Tumor-resident microorganisms as clinical biomarkers in primary liver cancer: A systematic review of current evidence

Table 1 Newcastle-Ottawa Scale quality assessment of included studies

Ref.	Selection (0-4)	Comparability (0-2)	Outcome (0-3)	Total score
Chai et al[31]	4	2	3	9
Chakladar et al[20]	4	1	3	8
Xue et al[8]	4	2	1	7
Chen et al[15]	4	1	2	7
He et al[21]	4	2	3	9
Huang et al[22]	4	2	3	9
Jiang et al[16]	4	2	3	9
Komiyama et al[32]	4	2	3	9
Liu et al[24]	3	1	2	6
Liu et al[23]	4	2	3	9
Lu et al[25]	4	2	3	9
Qu et al[33]	4	2	3	9
Schulz et al[26]	4	1	3	8
Shen et al[27]	4	2	3	9
Song et al[17]	3	1	3	7
Sun et al[28]	4	2	3	9
Sun et al[29]	3	2	1	6
Wang et al[11]	4	2	3	9
Xin et al[34]	4	2	3	9
Yang et al[30]	4	2	2	8

Selection (0-4 points): Representativeness of exposed cohort, selection of non-exposed cohort, ascertainment of exposure, demonstration that outcome was not present at start of study. Comparability (0-2 points): Comparability of cohorts based on design or analysis. Outcome (0-3 points): Assessment of outcome, adequacy of follow-up length, adequacy of follow-up completeness.

Table 2 Study characteristics and methodological comparison of included studies

Ref.	Country	Study design	Sample size	Cancer type	Methodology	Target region	Key microbial taxa	Primary outcome	AUC/HR (95%CI)	Validation method
Chai et al[31]	China	Cross-sectional	52 patients (94 samples)	ICC	16S rRNA + FISH + culture + TEM	V3-V4	Pseudomonas fungorum (anti-tumor), K. pneumoniae, S. capitis	Diagnostic/therapeutic biomarker	Negative correlation with CA19-9	FISH, culture, TEM, scRNA-seq
Chakladar et al[20]	United States (TCGA)	Cross-sectional	423 samples (373T + 50N)	HCC	Pathoscope 20 from RNA-seq	Whole transcriptome	Various by HBV/alcohol status	Prognostic associations	Multiple associations	Computational inference
Xue et al[8]	China	Cross-sectional	47 pairs	HCC	16S rRNA + multi-omics	V3-V4	Oscillospira, Mucispirillum, Helicobacter	Diagnostic biomarker	Not specified	Multi-omics integration
Chen et al[15]	China	Cross-sectional	72 pairs	HCC	16S rRNA + qRT-PCR	V4	Propionibacterium, Mycoplasma, Pseudomonas	DNA methylation correlation	Survival associations (TCGA)	qRT-PCR validation
He et al[21]	China	Cross-sectional	99 pairs	HCC	16S rRNA + FISH	V3-V4	Proteobacteria, Actinobacteria, Firmicutes	Diagnostic biomarker	Not specified	FISH validation
Huang et al[22]	China	Cross-sectional	100 patients (168 samples)	HCC	16S rRNA + FISH + culture + qPCR	V3-V4	Bacilli, γ-Proteobacteria	Diagnostic biomarker	Training: 1.00; Validation: 0.943	FISH, culture, qPCR
Jiang et al[16]	China	Cross-sectional	172 patients (344 samples)	HCC	16S rRNA + FISH + multi IF	V3-V4	Intestinimonas, Brachybacterium, Rothia	Prognostic biomarker	Independent prognostic factors	FISH, multi-IF
Komiyama et al[32]	Japan	Cross-sectional	65 patients	HCC + ICC	16S rRNA + qPCR	V3-V4	R. gnavus (viral HCC marker)	Diagnostic biomarker	Not specified	qPCR validation
Liu et al[24]	China	Cross-sectional	18 patients	HCC	16S rRNA + culture + animal model	V4	Rhodococcus sp. B513	Therapeutic target	Angiogenesis promotion	Culture, animal validation
Liu et al[23]	China	Cross-sectional	107 + 74 samples	HCC	16S rRNA + FISH + qPCR + culture	V4	S. maltophilia	Diagnostic biomarker	AUC: 0.739-0.78	FISH, qPCR, culture
Lu et al[25]	China	Cross-sectional	58 patients (300 samples)	Multifocal HCC	16S rRNA + multi-omics	V3-V4	IM vs MO specific taxa	Diagnostic biomarker	AUC: 0.795	Multi-omics integration
Qu et al[33]	China	Cross-sectional	28 patients (56 samples)	PLC (HCC + ICC + cHCC - CCA)	16S rRNA (FFPE)	V4	Pseudomonadaceae	Prognostic biomarker	Linear correlation with survival	FFPE validation
Schulz et al[26]	Europe + Turkey	Multi-center	20 patients (23 FFPE blocks)	HCC	16S rRNA + IHC (FFPE)	V1-V2	S. aureus, B. parvula, A. chinensis	Prognostic biomarker	Survival associations	IHC validation
Shen et al[27]	China	TCGA analysis	TCGA cohort	HCC	Computational mycobiome analysis	Whole genome	Malassezia	Prognostic biomarker	5-gene model	Multi-cohort validation
Song et al[17]	Global (TCGA)	Cross-sectional	352 patients	HCC	Computational from RNA-seq	Whole transcriptome	27 microbial genera	Prognostic biomarker	AUC: 0.849-0.822	TCGA validation
Sun et al[28]	China	Cross-sectional	91 patients (169 samples)	HCC	16S rRNA + FISH	V3-V4	Proteobacteria, Actinobacteria	Prognostic biomarker	HR: 0.296 (0.126-0.699)	FISH validation
Sun et al[29]	China	Cross-sectional	30 pairs	HCC	16S rRNA	V3-V4	Descriptive analysis	Descriptive study	Not applicable	None
Wang et al[11]	China	Cross-sectional + animal	127 patients + animal models	HCC	16S rRNA + shotgun + culture + FISH + qPCR	V3-V4	K. pneumoniae	Mechanistic/causative	Causal relationship	Multiple validations + animal models
Xin et al[34]	China	Cross-sectional	121 patients	ICC	16S rRNA + FISH + qRT-PCR + IHC	V4	11-genus risk score	Prognostic biomarker	Independent prognostic factor	FISH, qRT-PCR, IHC
Yang et al[30]	China	Cross-sectional	925 clinical samples	HCC	16S rRNA + 5R 16S + shotgun + FISH + IHC	V3-V4 + 5 regions	Streptococcus species, Delftia acidovorans	Diagnostic biomarker	AUC: 0.9405-0.9811	Multiple validations

FISH: Fluorescence in situ hybridization; IHC: Immunohistochemistry; HCC: Hepatocellular carcinoma; TCGA: The Cancer Genome Atlas; qRT-PCR: Quantitative Real-time Polymerase Chain Reaction; AUC: Area under the curve; HBV: Hepatitis B virus.