the GEO database were collected, and one dataset was used as the training set and the remaining five as the validation set. A variable importance analysis strategy based on penalized regression analysis was constructed to identify the optimal subset of variables, and a logistic regression analysis was performed to determine the six-gene model that was ultimately used for COO classification. Survival analysis was used to assess the relationship between the two COO subtypes predicted by the training and validation sets and clinical prognosis. Results The six-gene model predicted better in the training set [AUC(95% CI): 0.999 (0.997~1.000), discriminant slope and its 95% CI were 0.944 (0.920~0. 966)], and also showed better results in the validation set [AUC and its 95% CI fluctuated from 0.910 (0.820~0.999) to 1.000, and the discriminant slope and its 95% CI fluctuated from 0.506 (0.350~0. 966) to 0.927 (0.841~0.987)]. The prognostic modeling showed that the six genetically predicted subtypes were risk predictors in both the training and validation sets (all P < 0.05). Conclusions The six genes in the six-gene model have important clinical applications for the classification and prognosis of DLBCL. The gene ordering based on variable importance provides a reference basis for further-research on gene function and targeted drug research.
- Diffuse large B-cell lymphoma /
- Penalized regression /
- Prognosis /
- Subtype classification /
- Variable importance analysis

HTML全文

图 1 六基因模型性能评估

DS：判别斜率；ABC：活化B细胞；GCB：生发中心B细胞；DLBCL：弥漫性大B细胞淋巴瘤。

Figure 1. Six-gene model performance assessments

DS: discrimination slope; ABC: activated B-cell subtypes; GCB: germinal center B-cell; DLBCL: diffuse large B-cell lymphoma.

下载: 全尺寸图片幻灯片

图 2 训练数据集和验证数据集的6个基因的预测性能

ABC：活化B细胞；GCB：生发中心B细胞。

Figure 2. Prognostic performance of the six genes for the training and validation datasets

ABC: activated B-cell subtypes; GCB: germinal center B-cell.

下载: 全尺寸图片幻灯片

表 1 不同惩罚回归方法的变量排名表和秩整合的汇总排名列表

Table 1. Variable ranking lists of different penalized regression methods and aggregated ranking list by rank aggregation

最小绝对收缩选择算子 LASSO	自适应最小绝对收缩选择算子 aLASSO	弹性网 EN	岭回归 Ridge regression	极小极大凹惩罚 MCP	光滑链接绝对偏差惩罚 SCAD	秩整合 Rank aggregation
TNFRSF13B	TNFRSF13B	MYBL1	AFFX- HUMISGF3A/M97935_MB_at	MYBL1	MYBL1	MYBL1
MYBL1	MYBL1	TNFRSF13B	CCL5	TNFRSF13B	TNFRSF13B	TNFRSF13B
MAML3	MAML3	BATF	SCARB1	MAML3	MAML3	MAML3
CYB5R2	CYB5R2	CYB5R2	PXK	CYB5R2	CYB5R2	CYB5R2
BATF	BATF	ASB13	PXK	S1PR2	BATF	BATF
ASB13	ASB13	MAML3	C15orf40	ENTPD1	S1PR2	S1PR2
S1PR2	S1PR2	LIMD1	PDE7A	LMO2	ASB13	ASB13
LIMD1	LIMD1	S1PR2	CLEC12A	PALD1	LIMD1	LIMD1
SERPINA9	SERPINA9	SERPINA9	LACTB	ZBTB32	LMO2	SERPINA9
ENTPD1	ENTPD1	MME	LACTB	FUT8	SERPINA9	ENTPD1
LMO2	LMO2	PIM2	NLRP11	ASB13	ENTPD1	LMO2
FUT8	FUT8	FUT8	CDC42SE2	PTK2	FUT8	FUT8
ZBTB32	ZBTB32	ENTPD1	1552621_at	PIM2	ZBTB32	PALD1
PALD1	PALD1	ENTPD1	1552622_s_at	SERPINA9	PALD1	ZBTB32
BCL2L10	BCL2L10	CFLAR	TRNT1	ARHGAP24	BCL2L10	BCL2L10
PIM2	PIM2	BCL6	ARHGAP5	215164_at	STAG3	PDE7A
注：LASSO，最小绝对收缩选择算子；aLASSO，自适应最小绝对收缩选择算子；EN，弹性网；SCAD，光滑链接绝对偏差惩罚；MCP，极小极大凹惩罚。 Note：LASSO，least absolute shrinkage and selection operator；aLASSO，adaptive least absolute shrinkage and selection operator；EN，elastic net；SCAD，smoothly clipped absolute deviation penalty；MCP，minimax concave penalty.

下载: 导出CSV

表 2 训练数据集和验证数据集中的单变量和多变量Cox比例风险回归模型

Table 2. Univariate and multivariable Cox proportional hazards regression analyses in training and validation datasets

变量 Variables	单变量分析 Univariate analysis					多变量分析 Multivariable analysis
变量 Variables	β值 value	s_x	HR值value (95% CI)	χ²值 value	P值 value	β值 value	s_x	HR值value (95% CI)	χ²值 value	P值 value
训练集 GSE10846 (n=350)
六基因 Six-genes	1.040	0.179	2.830(1.993~4.018)	33.813	＜0.001	0.981	0.202	2.668(1.795~3.968)	23.514	＜0.001
国际预后指数 International prognostic index	1.091	0.192	2.978(2.044~4.340)	32.252	＜0.001	1.135	0.195	3.110(2.120~4.562)	33.688	＜0.001
处理 Treatment	-0.554	0.177	0.574(0.406~0.812)	9.851	0.002	-0.685	0.205	0.503(0.337~0.754)	11.077	0.001
性别 Gender	0.031	0.173	1.032(0.735~1.449)	0.032	0.858	―	―	―	―	―
验证集 Combat data (n=624)
六基因 Six-genes	0.457	0.157	1.579(1.161~2.147)	8.470	0.004	0.425	0.167	1.530(1.104~2.121)	6.500	0.011
国际预后指数 International prognostic index	0.966	0.165	2.627(1.900~3.632)	34.121	＜0.001	0.914	0.166	2.494(1.799~3.458)	30.090	＜0.001
性别 Gender	-0.086	0.158	0.918(0.674~1.251)	0.292	0.588	―	―	―	―	―
注：“―”表示数据无法获得。 Note："―" stands for date is not available.

下载: 导出CSV

参考文献(20)

[1]	Tilly H, Gomes da Silva M, Vitolo U, et al. Diffuse large B-cell lymphoma (DLBCL): ESMO clinical practice guidelines for diagnosis, treatment and follow-up [J]. Ann Oncol, 2015, 26(Suppl 5): v116-v125. DOI: 10.1093/annonc/mdv304.
[2]	Alizadeh AA, Eisen MB, Davis RE, et al. Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling [J]. Nature, 2000, 403(6769): 503-511. DOI: 10.1038/35000501.
[3]	Li SY, Young KH, Medeiros LJ. Diffuse large B-cell lymphoma [J]. Pathology, 2018, 50(1): 74-87. DOI: 10.1016/j.pathol.2017.09.006.
[4]	Zou H. The adaptive lasso and its oracle properties [J]. J Am Stat Assoc, 2006, 101(476): 1418-1429. DOI:10.1198/01621450600 735.
[5]	Zou H, Hastie T. Addendum: regularization and variable selection via the elastic net [J]. J Royal Stat Soc Ser B Stat Methodol, 2005, 67(5): 768. DOI: 10.1111/j.1467-9868.2005.00527.x.
[6]	Tibshirani R. Regression shrinkage and selection via the lasso [J]. J Royal Stat Soc Ser B Methodol, 1996, 58(1): 267-288. DOI: 10.1111/j.2517-6161.1996.tb02080.x.
[7]	Zhang CH. Nearly unbiased variable selection under minimax concave penalty [J]. Ann Statist, 2010, 38(2): 894-942. DOI: 10.1214/09-aos729.
[8]	Hoerl AE, Kennard RW. Ridge regression: biased estimation for nonorthogonal problems [J]. Technometrics, 1970, 12(1): 55-67. DOI: 10.1080/00401706.1970.10488634.
[9]	Pihur V, Datta S, Datta S. RankAggreg, an R package for weighted rank aggregation [J]. BMC Bioinformatics, 2009, 10: 62. DOI: 10.1186/1471-2105-10-62.
[10]	Lenz G, Wright G, Dave SS, et al. Stromal gene signatures in large-B-cell lymphomas[J]. N Engl J of Med, 2008, 359(22): 2313-2323. DOI: 10.1056/NEJMoa0802885.
[11]	Golay J, Broccoli V, Lamorte G, et al. The A-Myb transcription factor is a marker of centroblasts in vivo [J]. J Immunol, 1998, 160(6): 2786-2793.
[12]	Muppidi JR, Schmitz R, Green JA, et al. Loss of signalling via Gα13 in germinal centre B-cell-derived lymphoma [J]. Nature, 2014, 516(7530): 254-258. DOI: 10.1038/nature13765.
[13]	Flori M, Schmid CA, Sumrall ET, et al. The hematopoietic oncoprotein FOXP1 promotes tumor cell survival in diffuse large B-cell lymphoma by repressing S1PR2 signaling [J]. Blood, 2016, 127(11): 1438-1448. DOI: 10.1182/blood-2015-08-662635.
[14]	Onishi H, Yamasaki A, Kawamoto M, et al. Hypoxia but not normoxia promotes Smoothened transcription through upregulation of RBPJ and Mastermind-like 3 in pancreatic cancer [J]. Cancer Lett, 2016, 371(2): 143-150. DOI: 10.1016/j.canlet.2015.11.012.
[15]	Muppidi JR, Schmitz R, Green JA, et al. Loss of signalling via Gα13 in germinal centre B-cell-derived lymphoma [J]. Nature, 2014, 516(7530): 254-258. DOI: 10.1038/nature13765.
[16]	Devaney JM, Wang S, Funda S, et al. Identification of novel DNA-methylated genes that correlate with human prostate cancer and high-grade prostatic intraepithelial neoplasia [J]. Prostate Cancer Prostatic Dis, 2013, 16(4): 292-300. DOI: 10.1038/pcan.2013.21.
[17]	Lotem J, Sachs L. Epigenetics and the plasticity of differentiation in normal and cancer stem cells [J]. Oncogene, 2006, 25(59): 7663-7672. DOI: 10.1038/sj.onc.1209816.
[18]	Liu Q, Liu YX, Li WL, et al. Genetic, epigenetic, and molecular landscapes of multifocal and multicentric glioblastoma [J]. Acta Neuropathol, 2015, 130(4): 587-597. DOI: 10.1007/s00401-015-1470-8.
[19]	Care MA, Cocco M, Laye JP, et al. SPIB and BATF provide alternate determinants of IRF4 occupancy in diffuse large B-cell lymphoma linked to disease heterogeneity [J]. Nucleic Acids Res, 2014, 42(12): 7591-7610. DOI: 10.1093/nar/gku451.
[20]	Scott DW, Wright GW, Williams PM, et al. Determining cell-of-origin subtypes of diffuse large B-cell lymphoma using gene expression in formalin-fixed paraffin-embedded tissue [J]. Blood, 2014, 123(8): 1214-1217. DOI: 10.1182/blood-2013-11-536433.