Table 1.
Diagnostic application of liquid biopsy in recent 5 years
| Region | Year | Sample | Detection Method | Cohorts | Detected Abnormality |
Practice in clinical |
Result | Ref | |
|---|---|---|---|---|---|---|---|---|---|
| CTC | USA | 2021 | Peripheral blood |
VERSA Platform, Immunofluorescence |
29 RCC patients |
CK ( +) CTC counts |
Distinguishing progressing and responding patients |
AUC 0.79, Sensitivity 73% and Specificity 100% | [24] |
|
ctDNA/ cfDNA |
Japan | 2021 | Plasma |
NGS dPCR |
56 ccRCC patients 31 healthy control |
VHL |
Detection of RCC patients |
13 VHL mutations were found in 12 of 56 ccRCC patients (21.6%) with median variant frequency of 0.78% VHL cfDNA mutations were found in 8 of 28 patients (28.6%) with VHL tumor DNA mutations Patients with VHL cfDNA mutations tended to show a worse OS |
[48] |
| USA | 2020 |
Plasma Urine |
cfMeDIP–seq NGS |
99 RCC patients 28 healthy controls 15 UBC patients |
300 DMRs |
Detection of RCC patients |
67/69 RCC samples (97.1%) were of a higher median methylation score than all control samples with a mean AUC of 0.990 Same analyses were carried out to urine cfDNA from patients with RCC and healthy controls, with the mean AUC of 0.858 |
[60] | |
|
Distinguishing RCC and UBC |
Using methylation score to compare patients with RCC and UBC, resulting in a mean AUC of 0.979 | ||||||||
| USA | 2020 | Plasma |
cfMeDIP–seq Target sequencing |
Cohort 1: 40 mRCC patients Cohort 2: 38 RCC patients 34 healthy controls |
Methylation level of 21 cfDNA variants |
Detection of mRCC patients |
cfDNA variant analysis via targeted sequencing detected 21 candidate variants in 11 of 40 mRCC patients (28%), which can improve the sensitivity combined with tumor DNA variant analysis All of 34 mRCC patients are detected through cfMeDIP–seq (sensitivity 100%, specificity 88%), compared with that cfDNA variant analysis identified variants in 7 patients (21%) |
[61] | |
| Canada | 2020 | Plasma | Target sequencing | 55 mRCC patients | VHL, BAP1, PBRM1 et al | Detection of mRCC patients | 17 of 51 mRCC patients detected cfDNA variants. The most frequent mutated genes are VHL, BAP1 and PBRM1 (the frequency is 41%, 39%, 17%, respectively). The concordance of mutated genes profiling between cfDNA variants in plasma and tumor DNA variants in matched tissues is 77% | [49] | |
| Japan | 2018 | Plasma |
qPCR Microfluidics-based platform |
92 RCC patients 41 healthy controls |
Plasma cfDNA level |
Detection of RCC patients |
AUC 0.762, Sensitivity 63.0% and Specificity 78.1% | [41] | |
| cfRNA | Portugal | 2022 | Plasma | ddPCR |
124 RCC patients 15 oncocytomas patients 64 healthy controls |
miR-21-5p miR-155-5p |
Detection of RCC patients |
Sensitivity 89.52%, specificity 54.69% and accuracy 77.66% | [161] |
| 124 RCC patients |
miR-21-5p miR-155-5p |
Detection of early stages RCC | Sensitivity 92.42%, specificity 34.38% and accuracy 63.85% | ||||||
|
124 RCC patients 15 oncocytomas patients |
miR-126-3p miR-200b-3p |
Distinguishing ccRCC and other RCC subtypes |
Sensitivity 80.46%, specificity 56.76% and accuracy 73.39% | ||||||
| Canada | 2021 | Urine | qPCR |
76 ccRCC patients 8 benign renal tumor patients 16 healthy contrls" |
Circ-EGLN3 Circ-SOD2 |
Detection of RCC patients |
69% of samples detected urinary circEGLN3 and 60% of samples detected urinary circACAD11 circEGLN3 levels were significantly different between the healthy controls versus ccRCC patients (P < 0.05) The AUC of circEGLN3 and circSOD2 was of 0.71 and 0.68, respectively, for distinguishing cancer patients versus non-neoplastic patients Urinary circEGLN3 level of ccRCC patients was lower than that of healthy controls, while tissue circEGLN3 level was higher of ccRCC patients |
[174] | |
| China | 2021 | Serum | qPCR |
123 RCC patients 118 healthy controls |
miR-21-5p miR-150-5p miR-145-5p miR-146a-5p |
Detection of RCC patients |
AUC 0.938, sensitivity 90.79%, specificity 93.75% | [162] | |
| China | 2020 | Serum | qPCR |
113 RCC patients 79 healthy controls |
LncRNA-C00886 |
Detection of RCC patients |
AUC 0.803, sensitivity 67.09%, specificity 89.87% | [172] | |
| Detection of early stages RCC patients | AUC 0.800, sensitivity 71.05%, specificity 89.87% | ||||||||
| Detection of non-metastasis RCC patients | AUC 0.830, sensitivity 73.33%, specificity 89.87% | ||||||||
| Portugal | 2020 | Urine | qMSP |
Cohort 2: 38 ccRCC patients 15 metastasis ccRCC patients 57 healthy controls Cohort 3: 171 ccRCC patients 85 healthy controls |
Methylation level of miR-30a-5p |
Detection of ccRCC patients Detection of metastasis ccRCC patients |
Cohort 2: AUC 0.6873, sensitivity 83%, specificity 53% Cohort 3: AUC 0.6702, sensitivity 63%, specificity 67% Cohort 2: AUC 0.7684, sensitivity 80%, specificity 71% |
[79] | |
| China | 2020 | Serum | qPCR |
146 RCC patients 150 healthy controls |
miR-224-5p miR-34b-3p miR-182-5p |
Detection of RCC patients |
AUC 0.855, sensitivity80.3%, specificity66.3% | [163] | |
| China | 2020 | Serum | qPCR |
Testing cohort: 70 RCC patients 70 healthy controls Validating cohort: 40 RCC patients 40 healthy controls |
miR-20b-5p, miR-30a-5p, miR-196a-5p |
Detection of RCC patients |
Testing cohort: AUC 0.949, sensitivity 92.8%, specificity 80.0% Validating cohort: AUC 0.938, sensitivity 92.5%, specificity 80.0% |
[164] | |
| Canada | 2020 | Urine | qPCR |
30 oncocytomas patients 26 progressive ccRCC-SRM patients 24 non-progressive ccRCC-SRM patients |
9 miRNAs miR-328-3p |
Distinguishing RCC- SRM and oncocytoma Detection of ccRCC patients |
9 urinary miRNAs were differentially expressed between renal oncocytoma (≤ 4 cm) and ccRCC-SRMs (pT1a; ≤ 4 cm), where miR-432-5p and miR-532-5p showed the most measurable discriminatory ability (AUC 0.71, AUC 0.70, respectively) miR-328-3p was significantly down-regulated in progressive ccRCC-SRMs and showed significant discriminatory ability (AUC: 0.68) |
[79] | |
| China | 2018 | Serum | qPCR |
46 RCC patients 46 healthy controls |
LncRNA-GIHCG |
Detection of RCC patients |
AUC 0.920, sensitivity 87%, specificity 84.8% | [173] | |
| Detection of early stage RCC patients | AUC 0.886, sensitivity 80.7%, specificity 84.8% | ||||||||
| Ukraine | 2018 | Urine | qPCR |
52 RCC patients 15 oncocytoma patients 15 healthy controls |
miR-15a | Distinguishing RCC and benign renal tumor | AUC 0.955, sensitivity 100%, specificity 98.1% | [169] | |
| Germany | 2018 | Serum | qPCR |
86 ccRCC patients 55 benign renal tumor patients 28 healthy controls |
miR-122-5p miR-206 |
Detection of ccRCC patients | AUC 0.733, sensitivity 57.1%, specificity 83.8% | [165] | |
| Protein | India | 2021 | Serum | Elisa |
60 RCC patients 60 non-tumor controls |
GRP78 |
Detection of RCC patients |
AUC 0.739, sensitivity 71.7%, specificity 66.7% | [179] |
| USA | 2021 | Plasma | Elisa |
143 mRCC patients 137 18–25 years old healthy controls 252 50–80 years old healthy controls |
hPG80 | Detection of mRCC patients |
Compared to 18–25 years old healthy group: AUC 0.93, accuracy 0.89 Compared to 50–80 years old healthy group: AUC 0.84, accuracy 0.77 |
[180] | |
| Canada | 2020 | Urine | LC–MS/MS |
27 oncocytoma (≤ 4 cm) patients 23 progressive ccRCC-SRM patients 21 non-progressive ccRCC-SRM patients 20 healthy controls |
GLRx、CST3、SLC9A3R1、HSPE1、FKBP1a、EEF1G et al |
Detection of early-stage ccRCC patients |
GLRx (AUC = 0.72, P = 0.0047) showed the most significant discriminatory ability between ccRCC-SRM and healthy controls, followed by SLC9A3R1 (AUC = 0.70), HSPE1 (AUC = 0.70), FKBP1A (AUC = 0.65) and EEF1G (AUC = 0.65) (P < 0.05) Diagnostic model based on the expression of 7 proteins (DDT, EEF1G, EPB41L3, HSPE1, MUC4, RAP1B and SLC9A3R1) showed the most significant discriminatory ability (AUC: 0.82), outperforming all single protein markers |
[178] | |
|
Distinguishing renal oncocytoma (≤ 4 cm) and early-stage ccRCC |
C12orf49 (AUC = 0.77, P = 0.0001) showed the most significant discriminatory ability between ccRCC-SRM and renal oncocytoma, followed by EHD4 (AUC = 0.64, p = 0.049) Diagnostic model based on the expression of 3 proteins (C12orf49、EHD4 and PPA1) showed the most significant discriminatory ability (AUC: 0.85), outperforming all single protein markers |
||||||||
|
Distinguishing progressive and non-progressive early-stage ccRCC |
EPS8L2 (AUC = 0.76, p = 0.0037) showed the most significant discriminatory ability between progressive and non-progressive ccRCC-SRM, followed by CHMP2A (AUC = 0.70, p = 0.034), PDCD6IP (AUC = 0.68), CNDP2 (AUC = 0.63) and CEACAM1 (AUC = 0.66)(P < 0.05) Diagnostic model based on the expression of 2 proteins (EPS8L2 and CCT6A) showed the most significant discriminatory ability (AUC = 0.81), outperforming all single protein markers |
||||||||
| USA | 2019 | Urine | Plasmonic biosensor |
20 RCC patients 20 healthy controls 8 BLCA patients 10 diabetic nephropathy patients |
PLIN-2 |
Detection of RCC patients |
Median urine PLIN-2 concentrations in ccRCC patients (43 ng/mL) were significantly higher (P < 0.001) than healthy groups (0.3 ng/mL), BLCA patients (0.5 ng/mL) and diabetic nephropathy patients (0.6 ng/mL) | [128] | |
| Metabolites | Portugal | 2021 | Urine | HS–SPME–GC–MS |
75 ccRCC patients 75 health control |
6 volatiles metabolites |
Detection of RCC patients |
The diagnostic model was consisted of 6 volatile metabolites The diagnostic ability of ccRCC patients: AUC 0.869, sensitivity 83%, specificity 79%, accuracy 79% The diagnostic ability of stage I ccRCC patients: AUC 0.799, sensitivity 84%, specificity 73%, accuracy 76% The diagnostic ability of stage III-IV ccRCC patients: AUC 0.911, sensitivity 83%, specificity 84%, accuracy 84% |
[101] |
| Italy | 2021 | Urine |
GC/MS Gas sensor array |
40 ccRCC patients 8 healthy controls |
8 volatiles metabolites |
Detection of ccRCC patients |
8 volatile metabolites was differentially expressed in at least 70% ccRCC patients and consisted as a diagnostic model Analyzed through GC/MS, the diagnostic ability of the model: AUC 0.979, sensitivity 85.7%, specificity 100%, accuracy 92.9% (training cohort); AUC 0.875, sensitivity 83.3%, specificity 100%, accuracy 91.7% (testing cohort) Analyzed through Gas Sensor Array, the diagnostic ability of the model: AUC 0.979, sensitivity 100%, specificity 85.7%, accuracy 92.9% (training cohort); AUC 0.906, sensitivity 100%, specificity 83.3%, accuracy 91.7% (testing cohort) |
[102] | |
| Germany | 2021 | Urine |
LC–MS NMR |
41 early stage RCC patients 29 advanced stage RCC patients |
16 urinary metabolites | Distinguishing early and advanced stage RCC patients | A model consisting of 16 metabolites was used for distinguishing early and advanced stage RCC patients: AUC 0.95, sensitivity 80%, specificity 91%, accuracy 86% | [105] | |
| China | 2020 | Urine | LC–MS |
39 RCC patients 22 benign renal tumor patients 68 healthy controls |
6 urinary metabolites |
Distinguishing RCC and benign renal tumor | A model consisting of 3 metabolites (cortolone, testosterone and l-2-aminoadipate adenylate) was used for benign and malignant renal tumor distinction: AUC 0.868, sensitivity 75%, specificity 100% (tenfold cross-validation of testing cohort) | [175] | |
|
Detection of RCC patients |
A model consisting of 3 metabolites (aminoadipic acid, 2-(formamido)-N1-(5-phospho-d-ribosyl) acetamidine and alpha-N-phenylacetyl-l-glutamine) was used for detection RCC patients: AUC 0.841, sensitivity 75%, specificity 88.6% (tenfold cross-validation of testing cohort) | ||||||||
| Japan | 2020 | Urine | LC–MS |
69 stage I-II RCC patients 18 stage III-IV RCC patients 60 benign renal tumor patients |
9 urinary metabolites |
Distinguishing RCC and benign renal tumor | A model consisting of 5 metabolites (L-glutamic acid, lactate, D-sedoheptulose 7-phosphate, 2-hy-droxyglutarate and myoinositol) was used for detection RCC patients: AUC 0.966, sensitivity 93.1%, specificity 95% | [176] | |
| Poland | 2020 | Urine | AuNPET LDI MS |
50 RCC patients 50 healthy controls |
15 urinary metabolites |
Detection of RCC patients |
15 urinary metabolites were identified abnormal distribution in RCC patients' urine (7 upregulation and 8 downregulation), where 3,5-Dihydroxyphenylvaleric acid showed the most significant diagnostic value (AUC 0.844) A model consisting of all 15 metabolites was used for detecting RCC patients: AUC 0.915, efficiency 88%, efficiency 86% |
[103] | |
| China | 2019 | Urine | UPLC-MS |
146 BLCA patients 115 RCC patients 142 healthy controls |
16 urinary metabolites |
Detection of RCC patients |
A model consisting of 6 metabolites (α-CEHC, β-cortolone, deoxyinosine, flunisolide, 11b,17a,21-trihydroxypreg-nenolone and glycerol tripropanoate) was used for distinguishing cancer patients from healthy controls: AUC 0.950 (discovering group); AUC 0.867 (external validating group) | [104] | |
|
Distinguishing BLCA and RCC patients without hematuria |
A model consisting of 6 metabolites (4-ethoxymethylphenol, prostaglandin F2b, thromboxane B3, hydroxybutyrylcarnitine, 3-hydroxyphloretin and N'-formylkynurenine) was used for distinguishing BLCA and RCC patients without hematuria: AUC 0.829 in discovering group; AUC 0.76 in external validating group | ||||||||
|
Distinguishing BLCA and RCC patients with hematuria |
A model consisting of 4 metabolites (1-hydroxy-2-oxopropyl tetrahydropterin, 1-acetoxy-2-hydroxy-16-heptadecyn-4-one, 1,2dehydrosalsolinol and L-tyrosine) was used for distinguishing BLCA and RCC patients with hematuria: AUC 0.913 (discovering group) | ||||||||
| China | 2019 | Urine | LC–MS |
100 RCC patients 34 benign renal tumor patients 129 healthy controls |
18 urinary metabolites |
Detection of RCC patients |
A model consisting of 9 metabolites (N-Jasmonoyltyrosine, Tetrahydroaldosterone-3-glucuronide, Androstenedione, Dopamine 4-sulfate, 3-Methylazelaic acid, Cortolone-3-glucuronide, 7alpha-hydroxy-3-oxochol-4-en-24-oic Acid, Cortolone-3-glucuronide and Lithocholyltaurine) was used for distinguishing cancer patients from healthy controls: AUC 0.905 (testing cohort); AUC 0.885 (validating cohort) N'-formylkynurenine showed a significant discriminating ability of detecting RCC patients (AUC 0.808, sensitivity 84.8%, specificity 83.8%) |
[177] | |
|
Distinguishing RCC and benign renal tumor |
A model consisting of 3 metabolites (L-3-hydroxykynurenine, 1,7-dimethylguanosine and tetrahydroaldosterone-3-glucuronide) was used for distinguishing RCC patients from benign renal tumor patients: AUC 0.834 in testing cohort; AUC 0.816 for tenfold cross-validation | ||||||||
| Distinguishing early and late stages of RCC | A model consisting of 5 metabolites (thymidine, cholic acid glucuronide, alanyl-proline, isoleucyl-hydroxyproline, and myristic acid) was used for distinguishing early (stage I-II) from late stages (stage III-IV) of RCC: AUC 0.881 in testing cohort; AUC 0.813 for tenfold cross-validation | ||||||||
| Exosome | Spain | 2021 | Plasma |
Differential ultracentrifugation, qPCR, NGS, dPCR |
13 RCC patients 15 healthy controls |
Exosomal mtDNA VH1, CγB |
Detection of RCC patients |
dPCR and qPCR demonstrated that VH1 and CγB were of a significant discrimination ability for RCC and healthy group (F phase) VH1: AUC = 0.825, P < 0.0001 for VH1-short; AUC = 0.833, P < 0.0001 for VH1-long CγB: AUC = 0.755, P < 0.0001 for CγB-short; AUC = 0.810, P < 0.0001 for CγB-long |
[183] |
| China | 2020 | Plasma |
exoEasy maxi kit, qPCR |
22 RCC patients 16 healthy controls |
Exosomal miRNA miR-92a-1-5p, miR-149-3p, miR-424-3p |
Detection of RCC patients |
Compared with healthy controls, the levels of exsomal miR-149-3p and miR-424-3p were significantly up-regulated, while miR-92a-1-5p was down-regulated miR-149-3p: AUC 0.7188, sensitivity 75.0%, specificity 72.7% miR-424-3p: AUC 0.7727, sensitivity 75.0%, specificity 81.8% miR-149-3p: AUC 0.8324, sensitivity 87.5%, specificity 77.3% |
[181] | |
| China | 2019 | Urine |
Differential ultracentrifugation, Agilent 2100 Bioanalyzer, NGS |
70 early‐stage ccRCC patients 30 early‐stage PC patients 30 early‐stage BLCA patients 30 healthy controls |
Exosomal miRNA miR-30c-5p |
Detection of ccRCC patients |
Exosomal miRNA miR-30c-5p levels in ccRCC patients' urine were significantly lower than those in healthy controls, where was no significant differences between BLCA cancers, PC cancers and healthy controls The diagnostic value of exosomal miR-30c-5p for ccRCC patients: AUC 0.819, sensitivity 68.57%, specificity 100% |
[182] | |
| China | 2018 | Serum |
Total exosome isolation reagent, EpCAM isolation beads, Flow cytometry |
82 ccRCC patients received nephrectomy 80 healthy controls |
Exosomal miRNA miR-210, miR-1233 |
Detection of RCC patients |
Exosomal miRNA miR-210 and miR-1233 levels in ccRCC patients' serum were significantly lower than those in healthy controls miR-210: AUC 0.69, sensitivity 70%, specificity 62.2% miR-1233: AUC 0.82, sensitivity 81%, specificity 76% |
[153] |