Background
Materials and methods
Patient recruitment and sample preparation
RNA-seq data
Differential gene expression analysis
Gene set enrichment analysis with MSigDB gene sets
Co-clustering & survival analysis
Statistical testing
TCGA-LIHC RNA-seq data
Data visualization
R & Python packages and libraries versions
Assessment of fibrosis, steatosis and microvascular invasion
Data preparation for MLP classification
MLP classification model
Classification and prediction results
Results
Multi-region sequencing allows the identification of patient-specific dysregulated genes
Aggregation of per-patient analysis captures most conventional differential expression
Conventional all-patients analysis omits HCC subgroup-specific genes due to high inter-individual variability
Patient-specific differentially expressed genes with high DEPC scores are associated with metabolism, proliferation, and known cancer modules
Patient-specific transcriptomic profiles reveal novel HCC patient subgroups with a strong correlation to recurrence
Variable | PG0 | PG1 | P-value | significance | |
---|---|---|---|---|---|
N | 42 | 36 | |||
Sex | Female | 4 | 10 | 0.043 | * |
Male | 38 | 26 | |||
Ethnicity | Chinese | 31 | 20 | 0.363 | |
Filipino | 3 | 1 | |||
Indian | 1 | 2 | |||
Indonesian | 1 | 1 | |||
Malay | 2 | 2 | |||
Thai | 2 | 7 | |||
Others | 2 | 3 | |||
Significant Alcohol Consumption | Yes | 11 | 9 | 0.683 | |
No | 23 | 17 | |||
Unknown | 8 | 10 | |||
Child's Pugh score | A | 41 | 36 | 1.000 | |
B | 1 | 0 | |||
Diabetes | Yes | 17 | 12 | 0.636 | |
No | 25 | 24 | |||
Tumour Multiplicity | Yes | 7 | 8 | 0.588 | |
No | 35 | 28 | |||
Fibrosis Stage | 0 | 14 | 6 | 0.010* | ** |
1 | 3 | 4 | |||
2 | 1 | 9 | |||
3 | 11 | 4 | |||
4 | 9 | 11 | |||
Microvascular Invasion | Yes | 12 | 17 | 0.110 | |
No | 30 | 19 | |||
Edmondson Grade | 1 | 5 | 2 | 0.058 | |
2 | 26 | 15 | |||
3 | 11 | 17 | |||
4 | 0 | 2 | |||
Steatosis | 0–5% | 20 | 23 | 0.268 | |
5–33% | 14 | 11 | |||
33–66% | 3 | 0 | |||
Overall Survival | Alive | 36 | 25 | 0.110 | |
Dead | 6 | 11 | |||
Tumour Stage TNM V8 | I | 25 | 17 | 0.216 | |
II | 13 | 10 | |||
III | 4 | 9 | |||
Recurrence status | Yes | 24 | 14 | 0.125 | |
No | 18 | 22 | |||
HBV Status | positive | 23 | 29 | 0.030 | * |
negative | 19 | 7 | |||
HCV Status | positive | 4 | 2 | 0.681 | |
negative | 38 | 34 | |||
Max. Tumour Diameter (cm) | 6.84 ± 4.94 | 6.35 ± 3.93 | 0.876 | ||
Albumin (g/L) | 40.95 ± 4.32 | 41.3 ± 3.55 | 0.751 | ||
Bilirubin (umol/L) | 13.73 ± 4.57 | 13.12 ± 5.68 | 0.348 | ||
AST (U/L) | 50.65 ± 36.2 | 50.23 ± 52.29 | 0.854 | ||
ALT (U/L) | 50.47 ± 53.97 | 33.86 ± 19.92 | 0.196 | ||
Alkaline Phosphatase (U/L) | 108.85 ± 53.44 | 125.41 ± 121.7 | 0.943 | ||
Prothrombin Time (secs) | 10.91 ± 0.96 | 11.55 ± 1.35 | 0.025 | * | |
Platelets (× 10^9) | 232.82 ± 89.59 | 239.14 ± 71.48 | 0.344 | ||
AFP (ng/ml) | 1925.1 ± 9012.17 | 4362.23 ± 11,826.79 | 0.023 | * | |
Recurrence-free survival days | 675.76 ± 462.37 | 521.72 ± 433.32 | 0.132 |
Overall differences in the expression profiles between the patient groups with different recurrence rates
Patient-specific gene dysregulation may explain the low response rates of current HCC therapeutic treatments
Gene | Down-regulated DEPC | Up-regulated DEPC | Down-regulated in all-patients analysis | Up-regulated in all-patients analysis | Targeted by |
---|---|---|---|---|---|
VEGFR2/KDR | 15 | 11 | No | No | Sorafenib, Lenvatinib, Ramucirumab, Cabozantinib |
VEGFR1/FLT1 | 7 | 22 | No | No | Sorafenib, Lenvatinib, Cabozantinib |
VEGFR3/FLT4 | 14 | 3 | No | No | Sorafenib, Lenvatinib, Cabozantinib |
FLT3 | 11 | 5 | Yes | No | Sorafenib |
PDGFRA | 37 | 0 | Yes | No | Lenvatinib |
PDGFRB | 3 | 21 | No | No | Sorafenib |
FGFR1 | 31 | 2 | Yes | No | Lenvatinib |
FGFR2 | 34 | 7 | Yes | No | Lenvatinib |
FGFR3 | 20 | 6 | No | No | Lenvatinib |
FGFR4 | 4 | 14 | No | No | Lenvatinib |
KIT | 1 | 9 | No | No | Sorafenib, Lenvatinib |
RET | 28 | 2 | Yes | No | Sorafenib, Lenvatinib, Cabozantinib |
AXL | 32 | 0 | Yes | No | Cabozantinib |
MET | 3 | 10 | No | No | Cabozantinib |
BRAF | 1 | 5 | No | No | Sorafenib |
RAF1 | 1 | 1 | No | No | Sorafenib |