Ultrasensitive tumour‐agnostic non‐invasive detection of colorectal cancer recurrence using ctDNA methylation

结直肠癌 医学 甲基化 肿瘤科 DNA甲基化 癌症研究 内科学 癌症 生物 遗传学 基因 基因表达
作者
Yu Xiao,Xiaodong Wang,Hong Weng,Zhao Ding,Kaiyu Qian,Wan Jin,Sen Lu,Lingao Ju,Zhong He,Gang Wang,Xiaoyu Xie,Dongmei Li,Fan Zhang,Kai Wu,Sheng Li,Huanhuan Guo,Guofeng Qian,Wei Jiang,Yunji Leng,Jingjing Zhao,Xinyue Cao,Min‐Sheng Peng,Congqing Jiang,Li Li,Yi Zhang,Xinghuan Wang
出处
期刊:Clinical and translational medicine [Springer Science+Business Media]
卷期号:12 (9) 被引量:1
标识
DOI:10.1002/ctm2.1015
摘要

Dear Editor, Colorectal cancer (CRC) is the third most common cancer (10.0%) and the second leading cause of cancer-related deaths (9.4%) worldwide.1 Here we show cancer-derived DNA methylation (DNAm) enables sensitive and specific non-invasive diagnosis and monitoring of CRC. Cell-free DNA (cfDNA) containing tumour-derived biomarkers provide a promising avenue for non-invasive detection of minimal residual disease (MRD) after curative-intent surgery.2 Tumour-derived DNA, or circulating tumour DNA (ctDNA), could be detected by sequencing cfDNA from post-surgery or adjuvant therapy blood.3 We performed a prospective, observational, and multi-centre cohort study (Cancer HALLmark Epigenetics aNd GEnetics in CRC, Challenge-CRC) with 280 CRC patients (Stages I–IV) to assess the ability of a tumour-agnostic ctDNA assay to identify patients with MRD that would ultimately recur. Pre- and post-operative cfDNA were obtained from the patients. Somatic mutation and DNAm were sequenced and analysed. Statistics is performed to compare the performance to detect MRD from somatic mutation and DNAm (Figure 1). In this interim analysis, patients were divided into a surgery-naïve group (n = 64) and a post-curative-intent-surgery group (n = 40) (Tables S1 and S2). For the surgery-naïve group, 64 surgery-naïve, initially diagnosed, untreated CRC patients were enrolled. Enrolled patients include 37 (57.8%) colon cancer and 27 (42.2%) rectal cancer of clinical Stages I (3.1%, 2/64), II (20.3%, 13/64), III (34.4%, 22/64) and IV (42.2%, 27/64). Overall, 82.8% of patients (53/64) are lymph-node metastasis-free, and 68.8% (44/64) patients are distal metastasis-free. For the post-surgery group, enrolled patients include 10 (25.0%) colon cancer and 30 (75.0%) rectal cancer of clinical Stages II (10.0%, 4/40), III (27.5%, 11/40) and IV (62.5%, 25/40). Overall, 37.5% of patients (15/40) are lymph-node metastasis-free, whereas 62.5% (25/40) patients are with N1/N2. Overall, 22.5% (9/40) patients received neoadjuvant therapy and 67.5% (27/40) received adjuvant therapy. Overall, 22.5% (9/40) patients had surgery alone with no neoadjuvant or adjuvant therapy. Patients were followed up with a median of 239 days post-operatively. Overall, 32.5% (13/40) patients recurred with a median time to recurrence from surgery of 214 days (range, 6–384). Blood was drawn a median of 14 days (range, 6–307) post-operatively. In pre-operative blood cfDNA, with a limit-of-detection of 0.5% variant allelic frequency (VAF), 50.0% (32/64) patients were somatic mutation positive (Tables S3–S5). As expected, detected putative tumour-specific mutations include pathogenic mutations in TP53, KRAS, APC and SMAD4 (Figure 2A). Structural variants derived gene fusions of FGFR1 or NTRK1 were detected in two cases. These results closely resemble the reported mutation landscape in CRC.4 Mean frequency of CRC-specific DNAm haplotypes, or the mean VAF of tumour-specific somatic mutation (including structural variation) (Figure 2B), were independently used to derive tumour fraction (TF) in cfDNA. In samples with positive somatic mutation detected, methylation-derived TF linearly correlates with mutation-derived TF (adj. R2 = 0.9074, p < 2.2e − 16, Figure 3A). Furthermore, many samples without any detected somatic mutation were positive for methylation-derived TF (Figure 3A,B). Combining tumour-derived DNAm signal with tumour-associated, immune-related DNAm signal, age-related DNAm signal and cfDNA fragment size information further improved the performance of cfDNA methylation. Here, by leveraging DNAm correlating with multiple, orthogonal biological features, we constructed the MAFIT score to report an overall tumour-associated DNAm level in the sample (see the Supporting Information section). Overall, when considering both tumour-derived and tumour-associated, immune-related signals, DNAm was positive in 78.1% (50/64) of samples, whereas cfDNA mutation was positive in 50.0% (32/64) of samples. All samples positive with somatic mutation were also positive for DNAm, and DNAm additionally detected ctDNA signal in 56.3% (18/32) somatic-mutation-free samples. We conclude that compared to cfDNA mutation, DNAm is more sensitive for detecting presence of tumour (detection odds ratio 3.53 (1.56–8.36), p = 0.001595, Fisher's exact test). In post-operative blood cfDNA from 40 patients received curative-intent surgery, somatic mutation was positive in 47.5% (19/40) patients (Tables S4–S6), whereas DNAm was positive in 67.5% (27/40) patients (Figure 4A,C). In analysing progression-free survival (Figure 4B,D), 100% (13/13) of patients who are negative for DNAm, regardless of somatic mutation, were recurrence-free during the whole observation period (12 months). Tumour recurrence was found in 27.3% (3/11) of patients who are methylation-positive but somatic-mutation-negative with a median RFS of 384 days (95% CI: 356-Inf), or 62.5% (10/16) of patients who are double-positive for somatic mutation as well as DNAm, with a median RFS of 242 days (95% CI: 77-Inf.). Compared with somatic mutation predicted ctDNA positivity, methylation-derived ctDNA positivity has similar PPV (48.1% vs. 52.6%, p = 1, Fisher's exact test) but higher NPV (100% vs. 85.7%, p = 0.2701, Fisher's exact test). Overall, the cfDNA methylation assay shows a sensitivity of 100% and a specificity of 48.1%, whereas the cfDNA somatic mutation assay shows a sensitivity of 76.9% and a specificity of 66.7%. We conclude that methylation-derived ctDNA positivity was highly predictive for tumour recurrence and outperforms mutation-derived ctDNA positivity (p = 0.024, log-rank test) in our period of observation. Patients were grouped into risk strata according to their DNAm level (MAFIT score) in post-operative blood cfDNA as low- (<0.5), mid- (between 0.5 and 0.8) and high- (>0.8) risk. Overall, 100% low-risk group patients were progression-free within 400 days, whereas the medium survival days was 384 days for mid-risk group (28.6% recurred, 4/14) and 256 days for high-risk group (69.2% recurred, 9/13) (Figure 4F,G). We conclude that cfDNA methylation is capable of stratifying patients into risk groups with significant different risk of recurrence (Figure 4G and p = 0.0062, log-rank test). DNAm events predicting CRC have been reported in many studies and applied in several clinical practices for diagnosis and prognosis, for example, detecting methylated SEPT9 DNA in blood, or SDC2/BMP3/NDRG4 DNAm in faecal samples.5, 6 Diagnosis assays employing a genomic-epigenomic combined approach to detecting both tumour-specific somatic mutation (such as KRAS) and immune-cell-specific methylation (such as NDRG4)7 could enhance overall specificity of the assay and helped to discriminate between benign neoplasm and true malignancy, and monitor MRD.8 A major problem of all state-of-the-art MRD detection assay is sensitivity, that is, recurrence in the ctDNA-negative population. Here we found that cfDNA methylation-derived TF is linearly correlated with somatic mutation-derived TF. Furthermore, ctDNA positivity derived by methylation outperformed somatic mutation in terms of predicting recurrence in our period of observation. Superior performance of ctDNA by methylation could be due to two reasons: First, ctDNA detection by methylation is only cell-type specific and independent from the individualized mutation profile of each tumour, helping to normalize tumour-specific signal and to ‘rescue’ the detective power for tumours without canonical hotspot or targetable mutations. For example, certain tumour with a novel fusion driver or a DNAm driver (GIST with SDHC germline mutation) could be somatic mutation-free.9 Second, cfDNA methylation carried non-tumour-derived, tumour-associated signals such as remote immune cell activation to enhance detection sensitivity.10 The power of our conclusion is limited by a small number of patients with relatively short follow-up in this interim analysis. Additionally, such short follow-up has made the post-surgery cohort does not fully resemble pre-surgery cohort, which might impact the translation of prediction accuracy of DNAm for tumour load in cfDNA from the pre-surgery data towards post-surgery data. Long-term follow-up results from the full cohort would be reported in the future, and we expect that these limitations could be fully answered. As the study is carried out during the COVID-19 pandemic, blood collection times of the patients were not strictly restricted. However we did not notice any correlation between blood sampling time and tumour positivity detection by either somatic mutation or DNAm in our data. Furthermore, as we directly compared paired somatic mutation and DNAm sequencing results from the same blood draw, we did not consider the sampling time affects the power of this study. In conclusion, we found that DNAm on colorectal-cancer-specific DMR based on blood cfDNA sequencing enables an accurate detection of potential relapse of CRC. In the future, it is possible that cfDNAm test of pre-surgery blood could help one to distinguish between benign disease and truly malignant CRC, to reduce the unnecessary colonoscopy and even unnecessary surgery. Furthermore, post-surgery blood DNAm signature detects residual disease with high sensitivity to help in suggesting surveillance protocol and indicating time interval of follow-up or even guiding therapeutic interventions. The excellent technical assistance of Shanshan Zhang, Hongwei Peng, Wan Xiang, Ziyu Jiang, Wen Chen and Zongning Zhou at Hubei Biobank is gratefully acknowledged. Part of the analysis was performed on the High Performance Computing Platform of the Center for Life Science (Peking University). This study is partially supported by grants from the Improvement Project for Theranostic Ability on Difficulty Miscellaneous Disease (Tumour) from National Health Commission of China (ZLYNXM202006 and ZLYNXM202012), Medical Science Advancement Program (Clinical Medicine) of Wuhan University (TFLC2018002), Research Fund of Zhongnan Hospital of Wuhan University (SWYBK00-03, KY0100000109), Department of Science and Technology of Hubei Province (2018ZYYD023, 2022EJD001 and YYXKNL2022001) and Non-profit Central Research Institute Fund of Chinese Academy of Medical Sciences (2020-PT320-004). The funders played no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript. The authors declare that they have no conflict of interests. All the authors consent for publication. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.

科研通智能强力驱动
Strongly Powered by AbleSci AI
科研通是完全免费的文献互助平台,具备全网最快的应助速度,最高的求助完成率。 对每一个文献求助,科研通都将尽心尽力,给求助人一个满意的交代。
实时播报
CHEN完成签到 ,获得积分10
刚刚
驰骋完成签到 ,获得积分10
1秒前
maple完成签到 ,获得积分10
1秒前
阿冰完成签到 ,获得积分10
2秒前
酷炫映阳完成签到 ,获得积分10
3秒前
怡然的千儿完成签到,获得积分10
4秒前
大明完成签到,获得积分10
5秒前
harden9159完成签到,获得积分10
6秒前
10秒前
仁仁仁完成签到,获得积分10
11秒前
灯座完成签到,获得积分10
13秒前
u雨发布了新的文献求助10
14秒前
Voiceless的应助被Bin_Liu采纳,获得10
17秒前
不吃压力就是干完成签到 ,获得积分10
17秒前
Lingeek完成签到,获得积分10
22秒前
霸气芫完成签到 ,获得积分10
26秒前
26秒前
betty2009完成签到,获得积分10
27秒前
饱满千兰完成签到,获得积分10
28秒前
科研通AI6.3的应助被u雨采纳,获得10
32秒前
王kk完成签到 ,获得积分10
33秒前
哈哈哈完成签到 ,获得积分10
35秒前
邓代容完成签到 ,获得积分0
40秒前
奋斗诗云完成签到 ,获得积分10
42秒前
irisyan完成签到 ,获得积分10
44秒前
白云完成签到 ,获得积分10
45秒前
难过的研究牲完成签到 ,获得积分10
46秒前
大耳朵图图完成签到 ,获得积分10
47秒前
孤狼完成签到 ,获得积分10
51秒前
子良完成签到 ,获得积分10
51秒前
xinxin完成签到 ,获得积分10
52秒前
黑白完成签到 ,获得积分10
52秒前
一诺相许完成签到,获得积分10
54秒前
领导范儿的应助被文文采纳,获得10
56秒前
xh完成签到,获得积分10
57秒前
QwQ完成签到 ,获得积分10
58秒前
高兴的灰狼完成签到,获得积分20
58秒前
科研猫完成签到,获得积分10
59秒前
nanting完成签到 ,获得积分10
1分钟前
momo完成签到,获得积分10
1分钟前
高分求助中
(应助此贴封号)【重要!!请各用户(尤其是新用户)详细阅读】【科研通的精品贴汇总】 10000
拉伸高分子熔体中缠结行为的模拟和小角中子散射研究 1700
The World Atlas of Wine 9th edition 1000
奈特基础组织学:配套组织病理学(第4版)(Netter's Essential Histology With Correlated Histopathology) 800
John McDowells Bildungstheorie 800
Social Psychology in the Real World 600
八旗制度の研究 600
热门求助领域 (近24小时)
化学 材料科学 医学 生物 纳米技术 计算机科学 工程类 有机化学 化学工程 生物化学 物理 复合材料 内科学 催化作用 光电子学 人工智能 细胞生物学 无机化学 物理化学 数学
热门帖子
关注 科研通微信公众号,转发送积分 8180068
求助须知:如何正确求助?哪些是违规求助? 9570224
关于积分的说明 21808724
捐赠科研通 7776054
什么是DOI,文献DOI怎么找? 3433261
关于科研通互助平台的介绍 2557425
邀请新用户注册赠送积分活动 2479012