Liquid biopsy of plasma and cerebrospinal fluid for the detection of extracellular tumor DNA as a tool for glioma diagnosis and genotyping: a cross-sectional observational pilot study

Full Text

BACKGROUND

Gliomas account for approximately 80% of all central nervous system cancers [1]. Over 60% of these gliomas are glioblastomas, which have a 5-year survival rate of less than 7%. The relative 5-year survival rate for other high-grade gliomas is 30%–40% [1, 2]. Gliomas have largely similar morphologies; therefore, assessment by histopathological examination alone is insufficient for determining the tumor grade [1, 3–5]. That is why, since 2016, genetic testing of the tumor has been recommended for confirming the diagnosis, assessing the prognosis, and selecting the treatment strategy [1, 3–5]. According to the World Health Organization (WHO) classification, the majority of grade 4 gliomas show mutations in the TERT promoter (pTERT), whereas IDH1 mutations are reported only in 5%–10% of cases. In contrast, grade 1–3 gliomas more frequently demonstrate IDH1 mutations. Tumors with wild-type IDH mutations typically require more intensive treatment [1, 3–5]. Aberrations in the BRAF gene are an indication for BRAF (dabrafenib) and MEK (trametinib) inhibitor therapy [6]. The absence of MGMT gene methylation in patients with grade 4 glioma predicts high tumor resistance to temozolomide, which is a crucial chemotherapy component in gliomas, and may support its discontinuation [6]. Conversely, ATRX mutations and 1p/19q co-deletion are associated with a more favorable prognosis and better response to chemotherapy [5, 7, 8].

Currently, total tumor resection followed by chemotherapy and radiotherapy is the preferable treatment option, with simultaneous sampling for genotyping [3]. However, this surgery is impossible in 50%–70% of cases due to tumor properties or patient condition [3, 9]. Partial resection and tumor biopsy yield a limited amount of material, which may impede genotyping. Furthermore, neurosurgeries, including diagnostic ones to assess postoperative tumor progression, are associated with a risk of injury to functionally significant brain areas and subsequent nervous system disorders [3, 10].

Liquid biopsy is a promising technique for assessing genetic changes in gliomas. This term encompasses several approaches to analyzing tumor derivatives in body fluids, including tumor extracellular DNA (ecDNA), which may serve as a highly specific tumor marker [11]. The IDH1 R132H, pTERT C228T, and C250T single-nucleotide substitutions are prevalent and significant genetic changes that determine the course of the disease in gliomas. Therefore, analysis using droplet digital polymerase chain reaction (ddPCR) may determine the presence and potential progression of cancer with high sensitivity and specificity [3, 12]. Analyzing this marker in body fluids enables tumor monitoring and relapse prediction [13, 14]. Blood is the classical substrate for this technique; however, due to the blood–brain barrier, it contains only 6–7 ng/mL of tumor ecDNA in patients with glioma, which decreases the sensitivity of analysis [15, 16]. In contrast, the levels of this marker in the cerebrospinal fluid (CSF) can exceed 70 ng/mL, making CSF more suitable for analysis [17]. However, CSF tumor ecDNA levels significantly depend on the tumor grade and contact with CSF spaces [18]. Patients with WHO grade 1–3 gliomas located deep in the white matter may demonstrate limited levels of this marker, insufficient for analysis, even using ddPCR [18]. Preamplification increases the sensitivity of mutation detection by performing a preliminary PCR at the analyzed site. Zaytseva et al. emphasized the significance of preamplification in improving the efficacy of detecting the H3 K27M somatic variant in the CSF of patients with high-grade gliomas using ddPCR [18].

Margolick et al. performed preamplification before ddPCR for detecting cytomegalovirus, decreasing the minimal required DNA level twofold while maintaining the specificity of analysis [19]. Moreover, Diplas et al. used preamplification for detecting IDH mutations in cell line and frozen tissue samples from patients with diffuse gliomas [12].

The exact diagnostic and prognostic potential of liquid biopsy in gliomas remains unclear, due to inconsistent research findings. According to published data, the sensitivity and specificity of this technique may vary from 50% to 100% [8]. This is attributable to the wide variety of markers and analysis methods used and differences between study populations. This is supported by meta-analyses by McMahon et al. and Y. Kang et al. [20, 21].

The study aimed to assess the detectability and amount of tumor ecDNA with the IDH1 R132H, pTERT C228T, and C250T mutations in plasma and CSF of patients with gliomas using ddPCR with and without preamplification.

METHODS

An observational, cross-sectional study was conducted among hospitalized patients with glioma enrolled between April 2022 and July 2023 at the N.I. Pirogov National Medical and Surgical Center. The pilot study included plasma and CSF samples from 25 patients with newly diagnosed glioma (WHO grade 1–4).

The inclusion criteria were a confirmed glioma and surgery for total or partial tumor resection.

The exclusion criteria were an active cancer, a history of cancer, and cancer surgery or chemotherapy (before or during enrollment).

Patient clinical data were obtained from medical records and informed consent forms.

Histopathological examination of resected material was performed for morphological confirmation of the diagnosis. Spatial parameters of tumors were visualized and assessed using computed tomography, magnetic resonance imaging, and 18FDG positron emission tomography / computed tomography, as indicated. All relevant data were evaluated. The search for mutations was conducted on resected material by NGS in a commercial laboratory as part of the standard of care. All laboratory examinations were carried out at the Lomonosov Moscow State University Clinic. This pilot study’s findings will be used to conduct an open-label, prospective study in a larger patient population.

Biomaterial Collection, Treatment, And DNA Isolation

Peripheral venous blood samples of 10 mL were drawn from all patients into test tubes containing ethylenediaminetetraacetic acid. The biomaterial was collected prior to any invasive diagnostic procedure and treatment. Blood samples were stored at +4 °C for no more than 4 hours. Then, the samples were centrifuged at 4000 g for 12 minutes. Afterward, the samples were again centrifuged at 4000 g for 12 minutes to precipitate the remaining particles. Plasma samples were transferred to clean 5 mL tubes and frozen at −80 °C.

CSF samples of 2.5–5 mL were collected intraoperatively and immediately frozen at −80 °C. Samples were collected by lumbar puncture before tumor tissue or brain damage to prevent a false increase in blood and CSF tumor ecDNA levels.

After thawing, plasma and CSF samples were thoroughly mixed by pulse vortexing. Furthermore, DNA was isolated from 5 mL of plasma and 2.5–5 mL of CSF using the Circulating Nucleic Acid Kit (Qiagen, Germany) according to the manufacturer’s instructions. The elution volume was 100 µL.

Extracellular DNA Analysis

Tumor ecDNA was analyzed in all plasma and CSF samples to detect IDH1 R132H, pTERT C228T, and C250T mutations. The modified tumor ecDNA analysis protocol was tested using preamplification for the IDH1 R132H, pTERT C228T, and C250T mutations.

Preamplification was performed on a CFX96 amplifier (Bio-Rad, USA) according to the manufacturer’s instructions. IDH1 preamplification was conducted with a ready-to-use PCR master mix 5X qPCR mix HS (Evrogen, Russia) and a set of primers (forward primer 5ʹ-TATTGCCAACATGACTTA-3ʹand reverse primer 5ʹ-TGAAACAAATGTGGAAAT-3ʹ; concentration in the reaction mixture: 0.4 μM).

Moreover, pTERT preamplification was performed with a ready-to-use PCR master mix 5X qPCR mix HS (UDG) (Evrogen, Russia) and a set of primers (forward primer 5ʹ-AGCGCTGCCTGAAACTCG-3ʹ and reverse primer 5ʹ-CCTGCCCCTTCACCTTCCAG-3ʹ; concentration in the reaction mixture: 0.4 μM). The analyzed sequence was a GC-rich DNA region. Therefore, additional reagents were utilized for preamplification, such as dimethyl sulfoxide (DMSO) (5% of the reaction mixture volume), betaine (1.25 mol/L), formamide (5% of the reaction mixture volume), deaza-deoxyguanosine triphosphate (1% of the reaction mixture volume), Encyclo buffer for GC-rich DNA amplification (Evrogen, Russia), and DMSO + formamide or DMSO + betaine. Table A1 shows the composition of the reaction mixtures for pTERT preamplification.

The optimal primer annealing temperature was determined in a series of experiments with control DNA samples: 57°C for IDH1 and 54°C for pTERT. Tables А2 and А3 presents the preamplification protocols.

Prior to subsequent DNA analysis, the mixture was diluted using TE buffer (1:1).

The total ecDNA before and after preamplification was analyzed by ddPCR. The automated ddPCR system QX200 AutoDG ddPCR (Bio-Rad, USA) was used for droplet formation and data reading according to the manufacturer’s instructions. The IDH1 mutant allele was detected using the ddPCR™ Mutation Assay: IDH1 p.R132H (Bio-Rad). The following primers were used for ddPCR of pTERT mutations: forward primer 5ʹ-AGCGCTGCCTGAAACTCG-3ʹ, reverse primer 5ʹ-CCTGCCCCTTCACCTTCCAG-3ʹ, and LNA probes for the wild-type allele HEX-CCCC+C+T+CCGG-BHQ-1 and mutant allele (suitable for both substitutions) 6-FAM-CCC+C+T+T+CCGG-BHQ-1 [22]. Figure B1 shows the nucleotide sequence of primers and probes.

The primer and probe concentrations in the reaction mixture were 0.9 μM and 0.25 μM, respectively. The target DNA concentration in the final ddPCR reaction mixture was set at 8.8 μL per cell for the IDH1 R132H mutation and 9.9 μL for mutant pTERT C228T and C250T copies. Thermal cycling conditions were set according to the manufacturer’s protocols (see Tables А4 and А5).

The total DNA levels were determined as the sum of mutant and wild-type IDH1 alleles. The ddPCR findings are presented as absolute values (number of copies per 1 mL of biomaterial) and mutant allele fractions (MAFs) (tumor ecDNA level/total ecDNA level, %).

DNA levels in a body fluid were calculated using the following equation:

С_mL = С_μL × 2200/(V_sub × V_DNA)

where С_mL is plasma/CSF DNA concentration, copies/mL; С_μL is DNA concentration in the final reaction mixture, measured during ddPCR, copies/μL; V_sub is the volume of substrate from which the DNA was isolated, mL; and V_DNA is DNA volume in the final ddPCR reaction mixture, μL.

Preamplification causes heterogeneous changes in DNA levels. Therefore, DNA levels in the reaction mixture were not extrapolated to the substrate volume.

The threshold for false-positive tumor ecDNA signals for the mutant allele was set in a series of ddPCR tests using wild-type DNA samples and a negative control without a template.

Statistical Analysis

Statistical analysis was performed using IBM SPSS Statistics 26.0 (IBM Corp., USA).

The Shapiro–Wilk test was employed for normality testing in each group. Due to a non-normal distribution, the Mann–Whitney and Wilcoxon tests (for paired data) were used. Correlations were assessed using the Spearman’s rank correlation coefficient. Quantitative data are presented as a median (first quartile; third quartile). Differences were considered significant at p < 0.05.

RESULTS

Patient Characteristics

Plasma and CSF samples from 25 patients with glioma were examined. The mean age of patients at the time of the first blood sampling was 46.6 (28–71) years; the study included 12 males and 13 females. Table 1 shows the clinical characteristics of the study population.

 

Table 1. Demographic and clinical characteristics of study participants

Parameter	Values
Age, years	46,6 (28–71)1
Male/female, n (%)	12/13 (48/52)
BMI, kg/m2	25,85 (17,04–32,45)1
WHO grade:
1, n	1
2, n	1
3, n	9
4, n	14
Tumor site, n:
Frontal lobe	14
Parietal lobe	12
Temporal lobe	10
Insular lobe	6
Occipital lobe	4
Thalamus	1
Tumor volume, cm3:
<50	13
50–100	4
100–150	6
>150	1
Contrast uptake:
None, n	10/25
Moderate/heterogeneous, n	7/25
Intensive, n	1/25
Ring-shaped, n	7/25
IDH1 R132H	11

Note: ¹Data are presented as a mean (range). BMI, body mass index; WHO, World Health Organization.

 

Data on IDH1 mutations in tissues were available for 18/25 patients included in the study: 9/11 with WHO grade 1–3 and 9/14 with WHO grade 4. IDH1 R132H mutation was detected in 7/9 (77.8%) patients with grade 1–3 gliomas and 4/9 (44.4%) patients with grade 4 glioma according to sequencing. Table А6 presents more detailed information.

Diagnostic Efficacy of Analysis Without Preamplification

Plasma ecDNA was analyzed in 25 patients and CSF DNA in 24 patients (the CSF sample from 1 patient was missing). A series of experiments with wild-type DNA control samples and a no-template control established the thresholds for false-positive signals; ecDNA levels were presented as the number of copies per milliliter of substrate. Detailed findings are presented in Table А6.

According to the analysis without preamplification, CSF and plasma ecDNA levels were 1437.5 (661.8; 2320) and 1410.66 (915.6; 4890.11) copies/mL, respectively. No significant differences were found between total ecDNA levels in pairs of CSF and plasma samples (p > 0.05). Figure 1 presents the comparison of total ecDNA levels in pairs of CSF and plasma samples in the pilot study. The highest total CSF and plasma ecDNA levels were 31,215.1 and 145,038.04 copies/mL, respectively.

 

Fig. 1. Comparison of extracellular DNA levels in the cerebrospinal fluid and plasma of patients with glioma. CSF, cerebrospinal fluid. Outliers of 31,215.1 and 145,038.04 copies/mL for cerebrospinal fluid and plasma, respectively, were excluded to maintain figure proportions. P-value was calculated using the Wilcoxon test.

 

None of the plasma samples contained tumor ecDNA with IDH1 R132H and pTERT C228T/C250T mutations. CSF analysis without preamplification detected tumor ecDNA with IDH1 R132H (three samples) and pTERT C228Т (four samples), but not pTERT C250T. IDH1 R132H mutation was determined in 1/7 samples from patients with WHO grade 1–3 gliomas and 2/4 samples from patients with WHO grade 4 glioma, which were also positive for this mutation according to biopsy findings. Sequencing confirmed no false-positive mutation detection in tissues. All four pTERT mutations were observed in the CSF of patients with grade 4 glioma; however, none of the samples from these patients contained IDH1 R132H according to biopsy findings. The CSF level of tumor ecDNA with pTERT and IDH1 mutations was 195.11 (32.43; 276.45) copies/mL, and the MAF was 4.04% (3.29%; 9.43%) (see Fig. B2). No samples contained both mutant alleles. Table A6 shows the levels and MAFs for the positive samples mentioned above.

Diagnostic Efficacy of Analysis After Preamplification

Despite the wide range of reagents for the amplification of GC-rich sequences, preamplification of DNA with pTERT regions did not yield positive results. Various protocols typically used in these scenarios were tested, including the use of betaine, DMSO, 7-deaza-deoxyguanosine triphosphate, and uridine triphosphate; however, they failed to produce the expected result (see Table A1) [23]. Notably, the amplification of this region was challenging even with ddPCR and was only possible when a uridine triphosphate-containing PCR master mix was used. Therefore, only preamplification results for CSF DNA with IDH1 mutations are presented.

Experiments with preamplification revealed that wild-type DNA and negative control (without a template) samples showed no false-positive emulsion droplets during ddPCR. Figures 2a and 3b show 2D graphs with ddPCR findings for the same sample before and after preamplification.

 

Fig. 2. Comparison of extracellular DNA levels in the cerebrospinal fluid and plasma of patients with glioma before (a) and after (b) pre-amplification. CSF, cerebrospinal fluid. Outliers of 31,215.1 and 145,038.04 copies/mL for cerebrospinal fluid and plasma, respectively, were excluded to maintain figure proportions. P-value was calculated using the Wilcoxon test.

 

The exact number of DNA copies in the original sample is difficult to assess after preamplification; therefore, ecDNA levels are shown in copies per microliter of the reaction mixture.

As expected, the total CSF ecDNA levels after preamplification were significantly higher than before preamplification (1953.96 [627.93; 3331.3] vs. 23.7 [12.9; 36.3] copies/μL of the reaction mixture; р < 0.001) (see Fig. B3). Therefore, the CSF ecDNA level in the final reaction mixture increased by an average of 91.06 (74.77; 123.91) times. Target levels after preamplification showed a significant moderate correlation with those without preamplification (r = 0.746; p < 0.001) (see Fig. 3). Furthermore, an increase was noted in CSF tumor ecDNA levels in three samples with the previously detected IDH1 R132H mutation: from 0.515 to 3.9 (7.6 times), from 0.302 to 4.62 (15.3 times), and from 5.840 to 126.35 (21.6 times) copies/μL of the reaction mixture. However, the IDH1 R132H MAF in these samples decreased from 11.55% to 3.64%, from 4.71% to 4%, and from 3.03% to 2.06%.

 

Fig. 3. Scatter plot of extracellular DNA levels before and after pre-amplification in the cerebrospinal fluid of patients with glioma. The x-axis shows the total extracellular DNA concentration in cerebrospinal fluid before pre-amplification (copies/μL); the y-axis shows the total extracellular DNA concentration after pre-amplification (copies/μL). CSF, cerebrospinal fluid. Data are presented as copies per 1 μL of reaction mixture. P-value was calculated using Spearman’s test.

 

Overall, the detection rate of CSF tumor ecDNA with the IDH1 R132H mutation after preamplification increased in 1/7 samples from patients with WHO grade 1–3 gliomas and 2/4 samples from patients with WHO grade 4 glioma (to 4/7 and 3/4, respectively). Moreover, IDH1 R132H mutation not detected by biopsy was found in one patient with grade 1 glioma and in one patient with grade 4 glioma. IDH1 R132H mutation was detected in the CSF of one patient with grade 4 glioma; however, the biopsy findings for this patient were missing. Therefore, the detection rate of tumor ecDNA with IDH1 R132H mutation increased from 2/24 to 10/24 (see Table А6). In all samples with CSF tumor ecDNA with IDH1 R132H mutation, its level was 0.54 (0.3; 3.19) copies/μL, and MAF was 0.042% (0.031%; 0.108%).

Association Between ecDNA Levels and Clinical and Demographic Characteristics of Patients

Statistical analysis of ddPCR findings before and after preamplification revealed no significant correlations between total CSF and plasma ecDNA levels and clinical and demographic characteristics of patients (age, sex, height, body weight, and BMI) (p > 0.05). Scattering graphs with numerical values of correlation coefficients (R) and р values for each correlation are presented in Fig. B4. An increase was observed in ecDNA levels in groups formed by tumor grade and contrast uptake; however, no significant correlations between these parameters and tumor properties were found. Moreover, no correlation was noted between tumor site and ecDNA level. A correlation was only found between the total plasma ecDNA level and contrast uptake (р = 0.034) (see Fig. 4).

 

Fig. 4. Graphs of total extracellular DNA levels in biological fluids of patients stratified into groups: a, total extracellular DNA levels in plasma across patient groups according to contrast enhancement pattern; b, total extracellular DNA levels in cerebrospinal fluid of patients with detected extracellular tumor DNA, grouped by glioma grade; and c, total extracellular DNA levels in plasma of patients with detected extracellular tumor DNA, grouped by contrast enhancement pattern. WHO, World Health Organization. a, outliers of 31,215.1 and 145,038.04 copies/mL for cerebrospinal fluid and plasma, respectively, were excluded to maintain figure proportions. P-value was determined using the Kruskal–Wallis test.

 

After excluding presumably false-positive samples, no correlation was determined between plasma and CSF ecDNA levels and age, sex, BMI, or tumor site in patients with mutant DNA containing mutations in both genes, detected before and after preamplification. However, a correlation was observed between total CSF ecDNA levels before preamplification and tumor grade (р = 0.024) and between total plasma ecDNA levels and contrast uptake (р = 0.046) (Fig. 4). After excluding presumably false-positive samples and samples with negative IDH1 mutant alleles, no significant correlation was found between absolute tumor ecDNA levels or MAF and age, sex, or BMI in patients with mutant DNA detected in plasma and CSF before and after preamplification (р > 0.05). Figure B5 presents scattering graphs with numerical values of correlation coefficients (R) and р values for each correlation.

DISCUSSION

Analysis without preamplification revealed no significant differences between total plasma and CSF ecDNA levels. IDH1 preamplification increased both total ecDNA and tumor ecDNA levels, as expected. Ten preamplification cycles are expected to increase the amplicon content by 2¹⁰ (i.e., approximately 1000 times); however, preamplification only increased the total ecDNA level by 91.06 (74.77; 123.91) times. This indicates that an increase in DNA for each cycle was approximately 57% of the expected value. This may be caused by ineffective primer annealing or PCR inhibitors in the mixture [24]. The increase in tumor ecDNA was even lower (up to 21.6 times), a finding confirmed by a decrease in MAF in samples with IDH1 mutation detected before and after preamplification. Preamplification of the mutant allele is less effective due to competition between primers with excessive wild-type allele for complementary binding to the least prevalent mutant amplicon [25]. This can be addressed using various methods, such as allele-specific PCR, wild-type DNA amplification inhibition using selective agents, and high-resolution melting [25, 26]. However, comparing only two samples is insufficient for assessing changes in the tumor DNA to wild-type DNA ratio in the reaction mixture during preamplification.

According to ddPCR without preamplification, none of the plasma samples contained tumor ecDNA. In the CSF samples, the sensitivity and specificity of IDH R132H mutation detection by liquid biopsy were 14.3% and 100%, respectively, for patients with WHO grade 1–3 gliomas, and 50% and 100%, respectively, for patients with WHO grade 4 glioma. Notably, ddPCR without preamplification showed extremely low levels of IDH1 and pTERT mutant alleles. In some cases, ddPCR revealed only few emulsion droplets positive for mutant DNA. Therefore, the experiment was conducted at the lower sensitivity threshold of three droplets with mutant DNA, which may explain the low detection rate of mutant alleles in the study sample.

After preamplification, the sensitivity and specificity of IDH R132H detection were 57.1% and 50%, respectively, for patients with WHO grade 1–3 gliomas, and 75% and 80%, respectively, for patients with WHO grade 4 glioma. As previously mentioned, CSF tumor ecDNA levels depend on the glioma grade and site in the brain. Mutant DNA is more frequently detected in body fluids of patients with WHO grade 1–3 gliomas (glioblastoma, anaplastic astrocytoma, etc.) and in tumors located near the CSF spaces, which may explain intergroup differences in liquid biopsy sensitivity in our study [8, 17, 27]. Furthermore, CSF collected from ventricles located closer to the tumor may contain more ecDNA than samples obtained by lumbar puncture [28].

Liquid biopsy findings indicated a pTERT mutation in four CSF samples from patients with grade 4 glioma without IDH1 mutation. This finding may be explained by the fact that glioblastomas are the most common grade 4 gliomas, which typically lack IDH1 mutations but carry mutant pTERT alleles. Moreover, approximately 50% of grade 2 and 3 gliomas are astrocytomas, with an IDH1 mutation but without pTERT mutations [1, 3, 4].

According to published data, the sensitivity and specificity of liquid biopsy in detecting IDH1 and pTERT mutations in plasma and CSF are 50%–100% and 80%–100%, respectively [27, 29, 30]. In the present study, the ecDNA detection sensitivity in patients with both grade 1–3 and grade 4 gliomas falls within this range; however, the specificity in patients with grade 1–3 gliomas is significantly lower. These results are influenced by the small sample size, Samples from 11 patients with grade 1–3 gliomas were analyzed; however, only 2 samples had no IDH1 R132H mutations.

The detection of IDH1 R132H mutation after preamplification but not according to sequencing could be explained by a false-positive result, which was not observed in the analysis of wild-type samples, and undetected local metastases or low tumor allele frequency in tissues, insufficient for detection by sequencing. Xie et al. detected HER2 amplification by liquid biopsy in plasma samples in 28% of patients with breast cancer who were previously classified as HER2-negative based on ddPCR of biopsy samples [31]. Crucitta et al. detected IDH R132H mutation in plasma samples in 7.3% of patients using ddPCR; however, this mutation was not detected in tumor tissues by Sanger sequencing [29].

Moreover, ddPCR and pTERT preamplification used the same primers; however, preamplification of the pTERT gene did not yield positive results, unlike that of IDH1. This could be explained by the complex, GC-rich structure of this gene’s promoter, which is susceptible to forming quadruplex structures that inhibit Taq polymerase [32]. However, why amplification was successful with emulsion PCR, but not in a non-compartmentalized volume of the reaction mixture during preamplification, remains unclear [24].

Statistical analysis revealed a correlation between both ecDNA and tumor ecDNA levels and tumor volume, grade, and contrast uptake. No significant correlation was found between mutant and wild-type DNA levels in the overall sample and the tumor volume. This could be attributed to the small sample size, because similar studies have demonstrated a significant correlation between these parameters [30, 33].

However, a significant correlation was noted between plasma ecDNA levels and contrast uptake in the overall sample and in patients with detected tumor ecDNA (with IDH1 and pTERT mutant alleles). This may indicate the structure of blood vessels at the tumor site. No contrast uptake was observed in 8 of 12 patients with grade 1–3 gliomas, whereas ring-shaped contrast uptake was observed in patients with grade 4 glioma, indicating a more aggressive tumor growth [34]. Furthermore, patients with tumor ecDNA grouped by the glioma grade showed a significant difference in ecDNA levels.

Therefore, liquid biopsy of plasma and CSF samples is a promising approach for glioma diagnosis and genotyping. However, this technique still has limited sensitivity and specificity in detecting tumor ecDNA, especially in low-grade glioma, and currently cannot replace tumor tissue genotyping. Furthermore, despite the significance of molecular genetic testing in central nervous system tumors, histopathological examinations continue to play the key role in diagnosis. Liquid biopsy provides unique diagnostic opportunities that are of significant interest to researchers and clinicians, supporting further development of this technique.

Limitations

The study had several limitations that influenced the results. First, this was a pilot study; therefore, the sample size was limited to 25 patients, which decreases the significance of the findings. Second, data on the IDH1 mutation status according to sequencing were only available for 17/24, and information on the pTERT mutation status was lacking, preventing an accurate sensitivity and specificity assessment. Furthermore, access to tissue samples in this pilot study was limited; therefore, validation was based on comparisons with sequencing findings. Notably, NGS is approved for resected material genotyping in central nervous system tumors, and its sensitivity is comparable to that of ddPCR [3, 35]. However, the main study protocol included resected material collection and ddPCR analysis to confirm NGS findings.

CONCLUSION

This study assessed the sensitivity and specificity of liquid biopsy of plasma and CSF samples in patients with glioma. IDH1 and pTERT mutations were detected using ddPCR with or without preamplification. This technique has low sensitivity and specificity for detecting gliomas, especially low-grade ones. However, it evaluates significant tumor properties, such as size and contrast uptake, which may be useful in selecting treatment. Furthermore, this technique may provide new prospects for noninvasive tumor monitoring and growth prognosis.

ADDITIONAL INFORMATION

Supplement А: Tables А1–A6. DOI: 10.17816/KMJ679633-4395847

Supplement B: Figures B1–B5. DOI: 10.17816/KMJ679633-4395849

Author contributions: R.T.I.: validation, investigation, formal analysis, writing—original draft, writing—review & editing; J.M.: conceptualization, methodology, validation, investigation, formal analysis, writing—review & editing, project administration, funding acquisition; S.L.M.: writing—review & editing, project administration; A.I.M.: methodology, investigation, formal analysis, writing—review & editing; Z.A.A.: conceptualization, methodology, writing—review & editing, project administration. All the authors approved the version of the manuscript to be published and agreed to be accountable for all aspects of the work, ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Ethics approval: The study protocol was approved by the local Ethics Committee of the N.I. Pirogov National Medical and Surgical Center (minutes No. 3, dated March 16, 2022). Written informed consents were obtained from all the participants.

Informed consent: All presented data are anonymized; no photographs are published.

Funding sources: The study was financially supported by the state assignment Development of Innovative Laboratory Technologies, Including Molecular Genetic Ones, for Disease Diagnostics (Registration No. 123-032800010-0, 2024). The Russian Federation had no involvement in the design, planning, or conduct of the study; data collection, storage, analysis, or interpretation; manuscript preparation; publication decisions; or study oversight. The funding organizations imposed no restrictions on data utilization or the dissemination of study findings.

Disclosure of interests: The authors have no relationships, activities, or interests for the last three years related to for-profit or not-for-profit third parties whose interests may be affected by the content of the article.

Statement of originality: No previously obtained or published material (text, images, or data) was used in this study or article.

Data availability statement: All data obtained in this study are available in this article.

Generative AI: No generative artificial intelligence technologies were used to prepare this article.

Provenance and peer-review: This paper was submitted unsolicited and reviewed following the standard procedure. The peer review process involved two external reviewers, a member of the Editorial Board, and the in-house science editor.

About the authors

Tagir I. Rakhmatullin

Lomonosov Moscow State University

Author for correspondence.
Email: tagir.rakhmatullin@internet.ru
ORCID iD: 0000-0002-4601-3478
SPIN-code: 7068-1678

intern-researcher

Russian Federation, Moscow

Mark Jain

Lomonosov Moscow State University

Email: jain-mark@outlook.com
ORCID iD: 0000-0002-6594-8113
SPIN-code: 3783-4441

Cand. Sci. (Biology), senior research associate

Russian Federation, Moscow

Larisa M. Samokhodskaya

Lomonosov Moscow State University

Email: slm@fbm.msu.ru
ORCID iD: 0000-0001-6734-3989
SPIN-code: 5404-6202

MD, Cand. Sci. (Medicine), assistant professor

Russian Federation, Moscow

Ivan M. Alekseev

National Medical and Surgical Center named after N.I. Pirogov

Email: alexeev.im@yandex.ru
ORCID iD: 0000-0001-8107-3065
SPIN-code: 9947-1988

neurosurgeon

Russian Federation, Moscow

Andrey A. Zuev

National Medical and Surgical Center named after N.I. Pirogov

Email: mosbrain@gmail.com
ORCID iD: 0000-0003-2974-1462
SPIN-code: 9377-4574

MD, Dr. Sci. (Medicine), Professor

Russian Federation, Moscow

References

Supplementary files