Accessible Published by De Gruyter May 6, 2021

The assessment of circulating cell-free DNA as a diagnostic tool for breast cancer: an updated systematic review and meta-analysis of quantitative and qualitative ssays

Qingfeng Guo and Yuming Hua

Abstract

Objectives

This updated meta-analysis aimed to assess the diagnostic accuracy of circulating cell-free DNA (cfDNA) in breast cancer (BC).

Content

An extensive systematic search was performed in PubMed, Scopus, Embase, and Science Direct databases to retrieve all related literature. Various diagnostic estimates, including sensitivity (SE), specificity (SP), likelihood ratios (LRs), diagnostic odds ratio (DOR), and area under the curve (AUC) of summary receiver operating characteristic (sROC) curve, were also calculated using bivariate linear mixed models.

Summary

In this meta-analysis, 57 unique articles (130 assays) on 4246 BC patients and 2,952 controls, were enrolled. For quantitative approaches, pooled SE, SP, PLR, NLR, DOR, and AUC were obtained as 0.80, 0.88, 6.7, 0.23, 29, and 0.91, respectively. Moreover, for qualitative approaches, pooled SE and SP for diagnostic performance were obtained as 0.36 and 0.98, respectively. In addition, PLR was 14.9 and NLR was 0.66. As well, the combined DOR was 23, and the AUC was 0.79.

Outlook

Regardless of promising SE and SP, analysis of LRs suggested that quantitative assays are not robust enough neither for BC confirmation nor for its exclusion. On the other hand, qualitative assays showed satisfying performance only for confirming the diagnosis of BC, but not for its exclusion. Furthermore, qualitative cfDNA assays showed a better diagnostic performance in patients at the advanced stage of cancer, which represented no remarkable clinical significance as a biomarker for early detection.

Introduction

Breast cancer (BC) is known as the most common cancer with the highest mortality rate among women [1]. Early detection methods have helped in reducing BC mortality [2]. Currently, it was indicated that mammography as a gold standard method used for early detection of BC can reduce mortality up to 30% [3]. However, 13% of BC patients are missed by mammography due to some factors such as overlapping with dense fibroglandular tissue, tumor size, and patient’s age [4], [5]. Additionally, false-positive detection in mammograms may cause over-treatment as well as radiation-induced disease [6], [7]. The current serum-based biomarkers such as cancer antigen 15-3 (CA15-3) and carcinoembryonic antigen (CEA) also did not show satisfactory result regarding highly accurate diagnosis of BC [8], [9]. Therefore, it is necessary to develop novel minimally-invasive biomarkers, in order to achieve a sufficient diagnostic performance for early detection of BC in the clinical practice.

Circulating cell-free DNA (cfDNA) can be found in both healthy individuals and cancer patients; however, its level in cancer patients has been reported to be significantly higher compared to normal people [10]. The release of cfDNA into the bloodstream from cancer cells is called circulating tumor DNA (ctDNA). Correspondingly, the amount of cf DNA is determined by the level of necrosis and apoptosis [11], [12], [13]. Different valuable information on tumor drug resistance, clonal heterogeneity, and disease progression can be achieved by analyzing cfDNA via the evaluation of cfDNA for aberrant level, gene mutations, microsatellite alteration, loss of heterozygosity, single nucleotide polymorphism, and aberrant methylation [14], [15]. Different sensitive techniques such as sequencing-based methods [16], [17], [18] and digital PCR-based approaches [19], [20], [21] can also be applied to evaluate cfDNA in various cancer types, including hepatocellular carcinoma, lung, pancreatic carcinoma, colorectal, prostate, ovarian, and breast [16, 17, 19, 22]. Being repetitive, noninvasive, and easy to perform the analysis, are known as the major benefits of cfDNA as a diagnostic tool. Moreover, it is notable that CfDNA holds a great promise as a supplement or alternative for conventional biomarkers and tissue biopsy in early diagnosis of tumor, assessing tumor’s burden, and treatment monitoring.

Several studies have previously assessed the diagnostic accuracy of circulating cfDNA in terms of quantitative and qualitative changes among BC patients. By the term of Quantitative changes, it is referred to aberrant levels of cfDNA and circulating nucleosomes, while qualitative changes include epigenetic modifications, loss of heterozygosity (LOH), and single gene mutation. However, inconsistent results regarding its clinical use still exist in the previously performed works. Hence, this updated systematic review and meta-analysis aimed to integrate data from previous studies, in order to evaluate the diagnostic performance of cfDNA in BC diagnosis.

Methods

Literature search strategy

In this study, we performed an extensive search to find relevant articles in PubMed, Science Direct, Web of Science, and Scopus up databases to August 31, 2020, without applying any limit. A combination of keywords has been generated by arranging the medical subject heading (MeSH) terms and free words concerning search tips in each database as follows: (“circulating” OR “plasma” OR “serum” OR “cell-free”) AND (“DNA” OR “cfDNA” OR “ctDNA”) AND (“breast cancer” OR “breast neoplasm”) AND (“diagnostic accuracy” OR “diagnostic”). Furthermore, the reference lists of the obtained relevant literatures have been evaluated to find additional studies. Ineligible studies have been excluded by reading their titles and abstracts, then the full texts of each one of the remained study have been retrieved and comprehensively evaluated. This study has been performed in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) Statement [23].

Inclusion and exclusion criteria

The inclusion criteria were as follows: (a) all the patients recruited in the study must be diagnosed with BC using a gold standard test; (b) a case control design evaluating the diagnostic accuracy of cfDNA in plasma or serum for BC; and (c) availability of raw data on diagnostic accuracy measures including true positives (TP), true negatives (TN), false positives (FP), and false negatives (FN) or some estimates such as sensitivity (SE) and specificity (SP).

The main exclusion criteria were the followings: (a) duplicate studies; (b) meta-analyses, reviews, book chapters, conference abstracts, and letters to the editors; (c) unavailability of sufficient data; and (d) unavailability of full texts.

Data extraction and quality assessment

At this stage, we extracted the following data from the enrolled articles: authors’ name, year of publication, country of origin, design of the study, type of controls, type of samples, experimental methods, assay type, cutoff values, and number of the participants. Additionally, the numerical values for TP, FP, FN, and TN have been directly achieved or extracted from given SE and SP. The quality of the enrolled studies have been assessed in terms of the revised quality assessment of diagnostic accuracy studies-2 (QUADAS-2) criteria [24]. Finally, discrepancies have been resolved by consensus.

Biostatistical methods

Several diagnostic estimations were used to present the results of this meta-synthesis, including SE, SP, likelihood ratios (LRs), and diagnostic odds ratio (DOR). In addition, area under the curve of summary receiver operating characteristic curves (sROC-AUC) was used as a global measurement for testing the accuracy. Subgroup analysis and meta-regression were also performed to evaluate the effect of different covariates and potential sources of heterogeneity on the diagnostic measures [25]. Moreover, Fagan’s nomogram and likelihood ratio scattergram were used to present the clinical utility [26]. Publication bias was assessed using Deek’s funnel plot asymmetry test [27]. Accordingly, trim and fill analysis were applied in case of any publication bias to judge the effect of bias on the pooled estimates [28]. The threshold effect was assessed by Spearman’s correlation coefficient test, and I 2 and χ2 were used to evaluate the heterogeneity among the enrolled articles [29], [30]. Of note, all these test measures were calculated by bivariate mixed effects regression using Stata 14.2 which incorporates any possible correlation between SE and SP using a random effect model (version 14.2; Stata Corporation, College Station, TX) and Meta-DiSc (version 1.4, Ramon y Cajal Hospital, Madrid, Spain).

Results

Selection of studies

The procedures of literature recapturing and inclusion are presented in Figure 1. In this regard, we found 1,396 relevant articles from different databases. After excluding 282 duplicated articles, 955 articles were also excluded through the evaluation of their titles and abstracts due to having unrelated study design or unavailability of their data. Moreover, of the remaining 159 papers, 102 have been excluded due to details presented in Figure 1. Subsequently, this meta-analysis has been performed on 57 articles at last (130 assays).

Figure 1: 
PRISMA flowchart of the identification of eligible studies.

Figure 1:

PRISMA flowchart of the identification of eligible studies.

Baseline characteristics and quality assessment

Table 1 represents the characteristics of 57 included articles (130 assays) [21, 31], [32], [33], [34], [35], [36], [37], [38], [39], [40], [41], [42], [43], [44], [45], [46], [47], [48], [49], [50], [51], [52], [53], [54], [55], [56], [57], [58], [59], [60], [61], [62], [63], [64], [65], [66], [67], [68], [69], [70], [71], [72], [73], [74], [75], [76], [77], [78], [79], [80], [81], [82], [83], [84], [85], [86]. Accordingly, a total of 7,198 individuals, including 4246 BC patients and 2,952 controls, were enrolled in this diagnostic analysis. Healthy volunteers were selected as the single control group in 98 assays and patients with benign breast lesions were assigned as the only control group in nine assays. In addition, in 23 assays, a mix of healthy volunteers and patients with benign breast lesions was chosen as the control group. In terms of regions, 57 assays were conducted in Asia (44%), 11 in Africa (9%), 46 in Europe (35%), and 16 in America (12%). Of note, out of the selected articles, four articles had a prospective design, 28 articles had a retrospective design, and 25 articles had no clear design. One hundred and one assays evaluated cfDNA as qualitative indicators including loss of heterozygosity (LOH) (n=3), allelic imbalance (AI) (n=1), single gene mutation (Mut) (n=4), microsatellite instability (MI) (n=1), repetitive elements (n=1) and aberrant methylation (M) (n=91). Meanwhile, 29 assays evaluated cfDNA integrity (DI) index (n=7) and aberrant concentrations of cfDNA as a quantitative indicator (n=22). In 60 assays the sample size was ≥100 cases while in the remaining assays the sample size was <100 cases. The sample types were categorized as plasma (n=50 assays) and serum (n=80 assays).

Table 1:

Characteristics of the included studies.

Study ID Country Sample Design Controls Method Assay type Cutoff No. P/C TP/FP/FN/TN
Author, year
Agassi et al. (2015) Isreal Serum Retrospective Healthy Fluorescent assay Q 600 ng/mL 36/16 26/4/10/12
Agostini et al. (2012) Italy Plasma Retrospective Healthy RT-PCR Q (ALU 247) 2 ng/mL 39/49 39/0/0/49
RT-PCR Q (ALU 115) 9.3 ng/mL 39/49 37/0/2/49
RT-PCR DI 0.6 49/49 36/11/13/38
MSP M (RASSF1A) NA 39/49 9/0/30/49
MSP M (MAL) NA 39/49 5/0/34/49
MSP M (SFPR1) NA 39/49 16/0/23/49
Beaver et al. (2014) USA Plasma Retrospective Healthy ddPCR Mut (PI3CA) NA 15/14 14/0/1/14
Beck et al. (2010) USA Serum Retrospective Healthy NGS Repetitive elements NA 10/87 9/4/1/83
Board et al. (2010) UK Serum Retrospective Healthy ARMS-PCR Mut (PI3CA) NA 14/16 0/0/14/16
ARMS-PCR Mut (PI3CA) NA 10/31 8/1/2/30
Catarino et al. (2008) Portugal Plasma Retrospective Healthy RT-PCR Q 106 ng/mL 175/80 76/7/99/73
Chen et al. (1999) Switzerland Serum NA Healthy QPCR LOH 30% 21/10 10/0/11/10
Chimonidou et al. (2013a) Greece Plasma NA Healthy RT-MSP M (Sox17) NA 138/49 43/1/95/48
Chimonidou et al. (2013b) Greece Plasma Retrospective Healthy MSP M (CST6) NA 123/37 49/0/74/37
Divella et al. (2009) Italy Plasma NA Mixed RT-PCR Q NA 122/80 61/8/61/72
Dulaimi et al. (2004) USA Serum NA Healthy MSP M (APC) NA 34/20 10/0/24/20
MSP M (RASSF1A) NA 34/20 19/0/15/20
MSP M (DAPK) NA 34/20 12/0/22/20
Ebeid et al. (2016) Egypt Serum NA Healthy MSP M (RASSF1A) NA 80/80 50/0/30/80
Egypt Serum NA Healthy MSP M (PCDH10) NA 80/80 34/0/46/80
Fu et al. (2015) China Plasma Retrospective Healthy MSP M (Sox17) NA 155/60 90/0/65/60
Gal et al. (2004) UK Serum Retrospective Healthy PCR Q NA 96/24 68/2/28/22
Geng et al. (2013) China Plasma Retrospective Healthy RT-PCR Q 2.49 106 copies/mL) 229/43 181/10/48/33
Gong et al. (2012) China Serum Retrospective Mixed PCR Q 471 ng/mL 200/200 178/10/22/190
Hashad et al. (2012) Egypt Serum NA Healthy RT-PCR Q 34 42/27 41/1/1/26
Hoque et al. (2006) Senegal Plasma Retrospective Healthy MSP M (GSTP1) 0 47/38 12/0/35/38
Retrospective Healthy MSP M (RAR β2) 2.6 × 10−8 47/38 12/3/35/35
Retrospective Healthy MSP M (RASSF1A) 6.5 × 10−3 47/38 15/2/32/36
Retrospective Healthy MSP M (APC) 3.2 47/38 8/0/39/38
Huang et al. (2006) China Plasma Retrospective Healthy RT-PCR Q 19 ng/mL 61/27 58/3/3/24
NBC RT-PCR Q 22 ng/mL 61/33 57/11/4/22
Hussein et al. (2019) Egypt Plasma Retrospective Healthy RT-PCR Q (ALU 115) >9.6 40/10 27/0/13/10
Healthy RT-PCR DI (ALU 247/ALU 115) >6.35 40/10 28/0/12/10
Jing et al. (2007) China Serum NA Healthy MSP M (BRCA1) NA 38/20 14/0/24/20
MSP M (p16) NA 38/20 13/0/25/20
MSP M (14-3-3ϱ) NA 38/20 33/0/5/20
Jing et al. (2008) China Serum NA Healthy MSP M (BRCA1) NA 102/20 33/0/69/20
China Serum NA Healthy MSP M (p16) NA 102/20 30/0/72/20
China Serum NA Healthy MSP M (14-3-3ϱ) NA 102/20 84/0/18/20
Jing et al. (2010) China Serum NA Mixed MSP M (RASSF1A) NA 50/50 37/3/13/47
MSP M (CDH1) NA 50/50 30/5/20/45
MSP M (RAR β2) NA 50/50 21/2/29/48
MSP M (BRCA1) NA 50/50 17/0/33/50
MSP M (p16) NA 50/50 15/0/35/50
MSP M (Erα) NA 50/50 23/5/27/45
MSP M (COX2) NA 50/50 6/1/44/49
MSP M (APC) NA 50/50 14/0/36/50
MSP M (DAPK) NA 50/50 9/0/41/50
MSP M (FHIT) NA 50/50 5/0/45/50
Kim et al. (2010) Korea Serum Retrospective Healthy QM-MSP M (HIN-1) 0.03% 119/125 37/0/82/125
QM-MSP M (RASSF1A) 0.08% 119/125 39/6/80/119
QM-MSP M (RAR-B) 1.04% 119/125 103/8/16/117
QM-MSP M (Twist) 0.86% 119/125 65/10/54/115
Kirkizlar et al. (2015) USA Serum NA Healthy NGS AI 0.45% 11/30 8/0/3/30
Zachariah et al. (2008) Switzerland Plasma Retrospective Healthy Multiplex RT-PCR Q (nDNA) 1866 GE/mL 52/70 42/22/10/48
Multiplex RT-PCR Q (mtDNA) 463,282 GE/mL 52/70 28/9/24/61
Leon et al. (1977) USA Serum Retrospective Healthy Radioimmunoassay Q 50 ng/mL 32/55 12/4/20/51
Li et al. (2016) China Plasma NA Healthy BS-seq M (EGFR) NA 86/67 70/41/16/26
BS-seq M (PPM1A) NA 86/67 65/37/21/30
Li et al. (2018) China Plasma Retrospective Mixed BS-seq M (NBPF1) NA 52/61 35/27/17/34
Mahmoud et al. (2015) Egypt Plasma Retrospective NBC RT-PCR Q (nDNA) 2,236 copy/μL 50/30 38/9/12/21
RT-PCR Q (mtDNA) 376 copy/μL 50/30 30/11/20/19
Martínez-Galán et al. (2008) Spain Serum NA Healthy QMS-PCR M (14-3-3ϱ) 0.05 106/74 79/35/27/39
QMS-PCR M (ESR1) 0.02 106/74 70/28/36/46
Matuschek et al. (2010) Germany Serum Prospective Healthy MSP M (APC) NA 85/22 25/2/60/20
MSP M (RASSF1A) NA 85/22 22/5/63/17
MSP M (GSTP1) NA 76/16 14/1/62/15
MSP M (ESR1) NA 85/22 32/5/53/17
Mirza et al. (2007) India Serum NA Healthy MSP M (Erα) NA 36/5 16/0/20/5
MSP M (PRB) NA 36/5 16/0/20/5
MSP M (BRCA1) NA 36/5 9/0/27/5
MSP M (TMS1) NA 36/5 4/0/32/5
Mirza et al. (2010) Canada Serum Prospective Healthy MSP M (Erα) NA 55/30 30/3/25/27
MSP M (PRB) NA 53/53 29/4/24/49
MSP M (SFN) NA 56/56 33/7/23/49
Müller et al. (2003) Austria Serum NA Healthy Fluorescent base PCR M (ESR1) NA 25/10 7/0/18/10
Fluorescent base PCR M (APC) NA 25/10 6/0/19/10
Fluorescent base PCR M (HIC-1) NA 25/10 10/1/15/9
Fluorescent base PCR M (RASSF1A) NA 25/10 6/1/19/9
Ng et al. (2011) China Plasma Retrospective Healthy MSRED-qPCR M (SLC19A3) 0.41 39/20 34/3/5/17
Nunes et al. (2018) Portugal Plasma Retrospective Mixed MSP M (SCGB3A1) NA 108/103 18/21/90/82
Oshiro et al. (2015) Japan Serum Retrospective Healthy dPCR Mut (PI3CA) NA 110/30 25/0/85/30
Panagopoulou et al. (2019) Greece Plasma Retrospective Healthy qMSP M (SOX17) NA 34/35 23/2/11/33
qMSP M (WNT5A) NA 34/35 14/3/20/32
qMSP M (KLK10) NA 34/35 19/2/15/33
qMSP M (MSH2) NA 34/35 16/1/18/34
qMSP M (GATA3) NA 35/35 3/1/32/34
Papadopoulou et al. (2006) Greece Plasma NA Healthy QPCR Q 10.8 ng/mL 68/54 33/3/35/51
MSP M (RASSF1A) NA 50/14 13/0/37/14
MSP M (ATM) NA 50/9 7/0/43/9
Rykova et al. (2004) Russia Plasma NA NBC MSP M (RASSF1A) NA 17/6 8/0/9/6
Russia Plasma NA NBC MSP M (RAR β2) NA 15/6 5/0/10/6
Russia Plasma NA NBC MSP M (APC) NA 10/6 4/0/6/6
Schwarzenbach et al. (2011) Germany Serum NA Healthy PCR-based fluorescence microsatellite analyses LOH NA 102/53 19/0/83/53
Shan et al. (2016) China Serum Retrospective Mixed MSP M (hMLH1) NA 268/481 75/87/193/394
MSP M (RASSF1A) NA 268/481 46/44/222/437
MSP M (p16) NA 268/481 60/78/208/403
MSP M (PCDHGB7) NA 268/481 149/223/119/258
MSP M (SFN) NA 268/481 197/287/71/194
MSP M (HOXD13) NA 268/481 37/7/231/474
Sharma et al. (2007) India Serum NA Healthy MSP M (Cyclin D2) NA 36/4 9/0/27/4
India Serum MSP M (p16) NA 36/4 13/0/23/4
India Serum MSP M (p14) NA 36/4 13/0/23/4
India Serum MSP M (Slit2) NA 36/4 21/0/15/4
Sharma et al. (2010) India Serum NA Healthy MSP M (BRCA1) NA 100/30 22/1/78/29
MSP M (MGMT) NA 100/30 26/1/74/29
MSP M (GSTP1) NA 100/30 22/1/78/29
Shaw et al. (2000) UK Plasma Prospective NBC QPCR LOH NA 71/9 22/0/49/9
UK Plasma Prospective NBC QPCR MI NA 71/9 8/0/63/9
Shukla et al. (2006) India Serum NA Healthy MSP M (RASSF1A) NA 20/20 15/0/5/20
Skvortsova et al. (2006) Russia Plasma NA Mixed MSP M (HIC-1) NA 20/25 11/6/9/19
MSP M (RASSF1A) NA 20/25 3/1/17/24
MSP M (RAR β2) NA 20/25 3/2/17/23
Stötzer et al. (2014) Germany Plasma NA Healthy Q-PCR DI (ALU 115) NA 112/28 71/0/41/28
Q-PCR DI (ALU 247) NA 112/28 74/0/38/28
Taback et al. (2006) USA Serum Prospective Healthy MSP M (RASSF1A) NA 33/10 5/0/28/10
MSP M (MGMT) NA 33/10 2/0/31/10
MSP M (RAR β2) NA 33/10 1/0/32/10
MSP M (APC) NA 33/10 1/0/32/10
Tang et al. (2018) China Serum Retrospective Healthy Q-PCR Q (ALU 115) 300.96 ng/mL 40/40 26/12/14/28
Q-PCR DI (ALU 247/ALU 115) 0.78 ng/m 40/40 37/5/3/35
Tangvarasittichai et al. (2015) Thailand Plasma NA Healthy QPCR Q 120 ng/mL 100/100 100/13/0/87
Umetani et al. (2006) USA Serum Retrospective Healthy Q-PCR DI (ALU 247/ALU 115) 0.17 83/51 57/10/26/41
Wang et al. (2017) China Plasma NA NBC QPCR DI 0.74 51/28 37/14/14/14
Wu and Tanaka (2015) USA Plasma NA Healthy QPCR Q (telomeric DNA) 91.4 47/42 43/10/4/32
Yamamoto et al. (2012) Japan Serum Retrospective Healthy OS-MSP M (GSTP1) NA 154/87 21/2/133/85
OS-MSP M (RASSF1A) NA 154/87 26/0/128/87
OS-MSP M (RAR-B) NA 154/87 26/4/128/83
Zhang et al. (2015) China Serum Retrospective Healthy Q-PCR Q 0.3 100/104 80/33/20/71

  1. M, methylation; Q, quantitative; LOH, loss of heterozygosity; AI, allelic imbalance; DI, DNA instability; Mut, mutation; PCR, polymerase chain reaction; MSP, methylation specific PCR; RFLP, restriction fragment length polymorphism; RT-PCR, reverse transcription polymerase chain reaction; ddPCR, droplet digital PCR; BS-Seq, bisulfite sequencing; OS-MSP, one step methylation specific PCR; RASSF1A, Ras association domain family member 1; GSTP1, glutathione S-transferase pi 1; CDKI2A, cyclin dependent kinase inhibitor 2A; BRCA1, BRCA1 DNA repair associated; RAR-β2, retinoic acid receptor beta 2; APC, adenomatous polyposis coli; MGMT, o-6-methylguanine-DNA methyltransferase; Slit2, slit guidance ligand 2; HOXD13, homeobox D13; SFN, stratifin; PCDHGB7, protocadherin gamma subfamily B, 7; hMLH1, MutL homolog 1; ATM, ATM serine/threonine kinase; GATA3, GATA binding protein 3; MSH2, MutS homolog 2; KLK10, Kallikrein related peptidase 10; WNT5A, Wnt family member 5A; PI3CA, phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit alpha; SCGB3A1, secretoglobin family 3A member 1; SLC19A3, solute carrier family 19 member 3; HIC-1, HIC ZBTB transcriptional repressor 1; ER1, Estrogen receptor 1; ERα, Estrogen receptor alpha; NBPF1, NBPF member 1; PPMA1, protein phosphatase, Mg2+/Mn2+ Dependent 1A; EGFR1, Epidermal Growth Factor Receptor; TWIST1, twist family BHLH transcription factor 1; PCDH10, protocadherin 10; DAPK1, death associated protein kinase 1; TP, true positive; FP, false positive; TN, true negative; FN, false negative; SE, sensitivity; SP, specificity; NA, not available.

The result of the quality assessment in terms of QUADAS-2 criteria is presented in Figure 2. Accordingly, the majority of the enrolled studies fulfilled these criteria.

Figure 2: 
Quality assessment of included studies based on the Quality Assessment of Diagnostic Accuracy Studies (QUADAS-2) criterion. Each bar represents the percent of studies considered as high risk, low risk, or unclear for both risk of bias and applicability concerns.

Figure 2:

Quality assessment of included studies based on the Quality Assessment of Diagnostic Accuracy Studies (QUADAS-2) criterion. Each bar represents the percent of studies considered as high risk, low risk, or unclear for both risk of bias and applicability concerns.

Diagnostic accuracy

The pooled SE, SP, PLR, NLR, and DOR are considered as indicators employed to measure the diagnostic performance. For quantitative approaches, the pooled SE, SP, PLR, NLR, DOR, and AUC were obtained as 0.80 (95% CI: 0.71–0.86, I 2: 92.49%) (Figure 3A), 0.88 (95% CI: 0.82–0.92, I 2: 85.26%) (Figure 3B), 6.7 (95% CI: 4.4–10.3, I 2: 80.85%) (Figure 3C), 0.23 (95% CI: 0.16–0.34, I 2: 92.49%) (Figure 3D), 29 (95% CI: 14–57, I 2: 99%) (Figure 4), and 0.91 (Figure 5A), respectively.

Figure 3: 
Forest plots for diagnostic performance estimates of quantitative circulating cell-free DNA assays. (A) Sensitivity (B) Specificity (C) Positive likelihood ratio (D) Negative likelihood ratio.

Figure 3:

Forest plots for diagnostic performance estimates of quantitative circulating cell-free DNA assays. (A) Sensitivity (B) Specificity (C) Positive likelihood ratio (D) Negative likelihood ratio.

Figure 4: 
Forest plot for the pooled diagnostic odds ratio of quantitative circulating cell-free DNA in discriminating patients with breast cancer from controls.

Figure 4:

Forest plot for the pooled diagnostic odds ratio of quantitative circulating cell-free DNA in discriminating patients with breast cancer from controls.

Figure 5: 
The summary receiver operating characteristic (SROC) curves for: (A) Diagnostic accuracy of quantitative circulating cell-free DNA assays in discriminating patients with breast cancer from controls. (B) Diagnostic accuracy of qualitative circulating cell-free DNA assays in discriminating patients with breast cancer from controls.

Figure 5:

The summary receiver operating characteristic (SROC) curves for: (A) Diagnostic accuracy of quantitative circulating cell-free DNA assays in discriminating patients with breast cancer from controls. (B) Diagnostic accuracy of qualitative circulating cell-free DNA assays in discriminating patients with breast cancer from controls.

Furthermore, for qualitative approaches, the pooled SE and SP of cfDNA in the detection BC were obtained as 0.36 (95% CI: 0.31–0.41, I 2: 93.26%) (Supplementary Material 1; Figure 1) and 0.98 (95% CI: 0.96–0.99, I 2: 95.82%) (Supplementary Material 1; Figure 2), respectively. Besides, PLR was calculated as 14.9 (95% CI: 9.2–4.2, I 2: 88.43%) (Supplementary Material 1; Figure 3) and NLR was estimated to be 0.66 (95% CI: 0.61–0.71, I 2: 90.35%) (Supplementary Material 1; Figure 4). The combined DOR was 23 (95% CI: 14–37; I 2: 99%) (Supplementary Material 1; Figure 5), and the AUC for sROC was obtained as 0.79 (Figure 5B).

Subgroups

At this stage, we evaluated the diagnostic performance estimations in different subgroups based on the ethnicity of participants (Asian, Caucasian, and African), BC type (primary and metastatic), tumor stage (size) (T1–T2 vs. T3–T4), histological Grade (G1–G2 vs. G3), lymph node status (positive compared to negative), qualitative assays (methylation compared to mutation), methylation genes (RASSF1A, RAR-β2, p16, GSTP1, BRCA1, and APC), study design (prospective compared to retrospective), sample type (plasma compared to serum), publication year (before 2010 compared to after 2010), and sample size (≥100 cases compared to <100 cases) (Table 2 and Supplementary Material 1; Table 1).

Table 2:

Subgroups for qualitative assays.

Subgroup SE (95% CI) I 2, % SP (95% CI) I 2, % PLR (95% CI) I 2, % NLR (95% CI) I 2, % AUC (95% CI) DOR (95% CI) I 2, %
Ethnicity
Asian 0.39 (0.32–0.47) 95.6 0.97 (0.94–0.99) 97.5 13.2 (6.7–26.3) 92.6 0.53 (0.47–0.62) 93.5 0.63 (0.56–0.71) 21 (10–43) 99
Caucasian 0.33 (0.27–0.41) 88.7 0.97 (0.95–0.99) 84.4 12.8 (6.6–24.8) 63.7 0.68 (0.62–0.78) 85.5 0.80 (0.78–0.83) 19 (9–37) 99
African 0.34 (0.23–0.47) 88 0.99 (0.90–1) 70 49.8 (2.7–925.6) 35 0.67 (0.55–0.81) 84.8 0.71 (0.61–0.74) 74 (4–1,533) 99
BC type
PBC 0.41 (0.39–0.42) 91.5 0.93 (0.92–0.94) 85.8 7.51 (5.1–10.8) 82.5 0.65 (0.60–0.70) 89.1 0.82 (0.71–0.74) 13 (9–18) 50.3
MBC 0.53 (0.37–0.63) 88.6 0.91 (0.78–0.97) 91.4 6 (2.6–14) 79.8 0.52 (0.38–0.70) 95.1 0.79 (0.72–0.82) 12 (5–28) 99
Tumor stage
T1–T2 0.31 (0.22–0.42) 82.2 0.93 (0.88–0.96) 48.2 4.5 (2.6–7.8) 2.78 0.74 (0.64–0.85) 75.7 0.82 (0.71–0.74) 6 (3–12) 99
T3–T4 0.40 (0.33–0.48) 49.7 0.94 (0.98–0.97) 41 6.4 (3.6–11.3) 0 0.65 (0.60–0.70) 15.1 0.72 (0.68–0.75) 10 (6–18) 49
Histo. Grade
G1–G2 0.30 (0.27–32) 93 0.94 (0.92–0.97) 91 5 (2.53–10.1) 82.9 0.73 (0.65–0.83) 94 0.69 (0.62–0.72) 7 (3–15) 49
G3 0.39 (0.30–0.48) 84.9 0.98 (0.94–0.99) 95.4 19.8 (6.4–61.2) 74.1 0.62 (0.54–0.73) 72 0.73 (0.69–0.77) 32 (10–102) 99
Nodal status
Positive 0.43 (0.33–0.53) 88.4 0.98 (0.94–0.99) 91.6 24.6 (7.3–83.6) 68.1 0.58 (0.49–0.70) 78.6 0.73 (0.69–0.77) 42 (12–153) 99
Negative 0.35 (0.26–0.45) 86.7 0.98 (0.94–1) 93 21 (5.8–75.7) 64.1 0.66 (0.57–0.76) 77.9 0.72 (0.67–0.76) 32 (8–118) 99
cfDNA assay 99
Methylation (overall) 0.36 (0.31–0.41) 93.45 0.97 (0.95–0.98) 95.9 12.6 (7.8–20.1) 87.8 0.66 (0.62–0.71) 89.8 0.76 (0.72–0.80) 19 (12–31)
RASSF1A 0.34 (0.24–0.45) 91.17 0.98 (0.93–0.99) 84.22 14.7 (4.6–47.6) 81.92 0.68 (0.57–0.80) 92.53 0.82 (0.78–0.85) 22 (6–76) 99
RAR-β2 0.28 (0.11–0.54) 97 0.95 (0.91–0.97) 0.0 5.3 (2.2–13) 5.71 0.76 (0.54–1) 89.9 0.82 (0.78–0.85) 7 (2–23) 99
p16 0.29 (0.23–0.35) 56.8 1 (0.35–1) 88.7 124.1 (0.2–100767.7) 80.1 0.71 (0.65–0.78) 85.2 0.38 (0.34–0.42) 124 (0–146861) 99

Subgroup SE (95% CI) I 2, % SP (95% CI) I 2, % PLR (95% CI) I 2, % NLR (95% CI) I 2, % AUC (95% CI) DOR (95% CI) I 2, %

GSTP1 0.19 (0.14–0.24) 37 0.98 (0.94–0.99) 0 8.1 (2.8–23.7) 0 0.83 (0.78–0.89) 36.2 0.93 (0.90–0.95) 10 (3–30) 54.9
BRCA1 0.29 (0.24–0.36) 18.8 1 (0.85–1) 0 70.8 (1.7–2,995.2) 0 0.71 (0.65–0.77) 31.5 0.47 (0.43–0.52) 100 (2–4,384) 91
APC 0.22 (0.16–0.30) 55 1 (0.45–1) 52.1 184.1 (0.2–170381) 0.92 0.78 (0.71–0.85) 57.1 0.40 (0.35–0.44) 236 (0–215209) 96.85
Mutation (PI3CA) 0.42 (0.04–0.93) 94.55 0.99 (0.85–1) 64.2 58.7 (2.8–1,237) 55.2 0.59 (0.18–1.95) 97.9 0.99 (0.98–1) 100 (3–3,328) 99
Study design
Retrospective 0.36 (0.27–0.46) 95.8 0.96 (0.93–0.98) 98.11 10 (5.3–18.7) 94.5 0.67 (0.58–0.76) 95.17 0.78 (0.74–0.82) 15 (18–29) 99
Prospective 0.25 (0.16–0.28) 89.2 0.94 (0.89–0.97) 32.3 4.4 (2.2–8.6) 0.0 0.79 (0.69–0.91) 77.6 0.86 (0.83–0.89) 6 (3–12) 98
NA 0.40 (0.34–0.47) 90 0.99 (0.97–1) 91.4 43.1 (14.1–137.1) 78.4 0.60 (0.54–0.67) 83.23 0.78 (0.74–0.82) 71 (23–218) 99
Sample type
Plasma 0.38 (0.29–0.48) 91.6 0.97 (0.92–0.99) 92.3 11.6 (5.2–25.8) 78.5 0.64 (0.56–0.75) 85.9 0.77 (0.74–0.81) 18 (8–41) 99
Serum 0.35 (0.30–0.41) 93.8 0.98 (0.96–0.99) 96.5 16.5 (9.1–30) 89.8 0.66 (0.61–0.72) 91.4 0.80 (0.77–0.84) 25 (13–46) 99
Pub. year
≥2010 0.39 (0.32–0.46) 94.8 0.96 (0.94–0.98) 97.4 11 (6.7–18.41) 93.6 0.63 (0.57–0.71) 94.4 0.78 (0.75–0.82) 17 (10–29) 99
<2010 0.32 (0.26–0.43) 89.6 0.99 (0.96–1) 82.2 43.4 (9.3–203.3) 29.7 0.68 (0.61–0.75) 77.3 0.80 (0.76–0.83) 64 (13–303) 97.1
Sample size
≥100 0.39 (0.32–0.46) 95.85 0.96 (0.92–0.98) 97.5 9.4 (5.2–7.16.8 92 0.64 (0.58–0.71) 93.2 0.73 (0.69–0.77) 15 (8–27) 99
<100 0.34 (0.27–0.41) 85.6 0.98 (0.96–0.99) 46.2 16.7 (8.9–31.4) 0 0.68 (0.61–0.75) 83.3 0.92 (0.82–0.94) 25 (13–48) 94.7

  1. SE, sensitivity; SP, specificity; PLR, positive likelihood ratio, NLR, negative Likelihood ratio; DOR, diagnostic odds ratio; PBC, primary breast cancer; MBC, metastatic breast cancer; RASSF1A, Ras association domain family member 1; GSTP1, glutathione S-transferase pi 1; p16, cyclin dependent kinase inhibitor 2A; BRCA1, BRCA1 DNA repair associated; RAR-β2, retinoic acid receptor beta 2; APC, adenomatous polyposis coli; NA, not available.

For qualitative assays, regarding the histological grade, the results show that cfDNA has a better diagnostic performance in patients with G3 than in patients with G1–G2 (DOR: 32, AUC: 0.73 vs. DOR: 7, AUC: 0.69, respectively). In those assays performed using plasma, DOR was obtained as 18 and AUC was 0.77, while the same estimations for those assays that were performed using serum, were obtained to be 25 and 0.80, respectively. The articles published before 2010 showed superior performance in diagnosing BC compared to the articles published after 2010 (DOR: 64, AUC: 0.80 vs. DOR: 17, AUC: 0.78, respectively). Additionally, the assays with a sample size <100 cases (DOR: 25, AUC: 0.92) indicated slightly better performance when compared to the assays with a sample size ≥100 cases (DOR: 15, AUC: 0.73). Amongst the methylated genes, APC, p16, and BRCA1 showed the highest DOR (as 236, 124, and 100, respectively). However, the amounts of AUC for these genes were relatively low (as 0.40, 0.38, and 0.47, respectively (Table 2).

For quantitative assays, the assays performed using plasma indicated better diagnostic performance compared to those assays that were performed using serum (DOR: 34, AUC: 0.92 vs. DOR: 24, AUC: 0.90, respectively). Regarding the factor of publication year, the articles published after 2010 showed better performance when compared to the articles published before 2010 (DOR: 32, AUC: 0.91 vs. DOR: 22, AUC: 0.90, respectively). Moreover, in this meta-analysis, the assays with a sample size <100 showed superior performance compared to the assays with a sample size ≥100 cases (DOR: 37, AUC: 0.92 vs. DOR: 22, AUC: 0.90, respectively) (Supplementary Material 1; Table 1).

Meta-regression

In this phase of the study, we implemented meta-regression to assess the source of heterogeneity based on the subgroup’s results. For qualitative assays, the association between several covariates, including ethnicity, sample type, study design, sample size, histological type, and assay indicator in 6 different models, and the diagnostic performance was evaluated (Table 3). Based on the available reported data, we used the following covariates to perform meta-regression for quantitative assays: ethnicity, sample type, article design, sample size, and assay indicator in 5 different models (Supplementary Material 1; Table 2). Finally, meta-regression results demonstrated that none of the aforementioned covariates is a significant source of heterogeneity neither in quantitative nor in qualitative assays.

Table 3:

Meta-regression for covariates in qualitative cell-free DNA assays (inverse variance weights).

Var Coeff. Std. Err. p-Value RDOR [95% CI]
Model 1
Cte. 2.225 0.6824 0.0016
S −0.117 0.0712 0.1032
Ethnicity −0.215 0.3474 0.5369 0.81 (0.40; 1.61)
Sample type −0.123 0.3556 0.7296 0.88 (0.44; 1.79)
Indicator 0.084 0.2264 0.7113 1.09 (0.69; 1.71)
Study design 0.096 0.1643 0.5618 1.10 (0.79; 1.53)
Histo. type −0.213 0.4229 0.6161 0.81 (0.35; 1.87)
Sample size −0.385 0.3387 0.2588 0.68 (0.35; 1.33)
Model 2
Cte. 1.726 0.5170 0.0012
S −0.137 0.0698 0.0520
Ethnicity −0.108 0.3376 0.7494 0.90 (0.46; 1.75)
Sample type 0.021 0.3361 0.9497 1.02 (0.52; 1.99)
Indicator 0.099 0.2277 0.6648 1.10 (0.70; 1.74)
Study design 0.117 0.1652 0.4801 1.12 (0.81; 1.56)
Histo. type −0.274 0.4258 0.5221 0.76 (0.33; 1.77)
Model 3
Cte. 1.806 0.5072 0.0006
S −0.132 0.0696 0.0613
Ethnicity −0.054 0.3295 0.8697 0.95 (0.49; 1.82)
Sample type −0.016 0.3330 0.9613 0.98 (0.51; 1.91)
Indicator 0.098 0.2281 0.6687 1.10 (0.70; 1.73)
Histo. type −0.341 0.4163 0.4148 0.71 (0.31; 1.63)
Model 4
Cte. 1.743 0.3601 0.0000
S −0.131 0.0688 0.0598
Sample type −0.025 0.3265 0.9403 0.98 (0.51; 1.87)
Indicator 0.090 0.2209 0.6855 1.09 (0.71; 1.70)
Histo. type −0.368 0.3824 0.3383 0.69 (0.32; 1.48)
Model 5
Cte. 1.734 0.3506 0.0000
S −0.132 0.0683 0.0567
Indicator 0.089 0.2196 0.6860 1.09 (0.71; 1.69)
Histo. type −0.377 0.3645 0.3040 0.69 (0.33; 1.41)
Model 6
Cte. 1.829 0.2562 0.0000
S −0.133 0.0681 0.0537
Histo. type −0.389 0.3622 0.2861 0.68 (0.33; 1.39)

  1. Var, covariate; coeff., coefficient; Std. Err., standard error; RDOR, relative diagnostic odds ratio; Cte., constante; S, threshold effect in the model.

Besides, analyzing the threshold effect using the Spearman’s correction coefficient implies that the threshold possibly is not a source of heterogeneity (coefficient: 0.099 and p=0.6 for quantitative assays and coefficient; −0.143 and p=0.056 for qualitative assays).

Sensitivity analysis by single study omission

Regardless of extensive subgroup analysis and meta-regression, the source of heterogeneity was not found. Therefore, to further investigate the source heterogeneity, a sensitivity analysis has been performed by single-study omission for SE and SP in both quantitative and qualitative assays (Supplementary Material 1; Figure 6).

Based on the sensitivity analysis, the following articles were found as potential sources of heterogeneity: SE in quantitative assays: Tangvarasittichai 2013 [80], Agostini 2012 [32], and Tang 2018 [79]; SP of quantitative assays: Agostini 2012 [32] and Stötzer 2014 [77]; SE of qualitative assays: Hoque 2006 [48], Kim 2010 [54], and Taback 2006 [78]; and SP of qualitative assays: Ebeid 2016 [42], Jing 2008 [51], and Jing 2010 [52]. Thereafter, all the main effect measures have been re-calculated after excluding the above-mentioned articles. Eliminating these studies has consequently led to a considerable decrease in heterogeneity in all measures of the main analysis, including SE, SP, PLR, NLR, and DOR (Table 4). Although no changes have been observed in their direction, a considerable drop has been observed in the amount of DOR in both qualitative (from 23 to 16.08) and quantitative assays (from 29 to 15.15). Therefore, to investigate how the above-mentioned articles affected the other analyses, all of them, including subgroups, clinical utility, meta-regression, and publication bias, have been re-performed after excluding those articles.

Table 4:

The main effect measures after deleting studies.

Main effect measures for qualitative assays Main effect measures for quantitative assays
Value 95% Confidence interval I 2, % Value 95% Confidence interval I 2, %
SE 0.37 0.32–0.43 73.37 SE 0.75 0.68–0.82 71.85
SP 0.96 0.94–0.98 75.87 SP 0.84 0.78–0.89 64.08
PLR 10.43 6.24–12.43 66.74 PLR 4.81 3.39–6.82 55.40
NLR 0.65 0.59–0.71 69.36 NLR 0.30 0.23–0.39 61.05
DOR 16.08 10.39–23.54 75.87 DOR 15.15 8.93–24.67 80
AUC 0.76 0.72–0.80 - AUC 0.87 0.84–0.90 -

After deleting the articles at the previous stage, heterogeneity significantly dropped in almost all the subgroups; however, according to Cochrane guidelines, the I 2 for some subgroups might still represent substantial heterogeneity (I 2: 50–90%) [87]. It should be noted that, the effect measures have changed in almost all the subgroups (Supplementary Material 2; Tables 1 and 2), but the drawn conclusions remained unchanged. As well, meta-regression analysis did not recognize any of the covariate models as a potential source of heterogeneity (Supplementary Material 2; Tables 3 and 4). In the Fagan’s nomograms, the intervals have slightly decreased between positive and negative post-test probabilities (Supplementary Material 2; Figures 1 and 2). Moreover, the result of Deek’s funnel plot asymmetry test were similar for both of them (Supplementary Material 2; Figures 1 and 2). Of note, the overall results obtained from likelihood ratio scattergrams were similar (Supplementary Material 2; Figure 3). Therefore, regardless of the considerable decrease in the heterogeneity, the main conclusions from all the analyses remained unchanged. Correspondingly, this implies that, the effect of the deleted studies on the final results was minimal, so the results can be considered as stable.

Clinical utility

For quantitative cfDNA assays, the Fagan’s nomogram demonstrated that when pretest probability was set at 25%, cfDNA increased the post-test probability of a positive value up to 69%, while there was 7% probability of cfDNA in BC patients with negative results (Figure 6B). With 50% pretest probability, the correct detection succeeding a positive report increased to 87%, while there was 19% chance of ignoring a patient with negative test result (Figure 6C). With 75% pretest probability, the likelihood of a true diagnosis succeeding a positive test value was calculated as 95%; however, the possibility of ignoring a patient with negative test result was 41% (Figure 6D).

Figure 6: 
Publication bias analysis and clinical utility for quantitative assays. (A) Deeks’ funnel plot asymmetry test for quantitative studies (B) Fagan monogram with 25% pretest probability (C) Fagan monogram with 50% pretest probability (D) Fagan monogram with 75% pretest probability.

Figure 6:

Publication bias analysis and clinical utility for quantitative assays. (A) Deeks’ funnel plot asymmetry test for quantitative studies (B) Fagan monogram with 25% pretest probability (C) Fagan monogram with 50% pretest probability (D) Fagan monogram with 75% pretest probability.

For qualitative cfDNA assays, with a pretest probability of 25%, the probability of a true detection succeeding a positive test report was estimated to be 83%; however, the chance of ignoring a BC case after a negative result was 18% (Figure 7B). With 50% pretest probability, the correct detection leading to a positive result, increased up to 94%, while there was 40% chance of ignoring a BC case with a negative result (Figure 7C). With 75% pretest probability, the true diagnosis leading to a positive result, increased to 98%; however, there was 66% chance of ignoring a patient with a negative test result (Figure 7D).

Figure 7: 
Publication bias analysis and clinical utility for qualitative assays. (A) Deeks’ funnel plot asymmetry test for qualitative studies (B) Fagan monogram with 25% pretest probability (C) Fagan monogram with 50% pretest probability (D) Fagan monogram with 75% pretest probability.

Figure 7:

Publication bias analysis and clinical utility for qualitative assays. (A) Deeks’ funnel plot asymmetry test for qualitative studies (B) Fagan monogram with 25% pretest probability (C) Fagan monogram with 50% pretest probability (D) Fagan monogram with 75% pretest probability.

To evaluate the clinical values of different assays, we used a likelihood ratio scattergram [88]. Accordingly, this matrix is divided into the following four quadrants: the right upper quadrant (RUQ) (positive likelihood ratio (PLR) >10, negative likelihood ratio (NLR) >0.1: confirmation only), the left upper quadrant (LUQ) (PLR >10, NLR <0.1: exclusion and confirmation), the right lower quadrant (RLQ) (PLR <10, NLR >0.1: no exclusion or confirmation), and the left lower quadrant (LLQ) (PLR <10, NLR <0.1: exclusion only) [88]. In the present study, the summary of both PLR and NLR for quantitative cf-DNA assays was plotted in the RUQ (Supplementary Material 1; Figure 6A). Moreover, for quantitative cfDNA assays, summary of both PLR and NLR was plotted in the RLQ (Supplementary Material 1; Figure 6B).

Publication bias

Deek’s funnel plot asymmetry test have been employed to investigate publication bias. A coefficient of 11.89 (95% CI: 5.33–18.45; p=0.001) implied significant study bias among studies that used a qualitative methodological approach in order to assess cfDNA (Figure 7A). Accordingly, trim and fill analysis was performed to correct funnel plot asymmetry resulted from publication bias. The pooled point estimate was 1.78 (95% CI: 1.46–2.11) prior to trimming the omitted studies. After conducting the test, the outcome changed to 0.904 (95% CI: 0.56–1.23) (Figure 8). For those studies with a quantitative methodological approach to assess cfDNA, Deek’s funnel plot test showed no publication bias with a coefficient of 2.84 (95% CI: −0.61 to 6.29; p=0.102) (Figure 6A).

Figure 8: 
Trim and fill analysis for qualitative studies.

Figure 8:

Trim and fill analysis for qualitative studies.

Discussion

Currently, BC screening through mammography was shown to reduce mortality up to 30% [3]. However, 13% of BC patients are missed by mammography results due to different factors [4], [5]. Additionally, false-positive detection in mammograms may cause over-treatment as well as radiation-induced disease [89]. Some common serum-based biomarkers such as cancer antigen CA15-3 and CEA, also did not show highly accurate result for the diagnosis of BC [9, 89]. Therefore, there is a need for finding novel noninvasive biomarkers to overcome the limitations resulted from conventional biomarkers and diagnostic approaches. The present updated meta-analysis aimed to address controversies in the previous published results regarding the diagnostic accuracy of cfDNA for BC.

In the present study, the pooled SE, SP, and sROC-AUC for quantitative approaches in 29 studies were obtained as 80%, 88%, and 0.91, respectively. In a meta-analysis conducted by Ling et al. [90] on 14 quantitative cfDNA studies for detecting BC, the pooled SE, SP, and AUC were estimated as 78%, 83%, and 0.93. The higher rates of SE and SP in our study might be resulted from the advancement in the detection methods, which have been employed by the extra studies enrolled in the data synthesis. As well, these findings are comparable to the accuracy of conventional serum-based biomarkers for the detection of BC such as CEA, CA15-3, and human epidermal growth factor receptor 2 (HER2). Notably, in this study, CA15-13, as a common marker, showed a SE, SP, and AUC of 73%, 85%, and 0.78, respectively [91]. Accordingly, in another study, the same estimations for HER2 have been evaluated to be 51%, 86%, and 0.65, respectively [92]. The valuation of CEA in the breast ductal secretions also revealed SE and AUC rates at 58% and 0.87 for detecting BC, respectively [93]. Lawicki et al. [94] in their study assessed the combined diagnostic accuracy of four markers, including macrophage colony-stimulating factor (M-CSF), matrix metallopeptidase-9 (MMP-9), TIMP metallopeptidase Inhibitor-1 (TIMP-1), and CA 15-3, for the BC detection. This panel consisting of four biomarkers yielded SE, SP, and AUC rates as 84%, 83%, and 0.91, respectively [94]. Therefore, quantitative cfDNA assays indicated a higher diagnostic performance when compared to the conventional serum-based biomarkers. However, few studies directly compared the accuracies of cfDNA and other serum based biomarkers.

For qualitative cfDNA assays, SE intensely reduced to 36%, and AUC reduced to 0.79, while SP enhanced up to 98%. Among qualitative assays, the evaluation of the methylation status of single genes was the most common assay. By considering SE, SP, and AUC rates as 36%, 97%, and 0.78, respectively, a slight difference was found between methylation assays and the pooled estimations for the overall qualitative assays. The main reason for the poor sensitivity rate of methylation analysis by cfDNA, regardless of high concordance with tumor tissue methylation status, can be considered to be the poor association of the selected genes with the pathogenesis of the BC. For example, in a meta-analysis done by Li et al. [95] on the accuracy of RASSF1A methylation using serum/plasma and tissue DNA for detecting BC, the overall SE, SP, and AUC were obtained as 49%, 95%, and 0.83, respectively. There was a slight difference in terms of SE, SP, and AUC between the sample types (serum vs. tissue), indicating a low association of RASSF1A BC [95]. Additionally, in the present study, the pooled SE for hyper-methylated RASSF1A, RAR-β2, p16, GSTP-1, BRCA1, and APC ranged from 19 to 34%; however, there was a specificity between 95 and 100%. The same results have been also observed for other qualitative approaches, including LOH, MI, and repeated element analysis. Regardless of low SE, it was shown that the excellent rate of SP for cfDNA methylation markers and other qualitative cfDNA assays can be used as auxiliary tools for enhancing the SP of the currently applied screening tools with lower SP in the risk assessment or the detection of BC.

DOR, as a single indicator of diagnostic performance, was also assessed in the current research. Generally, a DOR value of >10 is considered as good discriminatory performance [96], [97]. In the present study, the pooled DOR for quantitative and qualitative cfDNA assays was obtained as 29 and 23, respectively. Moreover, for quantitative approaches, plasma-based assays yielded a better diagnostic performance compared to serum-based cfDNA assays (DOR: 34 vs. DOR: 24).

To the best of our knowledge, in comparison to the previous works performed with similar purposes [90, 98], [99], [100], [101], [102], the present study for the first time enrolled an adequate number of studies to compare the diagnostic performance of qualitative assays statistically between low grade and high grade BC patients. In this study, we evaluated the accuracy of qualitative cfDNA assays in different subgroups based on the tumor stage (size), histological grade, and lymph node’s involvement. The results show that the qualitative assays can better differentiate patients at T3–T4 stages, with histological grade G3, and positive nodal involvement with higher accuracy (higher SE, SP, and DOR) compared to patients at T1–T2 stages, with histological grade G1–G2, and negative nodal involvement (Table 2). In other words, the qualitative cfDNA assays showed a better diagnostic performance in patients with advanced BC than patients with lower stages of BC. Based on this, these assays do not represent a remarkable clinical significance as a biomarker for early detection of BC [103]. These findings are of great importance as they suggest that the previous conclusion regarding the potential application of overall cfDNA assays as screening tools for early detection of BC should be interpreted with cautious [90]. However, it should be noted that the same analysis for quantitative assays was not possible due to lack of adequate data reported in the enrolled studies. Therefore, future studies by focusing on the evaluation of the diagnostic accuracy of quantitative cfDNA assays in BC patients at early and late stages are vital.

In the present study, another evaluated diagnostic estimation was likelihood ratio. PLR and NLR merge both SE and SP, so they are more clinically applicable than SE and SP. As a general criterion, PLR of >10 and NLR of <0.1 have been accepted as high accuracy performances [104]. Quantitative assays did not fulfill this criterion neither in overall nor in subgroup analysis. On the other side, qualitative assays showed an acceptable PLR (Overall: 14.94), but same as quantitative assays, they achieved no satisfying NLR. To provide a more clinically applicable view, likelihood scattergrams have been plotted using both PLR and NLR values [105]. The likelihood scattergram for quantitative assays clearly suggested that these assays are not robust enough neither for BC confirmation nor for exclusion. The likelihood scattergram for qualitative assays indicated that these assays have a satisfying performance only for confirming the diagnosis of BC, but not for its exclusion.

There are some concerns on publication bias in qualitative assays. Results may be biased, since studies with greater sample size or positive results have higher chance of publishing. Nevertheless, after performing trim and fill analysis, no significant change has been observed. Besides, no publication bias was observed among quantitative assays based on the Deek’s funnel plot asymmetry test. In this regard, meta-regression results suggested that none of the aforesaid covariates is a significant source of heterogeneity neither in quantitative nor in qualitative assays.

The major limitations of this study were as follows: regardless of comprehensive literature search, the number of quantitative cfDNA assays were small, so for more robust results, further studies must be done in the future. Additionally, we did not perform subgroup analysis regarding tumor stage, histological type, and nodal status in the included studies for quantitative assays, due to lack of enough data reported in the enrolled studies. Regardless of meta-regression in different models, in depth analysis of heterogeneity was not possible, since most of the included studies did not provide any proper information on age of participants, blinding methods, consecutive or successive sampling, and specimen preparation protocols. As well, lack of standardization of the detection techniques from sample acquisition to different cutoff values and interpretation of the results, may lead to hidden bias in the pooled estimations and a significant observed heterogeneity. Furthermore, a restriction for including English language literature may have introduced bias to the pooled analysis. Therefore, further larger prospective analyses with a focus on the standardized protocols either in technical aspects or in reporting criteria in the future are indispensable for producing the robust results and primary clinical considerations.

Conclusions

In the present study, we evaluated the diagnostic performance of both quantitative and qualitative cfDNA assays for discriminating patients with breast cancer from noncancerous/healthy controls. Correspondingly, the obtained key findings are summarized as follows:

  • – The diagnostic performance of quantitative approaches in terms of SE, SP, sROC-AUC, and DOR were comparable to those of the conventional screening tools and serum-based biomarkers.

  • – The poor PLR and NLR values of quantitative assays suggested that these assays are not robust enough neither for BC confirmation nor for its exclusion.

  • – Qualitative cfDNA assays indicated poor sensitivity and also high value of NLR.

  • – The SP of qualitative assays were promising and these assays showed a satisfying performance only for confirming the diagnosis of BC, but not for its exclusion.

  • – Qualitative assays showed superior accuracy in distinguishing patients at T3–T4 stages, with histological grade G3, and positive nodal involvement compared to patients at T1–T2 stages, with G1–G2, and negative nodal involvement when compared to normal controls. Accordingly, this does not have a remarkable clinical significance, as a biomarker for early detection, since most patients at the advanced stage of cancer can be well-diagnosed using the existing tools.

Performing consistent prospective studies in the future is needed to evaluate diagnostic performance of quantitative cfDNA assays in combination with conventional biomarkers. As well, incorporating qualitative cfDNA approaches along with a strong association with BC pathogenesis are mandatory to address the poor SE of qualitative assays.


Corresponding author: Qingfeng Guo, Department of General Surgery, Affiliated Hospital of Jiangnan University (Original Area of Wuxi No. 3 People’s Hospital), Wuxi, 210000, P.R. China, E-mail:

  1. Research funding: This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

  2. Author contributions: All authors have accepted responsibility for the entire content of this manuscript and approved its submission.

  3. Competing interests: Authors state no conflict of interest.

  4. Informed consent: Not applicable.

  5. Ethical approval: Not applicable.

  6. Data availability: The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.

References

1. Bray, F, Ferlay, J, Soerjomataram, I, Siegel, RL, Torre, LA, Jemal, A. Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. Cancer J Clin 2018;68:394–424. https://doi.org/10.3322/caac.21492. Search in Google Scholar

2. Böhm, D, Keller, K, Wehrwein, N, Lebrecht, A, Schmidt, M, Kölbl, H, et al.. Serum proteome profiling of primary breast cancer indicates a specific biomarker profile. Oncol Rep 2011;26:1051–6. https://doi.org/10.3892/or.2011.1420. Search in Google Scholar

3. Yu, D, Tong, Y, Guo, X, Feng, L, Jiang, Z, Ying, S, et al.. Diagnostic value of concentration of circulating cell-free DNA in breast cancer: a meta-analysis. Front Oncol 2019;9:95. https://doi.org/10.3389/fonc.2019.00095. Search in Google Scholar

4. Oeffinger, KC, Fontham, ET, Etzioni, R, Herzig, A, Michaelson, JS, Shih, YC, et al.. Breast cancer screening for women at average risk: 2015 guideline update from the American Cancer Society. Jama 2015;314:1599–614. https://doi.org/10.1001/jama.2015.12783. Search in Google Scholar

5. Lehman, CD, Arao, RF, Sprague, BL, Lee, JM, Buist, DS, Kerlikowske, K, et al.. National performance benchmarks for modern screening digital mammography: update from the breast cancer surveillance consortium. Radiol 2017;283:49–58. https://doi.org/10.1148/radiol.2016161174. Search in Google Scholar

6. Olsen, O, Gøtzsche, PC. Screening for breast cancer with mammography. Cochrane Database Syst Rev 2001;4:Cd001877. https://doi.org/10.1002/14651858.CD001877. Search in Google Scholar

7. Khan, SA, Reddy, D, Gupta, S. Global histone post-translational modifications and cancer: biomarkers for diagnosis, prognosis and treatment? World J Biol Chem 2015;6:333–45. https://doi.org/10.4331/wjbc.v6.i4.333. Search in Google Scholar

8. Lumachi, F, Norberto, L, Marino, F, Basso, SM, Marzano, B, Chiara, GB. Hormone receptor rate, MIB-1 score and serum tumour markers CEA and CA 15-3 relationship in elderly women with pT1-2 breast cancer. Anticancer Res 2010;30:4701–4. Search in Google Scholar

9. Molina, R, Auge, JM, Farrus, B, Zanón, G, Pahisa, J, Muñoz, M, et al.. Prospective evaluation of carcinoembryonic antigen (CEA) and carbohydrate antigen 15.3 (CA 15.3) in patients with primary locoregional breast cancer. Clin Chem 2010;56:1148–57. https://doi.org/10.1373/clinchem.2009.135566. Search in Google Scholar

10. Volckmar, AL, Sültmann, H, Riediger, A, Fioretos, T, Schirmacher, P, Endris, V, et al.. A field guide for cancer diagnostics using cell-free DNA: from principles to practice and clinical applications. Genes Chromosom Cancer 2018;57:123–39. https://doi.org/10.1002/gcc.22517. Search in Google Scholar

11. Diaz, LAJr., Bardelli, A. Liquid biopsies: genotyping circulating tumor DNA. J Clin Oncol 2014;32:579–86. https://doi.org/10.1200/jco.2012.45.2011. Search in Google Scholar

12. Hao, TB, Shi, W, Shen, XJ, Qi, J, Wu, XH, Wu, Y, et al.. Circulating cell-free DNA in serum as a biomarker for diagnosis and prognostic prediction of colorectal cancer. Br J Canc 2014;111:1482–9. https://doi.org/10.1038/bjc.2014.470. Search in Google Scholar

13. Zonta, E, Nizard, P, Taly, V. Assessment of DNA integrity, applications for cancer research. Adv Clin Chem 2015;70:197–246. https://doi.org/10.1016/bs.acc.2015.03.002. Search in Google Scholar

14. Wang, R, Li, X, Zhang, H, Wang, K, He, J. Cell-free circulating tumor DNA analysis for breast cancer and its clinical utilization as a biomarker. Oncotarget 2017;8:75742–55. https://doi.org/10.18632/oncotarget.20608. Search in Google Scholar

15. Salvi, S, Gurioli, G, De Giorgi, U, Conteduca, V, Tedaldi, G, Calistri, D, et al.. Cell-free DNA as a diagnostic marker for cancer: current insights. Onco Targets Ther 2016;9:6549–59. https://doi.org/10.2147/ott.s100901. Search in Google Scholar

16. Bettegowda, C, Sausen, M, Leary, RJ, Kinde, I, Wang, Y, Agrawal, N, et al.. Detection of circulating tumor DNA in early- and late-stage human malignancies. Sci Transl Med 2014;6:224ra24. https://doi.org/10.1126/scitranslmed.3007094. Search in Google Scholar

17. Dawson, SJ, Rueda, OM, Aparicio, S, Caldas, C. A new genome-driven integrated classification of breast cancer and its implications. EMBO J 2013;32:617–28. https://doi.org/10.1038/emboj.2013.19. Search in Google Scholar

18. Forshew, T, Murtaza, M, Parkinson, C, Gale, D, Tsui, DW, Kaper, F, et al.. Noninvasive identification and monitoring of cancer mutations by targeted deep sequencing of plasma DNA. Sci Transl Med 2012;4:136ra68. https://doi.org/10.1126/scitranslmed.3003726. Search in Google Scholar

19. Garcia-Murillas, I, Schiavon, G, Weigelt, B, Ng, C, Hrebien, S, Cutts, RJ, et al.. Mutation tracking in circulating tumor DNA predicts relapse in early breast cancer. Sci Transl Med 2015;7:302ra133. https://doi.org/10.1126/scitranslmed.aab0021. Search in Google Scholar

20. Beaver, JA, Jelovac, D, Balukrishna, S, Cochran, R, Croessmann, S, Zabransky, DJ, et al.. Detection of cancer DNA in plasma of patients with early-stage breast cancer. Clin Canc Res 2014;20:2643–50. https://doi.org/10.1158/1078-0432.ccr-13-2933. Search in Google Scholar

21. Oshiro, C, Kagara, N, Naoi, Y, Shimoda, M, Shimomura, A, Maruyama, N, et al.. PIK3CA mutations in serum DNA are predictive of recurrence in primary breast cancer patients. Breast Canc Res Treat 2015;150:299–307. https://doi.org/10.1007/s10549-015-3322-6. Search in Google Scholar

22. Leary, RJ, Sausen, M, Kinde, I, Papadopoulos, N, Carpten, JD, Craig, D, et al.. Detection of chromosomal alterations in the circulation of cancer patients with whole-genome sequencing. Sci Transl Med 2012;4:162ra154. https://doi.org/10.1126/scitranslmed.3004742. Search in Google Scholar

23. Liberati, A, Altman, DG, Tetzlaff, J, Mulrow, C, Gøtzsche, PC, Ioannidis, JPA, et al.. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate healthcare interventions: explanation and elaboration. BMJ 2009;339:b2700. https://doi.org/10.1136/bmj.b2700. Search in Google Scholar

24. Whiting, PF, Rutjes, AW, Westwood, ME, Mallett, S, Deeks, JJ, Reitsma, JB, et al.. QUADAS-2: a revised tool for the quality assessment of diagnostic accuracy studies. Ann Intern Med 2011;155:529–36. https://doi.org/10.7326/0003-4819-155-8-201110180-00009. Search in Google Scholar

25. Belias, M, Rovers, MM, Reitsma, JB, Debray, TPA, IntHout, J. Statistical approaches to identify subgroups in meta-analysis of individual participant data: a simulation study. BMC Med Res Methodol 2019;19:183. https://doi.org/10.1186/s12874-019-0817-6. Search in Google Scholar

26. Akobeng, AK. Understanding diagnostic tests 2: likelihood ratios, pre- and post-test probabilities and their use in clinical practice. Acta Paediatr 2007;96:487–91. https://doi.org/10.1111/j.1651-2227.2006.00179.x. Search in Google Scholar

27. Deeks, JJ, Macaskill, P, Irwig, L. The performance of tests of publication bias and other sample size effects in systematic reviews of diagnostic test accuracy was assessed. J Clin Epidemiol 2005;58:882–93. https://doi.org/10.1016/j.jclinepi.2005.01.016. Search in Google Scholar

28. Duval, S, Tweedie, R. Trim and fill: a simple funnel‐plot-based method of testing and adjusting for publication bias in meta‐analysis. Biometrics 2000;56:455–63. https://doi.org/10.1111/j.0006-341x.2000.00455.x. Search in Google Scholar

29. Zhang, M, He, Y, Zhang, X, Zhang, M, Kong, L. A pooled analysis of the diagnostic efficacy of plasmic methylated septin-9 as a novel biomarker for colorectal cancer. Biomed Rep 2017;7:353–60. https://doi.org/10.3892/br.2017.970. Search in Google Scholar

30. Myers, L, Sirois, MJ. Spearman Rank correlation coefficient. The concise encyclopedia of statistics. New York, NY: Springer New York; 2008:502–5 pp. Search in Google Scholar

31. Agassi, R, Czeiger, D, Shaked, G, Avriel, A, Sheynin, J, Lavrenkov, K, et al.. Measurement of circulating cell-free DNA levels by a simple fluorescent test in patients with breast cancer. Am J Clin Pathol 2015;143:18–24. https://doi.org/10.1309/ajcpi5yhg0ogfahm. Search in Google Scholar

32. Agostini, M, Enzo, M, Bedin, C, Belardinelli, V, Goldin, E, Del Bianco, P, et al.. Circulating cell-free DNA: a promising marker of regional lymphonode metastasis in breast cancer patients. Canc Biomarkers 2012;11:89–98. https://doi.org/10.3233/cbm-2012-0263. Search in Google Scholar

33. Beaver, JA, Jelovac, D, Balukrishna, S, Cochran, RL, Croessmann, S, Zabransky, DJ, et al.. Detection of cancer DNA in plasma of patients with early-stage breast cancer. Clin Canc Res 2014;20:2643. https://doi.org/10.1158/1078-0432.ccr-13-2933. Search in Google Scholar

34. Beck, J, Urnovitz, HB, Mitchell, WM, Schütz, E. Next generation sequencing of serum circulating nucleic acids from patients with invasive ductal breast cancer reveals differences to healthy and nonmalignant controls. Mol Canc Res 2010;8:335–42. https://doi.org/10.1158/1541-7786.mcr-09-0314. Search in Google Scholar

35. Board, RE, Wardley, AM, Dixon, JM, Armstrong, AC, Howell, S, Renshaw, L, et al.. Detection of PIK3CA mutations in circulating free DNA in patients with breast cancer. Breast Canc Res Treat 2010;120:461–7. https://doi.org/10.1007/s10549-010-0747-9. Search in Google Scholar

36. Catarino, R, Ferreira, MM, Rodrigues, H, Coelho, A, Nogal, A, Sousa, A, et al.. Quantification of free circulating tumor DNA as a diagnostic marker for breast cancer. DNA Cell Biol 2008;27:415–21. https://doi.org/10.1089/dna.2008.0744. Search in Google Scholar

37. Chen, X, Bonnefoi, H, Diebold-Berger, S, Lyautey, J, Lederrey, C, Faltin-Traub, E, et al.. Detecting tumor-related alterations in plasma or serum DNA of patients diagnosed with breast cancer. Clin Canc Res 1999;5:2297. Search in Google Scholar

38. Chimonidou, M, Strati, A, Malamos, N, Georgoulias, V, Lianidou, ES. SOX17 promoter methylation in circulating tumor cells and matched cell-free DNA isolated from plasma of patients with breast cancer. Clin Chem 2013;59:270–9. https://doi.org/10.1373/clinchem.2012.191551. Search in Google Scholar

39. Chimonidou, M, Tzitzira, A, Strati, A, Sotiropoulou, G, Sfikas, C, Malamos, N, et al.. CST6 promoter methylation in circulating cell-free DNA of breast cancer patients. Clin Biochem 2013;46:235–40. https://doi.org/10.1016/j.clinbiochem.2012.09.015. Search in Google Scholar

40. Divella, R, Tommasi, S, Lacalamita, R, Daniele, A, Abbate, I, Garrisi, VM, et al.. Circulating hTERT DNA in early breast cancer. Anticancer Res 2009;29:2845–9. Search in Google Scholar

41. Dulaimi, E, Hillinck, J, de Caceres, II, Al-Saleem, T, Cairns, P. Tumor suppressor gene promoter hypermethylation in serum of breast cancer patients. Clin Canc Res 2004;10:6189–93. https://doi.org/10.1158/1078-0432.ccr-04-0597. Search in Google Scholar

42. Ebeid, SA, Abd El Moneim, NA, Hewala, TI, Hemida, MA, Shehata, G, El-Taher, N. Assessment of hypermethylation of RASSF1A and protocadherin-10 tumor suppressor genes in breast cancer females: a six-year disease-free survival case-control study. MEJC 2016;7:9–20. Search in Google Scholar

43. Fu, D, Ren, C, Tan, H, Wei, J, Zhu, Y, He, C, et al.. Sox17 promoter methylation in plasma DNA is associated with poor survival and can be used as a prognostic factor in breast cancer. Medicine 2015;94:e637. https://doi.org/10.1097/md.0000000000000637. Search in Google Scholar

44. Gal, S, Fidler, C, Lo, Y, Taylor, M, Han, C, Moore, J, et al.. Quantitation of circulating DNA in the serum of breast cancer patients by real-time PCR. Br J Canc 2004;90:1211–5. https://doi.org/10.1038/sj.bjc.6601609. Search in Google Scholar

45. Geng, Y, Gao, Y, Ju, H, Yan, F. Diagnostic and prognostic value of plasma and tissue ubiquitin‐like, containing PHD and RING finger domains 1 in breast cancer patients. Canc Sci 2013;104:194–9. https://doi.org/10.1111/cas.12052. Search in Google Scholar

46. Gong, B, Xue, J, Yu, J, Li, H, Hu, H, Yen, H, et al.. Cell-free DNA in blood is a potential diagnostic biomarker of breast cancer. Oncol Lett 2012;3:897–900. https://doi.org/10.3892/ol.2012.576. Search in Google Scholar

47. Hashad, D, Sorour, A, Ghazal, A, Talaat, I. Free circulating tumor DNA as a diagnostic marker for breast cancer. J Clin Lab Anal 2012;26:467–72. https://doi.org/10.1002/jcla.21548. Search in Google Scholar

48. Hoque, MO, Feng, Q, Toure, P, Dem, A, Critchlow, CW, Hawes, SE, et al.. Detection of aberrant methylation of four genes in plasma DNA for the detection of breast cancer. J Clin Oncol 2006;24:4262–9. https://doi.org/10.1200/jco.2005.01.3516. Search in Google Scholar

49. Huang, ZH, Li, LH, Hua, D. Quantitative analysis of plasma circulating DNA at diagnosis and during follow-up of breast cancer patients. Canc Lett 2006;243:64–70. https://doi.org/10.1016/j.canlet.2005.11.027. Search in Google Scholar

50. Hussein, NA, Mohamed, SN, Ahmed, MA. Plasma ALU-247, ALU-115, and cfDNA integrity as diagnostic and prognostic biomarkers for breast cancer. Appl Biochem Biotechnol 2019;187:1028–45. https://doi.org/10.1007/s12010-018-2858-4. Search in Google Scholar

51. Jing, F, Jun, L, Yong, Z, Wang, Y, Fei, X, Zhang, J, et al.. Multigene methylation in serum of sporadic Chinese female breast cancer patients as a prognostic biomarker. Oncology 2008;75:60–6. https://doi.org/10.1159/000155145. Search in Google Scholar

52. Jing, F, Yuping, W, Yong, C, Jie, L, Jun, L, Xuanbing, T, et al.. CpG island methylator phenotype of multigene in serum of sporadic breast carcinoma. Tumor Biol 2010;31:321–31. https://doi.org/10.1007/s13277-010-0040-x. Search in Google Scholar

53. Jing, F, Zhang, J, Tao, J, Zhou, Y, Jun, L, Tang, X, et al.. Hypermethylation of tumor suppressor genes BRCA1, p16 and 14-3-3s in serum of sporadic breast cancer patients. Oncol Res Treat 2007;30:14–9. https://doi.org/10.1159/000096892. Search in Google Scholar

54. Kim, J-H, Shin, M-H, Kweon, S-S, Park, MH, Yoon, JH, Lee, JS, et al.. Evaluation of promoter hypermethylation detection in serum as a diagnostic tool for breast carcinoma in Korean women. Gynecol Oncol 2010;118:176–81. https://doi.org/10.1016/j.ygyno.2010.04.016. Search in Google Scholar

55. Kirkizlar, E, Zimmermann, B, Constantin, T, Swenerton, R, Hoang, B, Wayham, N, et al.. Detection of clonal and subclonal copy-number variants in cell-free DNA from patients with breast cancer using a massively multiplexed PCR methodology. Transl Oncol 2015;8:407–16. https://doi.org/10.1016/j.tranon.2015.08.004. Search in Google Scholar

56. Leon, S, Shapiro, B, Sklaroff, D, Yaros, M. Free DNA in the serum of cancer patients and the effect of therapy. Canc Res 1977;37:646–50. Search in Google Scholar

57. Li, D, Li, P, Wu, J, Yi, J, Dou, Y, Guo, X, et al.. Methylation of NBPF1 as a novel marker for the detection of plasma cell-free DNA of breast cancer patients. Clin Chim Acta 2018;484:81–6. https://doi.org/10.1016/j.cca.2018.05.030. Search in Google Scholar

58. Li, Z, Guo, X, Tang, L, Peng, L, Chen, M, Luo, X, et al.. Methylation analysis of plasma cell-free DNA for breast cancer early detection using bisulfite next-generation sequencing. Tumor Biol 2016;37:13111–9. https://doi.org/10.1007/s13277-016-5190-z. Search in Google Scholar

59. Mahmoud, EH, Fawzy, A, Ahmad, OK, Ali, AM. Plasma circulating cell-free nuclear and mitochondrial DNA as potential biomarkers in the peripheral blood of breast cancer patients. Asian Pac J Canc Prev 2015;16:8299–305. https://doi.org/10.7314/apjcp.2015.16.18.8299. Search in Google Scholar

60. Martínez-Galán, J, Torres, B, Del Moral, R, Muñoz-Gámez, JA, Martín-Oliva, D, Villalobos, M, et al.. Quantitative detection of methylated ESR1 and 14-3-3-sigma gene promoters in serum as candidate biomarkers for diagnosis of breast cancer and evaluation of treatment efficacy. Canc Biol Ther 2008;7:958–65. https://doi.org/10.4161/cbt.7.6.5966. Search in Google Scholar

61. Matuschek, C, Bölke, E, Lammering, G, Gerber, PA, Peiper, M, Budach, W, et al.. Methylated APC and GSTP1 genes in serum DNA correlate with the presence of circulating blood tumor cells and are associated with a more aggressive and advanced breast cancer disease. Eur J Med Res 2010;15:277–86. https://doi.org/10.1186/2047-783x-15-7-277. Search in Google Scholar

62. Mirza, S, Sharma, G, Parshad, R, Srivastava, A, Gupta, SD, Ralhan, R. Clinical significance of Stratifin, ERalpha and PR promoter methylation in tumor and serum DNA in Indian breast cancer patients. Clin Biochem 2010;43:380–6. https://doi.org/10.1016/j.clinbiochem.2009.11.016. Search in Google Scholar

63. Mirza, S, Sharma, G, Prasad, CP, Parshad, R, Srivastava, A, Gupta, SD, et al.. Promoter hypermethylation of TMS1, BRCA1, ERalpha and PRB in serum and tumor DNA of invasive ductal breast carcinoma patients. Life Sci 2007;81:280–7. https://doi.org/10.1016/j.lfs.2007.05.012. Search in Google Scholar

64. Müller, HM, Widschwendter, A, Fiegl, H, Ivarsson, L, Goebel, G, Perkmann, E, et al.. DNA methylation in serum of breast cancer patients: an independent prognostic marker. Canc Res 2003;63:7641–5. Search in Google Scholar

65. Ng, EK, Leung, CP, Shin, VY, Wong, CL, Ma, ES, Jin, HC, et al.. Quantitative analysis and diagnostic significance of methylated SLC19A3 DNA in the plasma of breast and gastric cancer patients. PLoS One 2011;6:e22233. https://doi.org/10.1371/journal.pone.0022233. Search in Google Scholar

66. Nunes, SP, Moreira-Barbosa, C, Salta, S, Palma de Sousa, S, Pousa, I, Oliveira, J, et al.. Cell-free DNA methylation of selected genes allows for early detection of the major cancers in women. Cancers (Basel) 2018;10:357. https://doi.org/10.3390/cancers10100357. Search in Google Scholar

67. Panagopoulou, M, Karaglani, M, Balgkouranidou, I, Biziota, E, Koukaki, T, Karamitrousis, E, et al.. Circulating cell-free DNA in breast cancer: size profiling, levels, and methylation patterns lead to prognostic and predictive classifiers. Oncogene 2019;38:3387–401. https://doi.org/10.1038/s41388-018-0660-y. Search in Google Scholar

68. Papadopoulou, E, Davilas, E, Sotiriou, V, Georgakopoulos, E, Georgakopoulou, S, Koliopanos, A, et al.. Cell-free DNA and RNA in plasma as a new molecular marker for prostate and breast cancer. Ann N Y Acad Sci 2006;1075:235–43. https://doi.org/10.1196/annals.1368.032. Search in Google Scholar

69. Rykova, EY, Skvortsova, TE, Laktionov, PP, Tamkovich, SN, Bryzgunova, OE, Starikov, AV, et al.. Investigation of tumor-derived extracellular DNA in blood of cancer patients by methylation-specific PCR. Nucleos Nucleot Nucleic Acids 2004;23:855–9. https://doi.org/10.1081/ncn-200026031. Search in Google Scholar

70. Schwarzenbach, H, Müller, V, Milde-Langosch, K, Steinbach, B, Pantel, K. Evaluation of cell-free tumour DNA and RNA in patients with breast cancer and benign breast disease. Mol Biosyst 2011;7:2848–54. https://doi.org/10.1039/c1mb05197k. Search in Google Scholar

71. Shan, M, Yin, H, Li, J, Li, X, Wang, D, Su, Y, et al.. Detection of aberrant methylation of a six-gene panel in serum DNA for diagnosis of breast cancer. Oncotarget 2016;7:18485–94. https://doi.org/10.18632/oncotarget.7608. Search in Google Scholar

72. Sharma, G, Mirza, S, Parshad, R, Srivastava, A, Gupta, SD, Pandya, P, et al.. Clinical significance of promoter hypermethylation of DNA repair genes in tumor and serum DNA in invasive ductal breast carcinoma patients. Life Sci 2010;87:83–91. https://doi.org/10.1016/j.lfs.2010.05.001. Search in Google Scholar

73. Sharma, G, Mirza, S, Prasad, CP, Srivastava, A, Gupta, SD, Ralhan, R. Promoter hypermethylation of p16INK4A, p14ARF, CyclinD2 and Slit2 in serum and tumor DNA from breast cancer patients. Life Sci 2007;80:1873–81. https://doi.org/10.1016/j.lfs.2007.02.026. Search in Google Scholar

74. Shaw, JA, Smith, BM, Walsh, T, Johnson, S, Primrose, L, Slade, MJ, et al.. Microsatellite alterations plasma DNA of primary breast cancer patients. Clin Canc Res 2000;6:1119–24. Search in Google Scholar

75. Shukla, S, Mirza, S, Sharma, G, Parshad, R, Gupta, SD, Ralhan, R. Detection of RASSF1A and RARbeta hypermethylation in serum DNA from breast cancer patients. Epigenetics 2006;1:88–93. https://doi.org/10.4161/epi.1.2.2679. Search in Google Scholar

76. Skvortsova, TE, Rykova, EY, Tamkovich, SN, Bryzgunova, OE, Starikov, AV, Kuznetsova, NP, et al.. Cell-free and cell-bound circulating DNA in breast tumours: DNA quantification and analysis of tumour-related gene methylation. Br J Canc 2006;94:1492–5. https://doi.org/10.1038/sj.bjc.6603117. Search in Google Scholar

77. Stötzer, OJ, Lehner, J, Fersching-Gierlich, D, Nagel, D, Holdenrieder, S. Diagnostic relevance of plasma DNA and DNA integrity for breast cancer. Tumor Biol 2014;35:1183–91. https://doi.org/10.1007/s13277-013-1158-4. Search in Google Scholar

78. Taback, B, Giuliano, AE, Lai, R, Hansen, N, Singer, FR, Pantel, K, et al.. Epigenetic analysis of body fluids and tumor tissues: application of a comprehensive molecular assessment for early-stage breast cancer patients. Ann N Y Acad Sci 2006;1075:211–21. https://doi.org/10.1196/annals.1368.029. Search in Google Scholar

79. Tang, Z, Li, L, Shen, L, Shen, X, Ju, S, Cong, H. Diagnostic value of serum concentration and integrity of circulating cell-free DNA in breast cancer: a comparative study with CEA and CA15-3. Lab Med 2018;49:323–8. https://doi.org/10.1093/labmed/lmy019. Search in Google Scholar

80. Tangvarasittichai, O, Jaiwang, W, Tangvarasittichai, S. The plasma DNA concentration as a potential breast cancer screening marker. Indian J Clin Biochem 2015;30:55–8. https://doi.org/10.1007/s12291-013-0407-z. Search in Google Scholar

81. Umetani, N, Giuliano, AE, Hiramatsu, SH, Amersi, F, Nakagawa, T, Martino, S, et al.. Prediction of breast tumor progression by integrity of free circulating DNA in serum. J Clin Oncol 2006;24:4270–6. https://doi.org/10.1200/jco.2006.05.9493. Search in Google Scholar

82. Wang, W, Liang, M, Ma, G, Li, L, Zhou, W, Xia, T, et al.. Plasma cell-free DNA integrity plus circulating tumor cells: a potential biomarker of no distant metastasis breast cancer. Neoplasma 2017;64:611–8. https://doi.org/10.4149/neo_2017_417. Search in Google Scholar

83. Wu, X, Tanaka, H. Aberrant reduction of telomere repetitive sequences in plasma cell-free DNA for early breast cancer detection. Oncotarget 2015;6:29795–807. https://doi.org/10.18632/oncotarget.5083. Search in Google Scholar

84. Yamamoto, N, Nakayama, T, Kajita, M, Miyake, T, Iwamoto, T, Kim, SJ, et al.. Detection of aberrant promoter methylation of GSTP1, RASSF1A, and RARβ2 in serum DNA of patients with breast cancer by a newly established one-step methylation-specific PCR assay. Breast Canc Res Treat 2012;132:165–73, https://doi.org/10.1007/s10549-011-1575-2. Search in Google Scholar

85. Zachariah, RR, Schmid, S, Buerki, N, Radpour, R, Holzgreve, W, Zhong, X. Levels of circulating cell-free nuclear and mitochondrial DNA in benign and malignant ovarian tumors. Obstet Gynecol 2008;112:843–50, https://doi.org/10.1097/aog.0b013e3181867bc0. Search in Google Scholar

86. Zhang, L, Song, X, Wang, X, Xie, Y, Wang, Z, Xu, Y, et al.. Circulating DNA of HOTAIR in serum is a novel biomarker for breast cancer. Breast Canc Res Treat 2015;152:199–208, https://doi.org/10.1007/s10549-015-3431-2. Search in Google Scholar

87. Higgins, JPT, Thomas, J, Chandler, J, Cumpston, M, Li, T, Page, MJ, et al., editors. Cochrane handbook for systematic reviews of interventions, 2nd ed. Chichester (UK): John Wiley & Sons; 2019. Search in Google Scholar

88. Yang, H, Wang, H, Zhang, C, Tong, Z. The accuracy of microRNA-210 in diagnosing lung cancer: a systematic review and meta-analysis. Oncotarget 2016;7:63283. https://doi.org/10.18632/oncotarget.11446. Search in Google Scholar

89. Kim, SJ, Chung, SY, Chang, JM, Cho, N, Han, W, Moon, WK. Ultrasound screening of contralateral breast after surgery for breast cancer. Eur J Radiol 2015;84:54–60. https://doi.org/10.1016/j.ejrad.2014.09.019. Search in Google Scholar

90. Lin, Z, Neiswender, J, Fang, B, Ma, X, Zhang, J, Hu, X. Value of circulating cell-free DNA analysis as a diagnostic tool for breast cancer: a meta-analysis. Oncotarget 2017;8:26625. https://doi.org/10.18632/oncotarget.15775. Search in Google Scholar

91. Thriveni, K, Deshmane, V, Ramaswamy, G, Krishnamoorthy, L. Diagnostic significance of CA15-3 with combination of HER-2/neu values at 85th percentiles in breast cancer. Indian J Clin Biochem 2013;28:136–40. https://doi.org/10.1007/s12291-012-0288-6. Search in Google Scholar

92. Leyland-Jones, B, Smith, BR. Serum HER2 testing in patients with HER2-positive breast cancer: the death knell tolls. Lancet Oncol 2011;12:286–95. https://doi.org/10.1016/s1470-2045(10)70297-7. Search in Google Scholar

93. Tang, S, Zhou, F, Sun, Y, Wei, L, Zhu, S, Yang, R, et al.. CEA in breast ductal secretions as a promising biomarker for the diagnosis of breast cancer: a systematic review and meta-analysis. Breast Canc 2016;23:813–9. https://doi.org/10.1007/s12282-016-0680-9. Search in Google Scholar

94. Ławicki, S, Głażewska, EK, Sobolewska, M, Będkowska, GE, Szmitkowski, M. Plasma levels and diagnostic utility of macrophage colony-stimulating factor, matrix metalloproteinase-9, and tissue inhibitor of metalloproteinases-1 as new biomarkers of breast cancer. Ann Lab Med 2016;36:223–9. https://doi.org/10.3343/alm.2016.36.3.223. Search in Google Scholar

95. Li, M, Wang, C, Yu, B, Zhang, X, Shi, F, Liu, X. Diagnostic value of RASSF1A methylation for breast cancer: a meta-analysis. Biosci Rep 2019;39. https://doi.org/10.1042/bsr20190923. Search in Google Scholar

96. Swets, JA. Measuring the accuracy of diagnostic systems. Science 1988;240:1285–93. https://doi.org/10.1126/science.3287615. Search in Google Scholar

97. Choupani, J, Nariman-Saleh-Fam, Z, Saadatian, Z, Ouladsahebmadarek, E, Masotti, A, Bastami, M. Association of mir-196a-2 rs11614913 and mir-149 rs2292832 polymorphisms with risk of cancer: an updated meta-analysis. Front Genet 2019;10:186. https://doi.org/10.3389/fgene.2019.00186. Search in Google Scholar

98. Cheuk, IWY, Shin, VY, Kwong, A. Detection of methylated circulating DNA as noninvasive biomarkers for breast cancer diagnosis. J Breast Cancer 2017;20:12. https://doi.org/10.4048/jbc.2017.20.1.12. Search in Google Scholar

99. Yu, D, Tong, Y, Guo, X, Feng, L, Jiang, Z, Ying, S, et al.. Diagnostic value of concentration of circulating cell-free DNA in breast cancer: a meta-analysis. Front Oncol 2019;9:95. https://doi.org/10.3389/fonc.2019.00095. Search in Google Scholar

100. Zhou, Y, Wang, C, Zhu, H, Lin, Y, Pan, B, Zhang, X, et al.. Diagnostic accuracy of PIK3CA mutation detection by circulating free DNA in breast cancer: a meta-analysis of diagnostic test accuracy. PLoS One 2016;11:e0158143. https://doi.org/10.1371/journal.pone.0158143. Search in Google Scholar

101. Ye, M, Huang, T, Ying, Y, Li, J, Yang, P, Ni, C, et al.. Detection of 14-3-3 sigma (σ) promoter methylation as a noninvasive biomarker using blood samples for breast cancer diagnosis. Oncotarget 2017;8:9230. https://doi.org/10.18632/oncotarget.13992. Search in Google Scholar

102. Bastami, M, Choupani, J, Saadatian, Z, Zununi Vahed, S, Ouladsahebmadarek, E, Mansoori, Y, et al.. Evidences from a systematic review and meta-analysis unveil the role of miRNA polymorphisms in the predisposition to female neoplasms. Int J Mol Sci 2019;20:5088. https://doi.org/10.3390/ijms20205088. Search in Google Scholar

103. Pepe, MS, Etzioni, R, Feng, Z, Potter, JD, Thompson, ML, Thornquist, M, et al.. Phases of biomarker development for early detection of cancer. J Natl Cancer Inst 2001;93:1054–61. https://doi.org/10.1093/jnci/93.14.1054. Search in Google Scholar

104. McGee, S. Simplifying likelihood ratios. J Gen Intern Med 2002;17:646–9. https://doi.org/10.1046/j.1525-1497.2002.10750.x. Search in Google Scholar

105. Stengel, D, Bauwens, K, Sehouli, J, Ekkernkamp, A, Porzsolt, F. Original Paper: a likelihood ratio approach to meta-analysis of diagnostic studies. J Med Screen 2003;10:47–51. https://doi.org/10.1258/096914103321610806. Search in Google Scholar

Supplementary Material

The online version of this article offers supplementary material (https://doi.org/10.1515/cclm-2021-0193).

Received: 2021-02-10
Accepted: 2021-04-23
Published Online: 2021-05-06
Published in Print: 2021-08-26

© 2021 Walter de Gruyter GmbH, Berlin/Boston