Table 3.

Monitoring treatment and predicting response by liquid biopsy in recent 5 years

	Region	Year	Detected Abnormality	Sample	Detection Method	Cohorts	Practice in clinical	Result	Ref
CTC	China	2021	CTC counts	Peripheral blood	CTC-BIOPSY system	41 RCC patients	Monitoring the postoperative condition of patients	• In the high CTC group, CTC counts decreased in 21 of 24 (87.5%) patients 1 week after surgery compared with the low CTC group (52.9%)	[26]
	USA	2021	PD-L1	Peripheral blood	VERSA Platform, Immunofluorescence	20 RCC patients treated with ICI	Evaluating the responses of ICI therapy	• PD-L1 expression in CAXII single positive CTC correlates with the efficacy of ICI treatment. Detection of progressing patients with ICI therapy: AUC 0.77, sensitivity 67% and specificity 88%	[24]
			HLA-1	Peripheral blood	VERSA Platform, Immunofluorescence	22 RCC patients treated with TKI	Evaluating the responses of TKI therapy	• PD-L1 expression in CAXII single positive CTC correlates with the efficacy of TKI treatment. Detection of progressing patients with TKI therapy: AUC 0.83, sensitivity 100% and specificity 88%
	Japan	2020	CTC counts	Peripheral blood	FISHMAN-R flow cytometer	54 RCC patients treated with NE or RE	Evaluating the responses of surgery	• Postoperative CTCs was significantly correlated with tumor diameter (P = 0.0004) and surgical approach (P = 0.016)	[189]
	China	2019	Beclin-1	Peripheral blood	Can Patrol CTC enrichment technique, RNA ISH	58 metastasis-free RCC patients 11 metastasis RCC patients	Monitoring the postoperative metastasis	• Beclin1-positive epithelial CTCs in the metastatic group at 12 months postoperatively was significantly higher than preoperatively • Beclin1-positive mesenchymal CTC in the metastatic group at 6 months postoperatively was significantly higher than preoperatively • Beclin1-positive mesenchymal CTC in the metastatic group at 6 months postoperatively was significantly higher than preoperatively; at 12 months postoperatively was significantly higher than 6 months postoperatively and preoperatively	[25]
ctDNA/ cfDNA	Italy	2022	TP53	Plasma	NGS	12 mccRCC patients with immunotherapy 36 mccRCC patients with TKI therapy	Evaluating the responses of TKI and immune therapy	• The most frequently mutated genes in cfDNA were TP53 (43%) and PDGFRA (21%), followed by mTOR, PI3K, BRAF etc. The used NGS panel did not include VHL • Patients with at ctDNA > 0.883 ng/μl had a shorter PFS and OS versus those with ctDNA > 0.883 ng/μl in overall population (P < 0.001, P < 0.008, respectively). The results were consistent with patients treated immunotherapy and TKI separately (P < 0.0365, P < 0.0035, respectively) • Patients with TP53 mutation have a shorter PFS than those who do not (P = 0.04) • Comprehensively evaluated of both ctDNA level and TP53 mutation status, patients with high cfDNA and mutated TP53 had the worst PFS, while patients with low cfDNA and no TP53 mutations had the longer PFS (P = 0.004)	[51]
							Predicting the best response to TKI and immunotherapy	• cfDNA level was associated with best response in the overall population (P = 0.006), which is consistent with patients with immunotherapy (P = 0.004) and TKIs (P = 0.003) • cfDNA cut point of ≥ 2.19 ng/μl for early progressors: Youden’s 0.75, sensitivity 100%, specificity 75% • cfDNA cut point of ≤ 1.35 ng/μl for long progressors: Youden’s 0.556, sensitivity 78%, specificity 78%
	Japan	2022	VHL, TP53, ATM, MET	Plasma	NGS	11 ccRCC patients with ICI therapy	Predicting response to ICI therapy in mRCC patients	• The commonly mutated genes were VHL (30.0%), TP53 (20.0%), ATM (10.0%), and MET (10.0%) • The coincidence rate of VHL (9 of 14 patients), TP53(2 of 14 patients) and MET (2 of 14 patients) between plasma ctDNA and tumor tissue DNA is 55.6%, 100%, 50%, respectively • For ICI-treated patients, ctDNA decreased in 4 of 5 responders and increased in 5 of 6 non-responders. A longer PFS is showed in the ctDNA-decreased group than ctDNA-increased group	[47]
	Spain	2021	GAPDH, hTERT	Plasma	qPCR	82 RCC patients 20 healthy controls	Evaluating surgery effects	• After nephrectomy, the mean level of GAPDH cfDNA was 16.9 fg/ml, which was significantly lower than preoperative level (29.3 fg/ml, P < 0.0001)	[190]
							Predicting patients' risk of death	• Univariate Cox Regression analysis showed that each fg/ml of GAPDH cfDNA increased the risk of progression by 14.8 postoperatively in mRCC patients • Each fg/ml of GAPDH cfDNA and hTERT cfDNA increased the risk of progression by 1.04 and 1.23 postoperatively, respectively
	Japan	2019	VHL, TP53, mTOR, TSC1, BAP1et al	Plasma	NGS ddPCR	53 RCC patients	Monitoring responses to surgery and TKI therapy in RCC patients	• The mutant allele frequency (MAF) of VHL, TP53 and other ctDNA decreased postoperatively, which reflected the changes of tumor burden • Patients with short fragment sizes of cfDNA showed significantly worse responsiveness (P = 0.011). For TKI-treated patients, positive ctDNA was significantly associated with weaker effect (P = 0.049), and short fragment sizes of cfDNA tended to be associated with worse outcome (P = 0.090)	[50]
cfRNA	Ukraine	2018	miR-15a	Urine	qPCR	52 RCC patients 15 oncocytoma patients 15 healthy controls	Evaluating surgery effects	• The tumor size related to the expression of miR-15a (Pearson correlation coefficient 0.873) • The mean expression of miR-15a in patients with nephrectomy decreased by 99.53% (P < 0.01) on the 8th day postoperatively	[169]
	China	2018	lnc-GIHCG	Serum	qPCR	20 RCC patients with total nephrectomy	Evaluating surgery effects	• Serum GIHCG level significantly decreased in postoperatively compared than preoperatively (P < 0.001)	[173]
Protein	USA	2021	KIM-1	Plasma	microbead-based assay	418 ccRCC patients with total nephrectomy	Predicting patients' postoperative survivals	• Higher post-nephrectomy baseline of KIM-1 was related to worse DFS, with a survival time ratio of 0.65 (in univariable lognormal AFT model) and 0.56 (in variable lognormal AFT model) for the 75th vs 25th percentile of baseline KIM-1 (p = 0.0004, P < 0.001, respectively) • Higher post-nephrectomy baseline was related to worse OS in a multivariable AFT model, with a survival time ratio 0.71 for 75th vs 25th percentile of KIM-1 (P < 0.001)	[191]
							Predicting patients' postoperative recurrence risk	• Added to either the SSIGN score or the UISS score, baseline KIM 1 improved the predictive value for recurrence after nephrectomy of both models (likelihood ratio test p = 0.078, p = 0.0022, respectively)
	UK	2021	CA9, HGF, MET, Gas6, Ax1, VEGF, VEGFR2, IL-8	Plasma	Luminex assay platforms Elisa	330 advanced RCC patients with cabozantinib 330 advanced RCC patients with everolimus	Predicting the survival and responses of cabozantinib	• The multivariable analysis showed that baseline levels of HGF were independent prognostic biomarker of PFS for cabozantinib, and HGF, GAS6, VEGF were independent prognostic biomarker for OS with cabozantinib • Decrease of AXL level were independent prognostic biomarker of PFS for cabozantinib, and decrease of levels of HGF, GAS6 were both independent prognostic biomarker for improved OS with cabozantinib	[193]
							Predicting the survival and responses of everolimus	• The multivariable analysis showed baseline levels of HGF were independent prognostic biomarker of longer PFS for everolimus. No biomarkers were independently prognostic for OS with everolimus • Decrease of HGF level were independent prognostic biomarker of PFS and OS for everolimus
	USA	2021	23 angiokines, including Ang-2, CD-73, HER-3, HGF, IL-6, OPN etc	Plasma	SP-X imaging and analysis system from Quanterix Ella System (Protein Simple)	53 non-ccRCC patients with everolimus 46 non-ccRCC patients with sunitinib	Predicting the survival and treatment benefit of RCC therapy	Exception of HER-3, SDF-1, TGFb-R3 and BMP-9, higher angiokine levels were associated with worse PFS (HR > 1) The univariate analysis showed that exception of HER-3, SDF-1, TGFb-R3, VEGF-R1, and VEGF-R2, higher angiokine levels were associated with worse OS (HR > 1) The multivariable analysis showed that HGF, OPN, TIMP-1, TSP-2, and VCAM-1 were independently associated with OS with HRs of 1.5, 1.3, 1.7, 1.8, and 2.2, respectively None of the angiokines are statistically significant to predictive benefit for patients receiving either sunitinib or everolimus	[122]
	Spain	2021	CXCL10, CXCL11, HGF, IL-6	Serum	The Luminex’s xMAP Technology multiplex system	51 mRCC patients with sunitinib therapy 4 mRCC patients with pazopanib therapy 5 mRCC patients with both	Predicting patients' survivals with anti-angiogenic therapy	• High basal HGF levels (over 649.1 pg/mL) were significantly correlated to worse PFS (P = 0.003) and OS (P = 0.0034). A reduction of HGF levels during the treatment was related to a lower PFS (p = 0.017), but not with OS • CXCL11 levels were significantly higher in patients who did not respond to treatment (P < 0.05) than in those who responded. High levels of CXCL11 (above 39.4 pg/mL) were significantly related to shorter PFS (P = 0.0003) and OS (P = 0.001). Patients with a reduction of CXCL11 levels after 3 months treatment had a significantly lower OS (P = 0.027), but not PFS	[192]
	France	2021	SAA2, CFB	Plasma	Elisa	59 mRCC patients with sunitinib or bevacizumab treatment	Predicting survivals of mRCC patients with TKI therapy	• The levels of SAA2 and CFB can subdivide the cohort with anti-angiogenic treatment into 3 different groups according to PFS and OS: CFB low/SAA2 low (PFS 13.23 months, OS: 20.8 months), CFB low/SAA2 high or CFB high/ SAA2 low (PFS 9.87 months, OS 16.52 months), and CFB high/SAA2 high (PFS 2.8 months, OS 8.33 months)	[186]
	USA	2021	30 cytokines, including IL-6, IL-1RA, CSF, IFN-γ, IL-12, VEGF etc	Plasma	Luminex FLEXMAP 3D System	33 mccRCC patients with immunotherapy 23 mccRCC patients with TKI therapy	Predicting and monitoring response to ICI therapy in mRCC patients	• 17 patients are identified in clinical benefits (CB) group and 16 patients are identified in non-clinical benefits (NCB) group • No significant cytokine differences were observed between CB and NCB patients	[195]
							Predicting and monitoring response to TKI therapy in mRCC patients	• 13 patients are identified in clinical benefits CB group and 10 patients are identified in non-clinical benefits NCB group • Patients in CB group had lower median levels of IL-6 (8.4 vs 13.5 pg/mL, p = 0.02), IL-1RA (178 vs 248 pg/mL, p = 0.03), and G-CSF (23.9 vs 38.3 pg/mL, p = 0.02) compared with patients in NCB group in pretreatment
	USA	2020	Ang-1, Ang-2, HGF, CXCL10, IL2, IL6, IL8, IL10, CXCL9, NGAL, OPN, TGFb, VEGF etc	Serum	Luminex instrument	52 advanced RCC patients with combined axitinib and pembrolizumab treatment	Predicting the survival and treatment benefit of RCC therapy	• Higher baseline of CXCL10 was correlated with objective response rate (ORR) (unadjusted P value = 0.0197) • Lower EOT level of CEACAM1, GRO-a, HGF, and TIMP-1 was correlated with objective response rate (ORR) (unadjusted P value = 0.0026, 0.0495, 0.0112, 0.0044, respectively) • At baseline, CEACAM1 levels ≥ median were associated with better PFS (P = 0.085). C2D1, GRO-a and HGF levels were associated with better PFS (P = 0.0034, P < 0.001, respectively). At EOT, HGF and TIMP-1 levels < median were associated with better PFS (P = 0.0034, 0.014, respectively)	[194]
	Italy	2020	PD-1, PD-L1, BTN3A1	Plasma	Elisa	Testing cohort: 21 mccRCC patients with nivolumab treatment Validating cohort: 20 mccRCC patients with nivolumab treatment 15 localized ccRCC patients	Predicting survivals of mccRCC patients with immunotherapy	• The mean pre-treatment levels of plasma ICs in long-responder group (> 18 months) was significantly higher than all patients (PD-1: 13.25 vs. 2.00 ng/mL, p = 0.01; PD-L1: 1.09 vs. 0.64 ng/mL, p = 0.02; BTN3A1: 11.03 vs. 6.84 ng/mL, p = 0.03) • High level of plasma PD-1 (> 2.11 ng/mL), PD-L1 (> 0.66 ng/mL) and BTN3A1 (> 6.84 ng/mL) were correlated to a shorter median PFS (PD-1: 20.7 vs. 6.9 months, P < 0.0001; PD-L1: 19 vs. 9 months, P < 0.0001; BTN3A1: 17.5 vs. 8.4 months, p = 0.002) • After 18 months of immunotherapy, plasma PD-1 and PD-L1 levels were lower than baseline in patients with FPS longer than 18 months (PD1: 1.23 vs. 13.25 ng/mL; PD-L1: 0.73 vs. 1.09 ng/mL) • The predictive value of PD1, PD-L1 and BTN3A1 in validating cohort: PD1: AUC = 1.0, P < 0.001; PD-L1: AUC = 0.944, P < 0.001; BTN3A1: AUC = 0.833, P < 0.03	[187]
Exosome	China	2018	Exsomal miRNA miR-210, miR-1233	Serum	Total exosome isolation reagent, EpCAM isolation beads, Flow cytometry	10 ccRCC patients received surgical tumor removal	Monitoring responses to surgery	• The levels of exosomal miR-210 and miR-1233 were significantly lower in postoperative than in preoperative samples (p = 0.004, p = 0.008, respectively)	[153]