各位好！今日与大家分享一项可切除早期NSCLC 术前血ctDNA队列研究的分析结果。研究基于800余例术前血的检测结果，为大家呈现了基线ctDNA的临床预测与预后价值。众所周知，ctDNA目前仍然存在许许多多有待解决的问题，例如panel的选择（固定化/个体化定制）、技术标准的推荐、连续性检测结果的临床价值与实际临床应用的推荐级别。那么到底早期NSCLC中基线血的ctDNA检测有哪些应用值得期待？
Clinical Utility of Tumor-Naïve Pre-surgical ctDNA Detection in Early-stage NSCLC
Tae Hee Hong, Soohyun Hwang, Abhijit Dasgupta, Chris Abbosh, Tiffany Hung, Jörg Bredno, Jill Walker, Xiaojin Shi, Tsveta Milenkova, Leora Horn, Joon Young Choi, Ho Yun Lee, Jong Ho Cho, Yong Soo Choi, Young Mog Shim, Shoujie Chai, Kate Rhodes, Manami Roychowdhury-Saha, Darren Hodgson, Hong Kwan Kim, Myung-Ju Ahn
JTO July 04, 2024
Purpose: The use of tumor-informed circulating tumor DNA (ctDNA) testing in early-stage patients before surgery is limited mainly due to restricted tissue access and extended turnaround times. This study aimed to evaluate the clinical value of a tumor-naïve, methylation-based cell-free DNA assay in a large cohort of patients with resected non-small cell lung cancer (NSCLC).
背景：手术前对早期患者进行肿瘤源-循环肿瘤DNA（ctDNA）检测的使用受到限制，主要是由于组织获取受限和转诊时间延长。这项研究旨在基于大样本队列评估基于甲基化的肿瘤处治cfDNA检测在可切除的非小细胞肺癌（SOC）患者中的临床价值。
METHODS: We analyzed pre-surgical plasma samples from 895 patients with EGFR and ALK-wild-type, clinical stage I or II NSCLC. The ctDNA status was evaluated for its prognostic significance in relation to tumor volume, metabolic activity, histology, histological subtypes, and clinical-to-pathological TNM upstaging.
方法：我们分析了895名EGFR和ALK野生型、临床I期或II期非小细胞肺癌患者的手术前血浆样本。评估了ctDNA状态与肿瘤体积、代谢活性、组织学、组织学亚型和临床到病理TNM分期上调相关的预后意义。
RESULTS: Pre-surgical ctDNA detection was observed in 55 out of 414 (13%) patients with clinical stage I lung adenocarcinoma (LUAD) and was associated with poor recurrence-free survival (RFS) (2-year RFS 69% versus 91%; log-rank P<0.001), approaching that of clinical stage II LUAD. Pre-surgical ctDNA detection was not prognostic in patients with clinical stage II LUAD or non-LUAD. Within LUAD, tumor volume and positron emission tomography avidity interacted to predict pre-surgical ctDNA detection. Moreover, pre-surgical ctDNA detection was predictive of the post-surgical discovery of IASLC G3 tumors (P<0.001) and pathological TNM upstaging (P<0.001). Notably, pre-surgical ctDNA detection strongly correlated with higher PD-L1 expression in tumors (positive rates 28% vs. 55%, P<0.001), identifying a subgroup likely to benefit from anti-PD-(L)-1 therapies.
结果：在414名临床I期肺腺癌（LUAD）患者中，55名（13%）观察到了手术前ctDNA检测，并与无复发生存率（RFS）差相关（2年RFS 69% vs 91%; log rank P<0.001），接近临床II期LUAD。对于临床II期LUAD或非LUAD患者，手术前ctDNA检测没有预后。在LUAD中，肿瘤体积和正电子发射断层扫描的代谢活性的相互作用来预测手术前的ctDNA检测。此外，手术前ctDNA检测可预测手术后发现IASLC G3肿瘤（P<0.001）和病理性TNI上调（P<0.001）。值得注意的是，手术前ctDNA检测与肿瘤中更高的PD-L1表达密切相关（阳性率28% vs 55%，P<0.001），从而确定了一个可能受益于抗PD-（L）-1疗法的亚组。
CONCLUSION: These findings support the integration of ctDNA testing into routine diagnostic workflows in early-stage NSCLC without the need of tumor tissue profiling. Furthermore, it is clinically useful in identifying high-risk patients who might benefit from innovative treatments, including neoadjuvant immune checkpoint inhibitors.
结论：这些发现支持将ctDNA检测整合到早期非小细胞肺癌的常规诊断工作流程中，而无需肿瘤组织分析。此外，它对于识别可能受益于新型治疗（包括新辅助免疫检查点抑制剂）的高危患者具有临床意义。
Keywords: NSCLC; Cell-free DNA; Cancer detection; Recurrence; Staging
1. 老规矩，先来了解下专家的观点，回顾下领域内的经典研究。
2. 其次看研究细节。
①研究由韩国学者开展（ps有点怀疑是不是财阀家搞学术的儿子做得。。。。），收集了三星医疗中心自2014-2020年期间的临床I-II期NSCLC术前血样，由于术前组织样本获取难度大、肿瘤的异质性，采用tumor-naive targeted cfDNA methylation assay方式检测。
②研究竭尽所能地探索了ctDNA在早期肺癌中与各种临床病理因素、预后的关联。仅基于单次基线血，做出这样的结果其实已经比较不错了。其中的很多数据为后续研究的样本量估算和设计提供了宝贵的基础。
③抛开结果，文章结构上整体割裂感很强。并不是一个具有系统模块化的story，特别是CTR部分没有和tumor volume/PET部分整合在一起，也没有做clinical utility的对比分析（甚至没有肺叶和亚肺叶的基线信息）和临床决策推荐（可能是结果不如意）。分析视角上是强调了现有临床证据下ctDNA的分层作用。最后一个部分强行套在PD-L1上的分析读起来是真的非常勉强，脱离生物学机制谈应用基本就是水内容了。Extended Data Fig. 4 还标的WHO Grade。。。。
3.再好的工具也要找到合适的场景。多数前瞻性研究为确保具有足够的分析角度，都会选择长程ctDNA而不是单次检测的结果来与结局进行分析。
目录
1. INTRODUCTION
2. METHODS
2.1 Study objectives
2.2 Description of study cohorts
2.3 Biomarker analyses
2.4 Tumor volume and PET SUV analyses
2.5 Statistics
2.6 Analytical validation approach
3. Results
3.1 ctDNA detection by targeted methylation analyses (Extended Data Fig. 1)
3.2 Tumor-naïve pre-surgical ctDNA detection in early-stage NSCLC (Fig. 1)
3.3 Tumor volume and PET avidity associated with ctDNA detection (Fig. 2)(Extended Data Table 1)
3.4 Pre-surgical ctDNA detection and pathological upstaging (Extended Data Table 2)(Table 1)
3.5 Pre-surgical ctDNA detection and clinical outcome (Fig. 3)(Extended Data Fig. 2-3)(Extended Data Table 3)
3.6 ctDNA status refines consolidation to tumor ratio (CTR) as a risk parameter (Fig. 4)
3.7 Pre-surgical ctDNA status predicts International Association for the Study of Lung Cancer (IASLC) LUAD histological grade (Extended Data Fig. 4)
3.8 Pre-surgical ctDNA-positive patients and PD-L1 status (Fig. 5)
4. Discussion
— 图表汇总—
3.1 ctDNA detection by targeted methylation analyses
Extended Data Fig. 1 Patients, samples, and assay.
a, Flow chart depicting patient selection process.
b, Distribution of cfDNA inputs in nanograms from the 895 patients included in this study. cfDNA values are log-transformed.
c, TMeF distribution plot by sample type (cancer vs non-cancer) of all tested participant sample replicates near LOD95 and all non-cancer sample replicates from the limit of blank study. Color was customized for each participant with cancer and unicolor was applied for all non-cancer participants.
d, Targeted methylation-based prognostic assay analytical validation results.
Analytical sensitivity at LOD95 (140 replicates from 5 LUAD participants at empirical LOD95 level) and analytical specificity (66 replicates from 62 non-cancer participants) were presented.
ca, cancer; cfDNA, cell-free DNA; ctDNA, circulating tumor DNA; LOD, limit of detection; LUAD, lung adenocarcinoma; nc, non-cancer; NSCLC, non-small-cell lung cancer; ppm, parts per million; TMeF, tumor methylated fraction; (c)/(p)TNM, (clinical)/(pathologic) tumor-node-metastasis.
3.2 Tumor-naïve pre-surgical ctDNA detection in early-stage NSCLC
Fig. 1 Clinical features associated with pre-surgical ctDNA analyses.
Clinical features in 895 clinical stage I–II patients with a, LUAD or b, non-LUAD NSCLC.
CSO row: cancer signal of origin data. Histology row: LUAD NSCLC categorized as solid and non-solid, non-LUAD NSCLC categorized as squamous or other histology. Clinical TNM row: clinical TNM (version 8) stage. Path TNM row: pathological TNM (version 8) stage. Volumetrics row: log10-transformed tumor volume (in cm3) measured using CT. Consolidation:Tumor: Measurement of consolidation to tumor ratio using pre-surgical CT. SUVmax: PET avidity measured on presurgical PET-CT; white represents no available PET data. PD-L1 row: PD-L1 percentage expression determined in resection specimen by immunhistochemistry. Methyl ctDNA level: measurement of ctDNA level in a sample based on targeted methylation assay. A non-zero methyl ctDNA level can occur in ctDNA-negative patients for whom the methyl signal was insufficient to result in confident detection of ctDNA above background healthy controls. c, Proportion of CSO calls by histology.
CSO, cancer signal of origin; ctDNA, circulating tumor DNA; CT, computed tomography; LUAD, lung adenocarcinoma; NSCLC, non-small-cell lung cancer; PD-L1, programmed death ligand-1; PET, positron emission tomography; SUV, standardized uptake value; TNM, tumor-node-metastasis.
3.3 Tumor volume and PET avidity associated with ctDNA detection
Fig. 2 Tumor volume and maximum PET avidity within a tumor associate with pre-surgical ctDNA detection using a tumor-naïve targeted methylation approach.
Tumor volume (cm3, measured on pre-surgical CT scans) is higher in pre-surgical a, (left facet) ctDNA-positive LUAD and (right facet) ctDNA-positive non-LUAD NSCLC. Tumor PET avidity as measured by SUVmax is higher in b, (left facet) ctDNA-positive LUAD and (right facet) ctDNA-positive non-LUAD NSCLC. Hinges correspond to first and third quartiles, whiskers extend to the largest/smallest value no further than 1.5x the interquartile range. Center lines represent medians. Each dot on the plot represents outlier values. P-value on plot represents Wilcoxon Test. c,d, Interaction between PET avidity and tumor volume in c, LUAD and d, non-LUAD NSCLC, determined separately for each histologic subgroup using GAM (generalized additive model) modelling. Colors on each heatmap represent the predicted probability of ctDNA detection given a set of SUVmax (y-axis, SUVmax) and tumor volume (x-axis, cm3) parameters. Gradient lines highlight probability of detection in 0.1 intervals.
ctDNA, circulating tumor DNA; CT, computed tomography; GAM, generalized additive model. LUAD, lung adenocarcinoma; NSCLC, non-small-cell lung cancer; PET, positron emission tomography; SUV, standardized uptake value.
Extended Data Table 1. Logistic regression of ctDNA-positive status on tumor volume and avidity.
CI, confidence interval; ctDNA, circulating tumor DNA; LUAD, lung adenocarcinoma; OR, odds ratio; SUV, standardized uptake value.
3.4 Pre-surgical ctDNA detection and pathological upstaging
Extended Data Table 2 Associations between ctDNA status and upstaging in clinical stage I–II NSCLC.
CI, confidence interval; ctDNA, circulating tumor DNA; LUAD, lung adenocarcinoma; NSCLC, non-small-cell lung cancer.
Table 1 Associations of ctDNA positivity and upstaging risk.
ctDNA, circulating tumor DNA; LUAD, lung adenocarcinoma; NPV, negative predictive value; PPV, positive predictive value.
3.5 Pre-surgical ctDNA detection and clinical outcome
Fig. 3 Pre-surgical ctDNA associations with clinical outcome in clinical stage I–II LUAD.
a,b, RFS and c,d, OS were associated with pre-surgical ctDNA status. N = 414 patients with clinical stage I LUAD analyzed, 55/414 were ctDNA positive. N = 140 patients with clinical stage II LUAD analyzed, 66/140 were ctDNA positive. P-values on Kaplan‒Meier graphs represent log-rank P-values.
ctDNA, circulating tumor DNA; LUAD, lung adenocarcinoma; NSCLC, non-small-cell lung cancer; OS, overall survival; RFS, relapse-free survival.
Extended Data Fig. 2 RFS by clinical stage and ctDNA status.
Kaplan–Meier analysis of RFS in clinical stage I LUAD (stratified by ctDNA status) and clinical stage II LUAD (all comers).
ctDNA, circulating tumor DNA; LUAD, lung adenocarcinoma; RFS, relapse-free survival.
Extended Data Fig. 3 Pre-surgical ctDNA associations with clinical outcome in clinical stage I or II non-LUAD histologies.
a,b, RFS and c,d, OS were associated with pre-surgical ctDNA status. Of 188 patients with clinical stage I non-LUAD analyzed, 133/188 were ctDNA positive. Of 153 patients with clinical stage II non-LUAD analyzed, 143/153 were ctDNA positive. P-values on Kaplan–Meier graphs represent log-rank P-values.
e,f, RFS and g,h, OS were also associated with TMeF level (above or below the median in ctDNA-positive cases, median TMeF = 52 ppm).
ctDNA, circulating tumor DNA; LUAD, lung adenocarcinoma; OS, overall survival; RFS, relapse-free survival; TMeF, tumor methylated fraction.
Extended Data Table 3. Multivariable Cox regression analyses of pre-surgical ctDNA status, adjusted for stage sub-category, age, gender, and smoking status, and association with RFS and OS in clinical stage I LUAD. N = 414 patients subject to analysis.
CI, confidence interval; ctDNA, circulating tumor DNA; HR, hazard ratio; LUAD, lung adenocarcinoma; NSCLC, non-small-cell lung cancer; OS, overall survival; RFS, relapse-free survival.
3.6 ctDNA status refines consolidation to tumor ratio (CTR) as a risk parameter
Fig. 4 Integration of pre-surgical ctDNA status with CT-determined CTR.
a, Bar chart demonstrating the existence of pre-surgical ctDNA positive patients with LUAD NSCLC across IA1–IB clinical stage brackets (TNM version 8).
b, RFS and c, OS were associated with categories based on integrating pre-surgical ctDNA status and solid nodule status in patients with clinical stage I LUAD. P-values represent log-rank P-values.
N = 414 patients with clinical stage I LUAD analyzed; 242/414 had non-solid nodules and were ctDNA negative, 130/414 had solid nodules and were ctDNA negative, and 42/414 had solid nodules and were ctDNA positive.
ctDNA, circulating tumor DNA; CTR, consolidation to tumor ratio; LUAD, lung adenocarcinoma; neg, negative; NSCLC, non-small-cell lung cancer; OS, overall survival; pos, positive; RFS, relapse-free survival; TNM, tumor-node-metastasis.
3.7 Pre-surgical ctDNA status predicts International Association for the Study of Lung Cancer (IASLC) LUAD histological grade
Extended Data Fig. 4 RFS, OS, and clinical stage I patients by histologic grade.
IASLC histologic grade (G1‒3) associations with a, RFS and b, OS in 345 patients with pathologically confirmed stage I invasive LUAD. P-values on Kaplan–Meier graphs represent log-rank P-values. c, Bar chart showing that pre-surgical ctDNA-positive status enriches in grade 3 histology.
3.8 Pre-surgical ctDNA-positive patients and PD-L1 status
Fig. 5 Pre-surgical ctDNA status associates with PD-L1 expression levels observed in resected LUAD specimens.
N = 414 patients with clinical stage I LUAD analyzed and N = 140 patients with clinical stage II LUAD analyzed.
ctDNA, circulating tumor DNA; LUAD, lung adenocarcinoma; PD-L1, programmed death ligand-1.
