Fat Attenuation Index of Renal Cell Carcinoma Reveals Biological Characteristics and Survival Outcome

Figures

• 

  Figure 1. The study design, inclusion criteria, and participants. (A) The validation cohorts were composed by patients from the FUSCC database (FAI cohort A) and TCGA/TCIA database (FAI cohort B). We excluded patients who did not have CT scans and those who had extreme values of VFA. The final included patients were 129 from the FUSCC cohort and 218 from the TCGA/TCIA cohort. (B) The imaging genomics cohort prospectively included consecutive unselected patients from FUSCC who had nephrectomy or nephron sparing surgery. After excluding 8 patients who were not clear cell RCC and 2 patients with unqualified tissue samples, the final cohort included 19 patients. (C) The treatment response cohort included patients from a FUSCC cytoreductive surgery database. All patients had CT scans of the kidney tumor before and after two months of neoadjuvant/palliative targeted therapy, and then they undergone nephrectomies.

• 

  Figure 2. Kaplan–Meier curves of patients with different FAIPTAT. Fat attenuation index (FAI) was calculated and grouped according to the median FAT value of the FUSCC and TCGA cohorts. (A) Overall survival of the TCGA cohort, the 5-year median overall survival was 64 months for high FAIPTAT group and not reached for low FAIPTAT group (HR = 2.24, p <0.001). (B) Overall survival of the FUSCC cohort, the 5-year median overall survival was not reached for both groups in FUSCC cohort (HR = 2.28, p = 0.01). The 3-year survival rate was 88%/93% for high FAIPTAT and low FAIPTAT groups in FUSCC cohort (HR = 1.81) and 72%/93% in TCGA cohort (HR = 5.65).

• 

  Figure 3. Pathway alterations in the tumor tissue from the imaging genomics cohort. (A) KEGG enrichment analysis of annotated different expressed gene was performed in the imaging genomics cohort. Pathway analysis showed that the neuroactive ligand–receptor interaction, cytokine–cytokine receptor interaction, complement and coagulation cascades were altered in high FAIPTAT patients compared with low FAIPTAT patients. (B–E) The PPI networks were drawn using Cytoscape and key genes from those pathways such as GCGR, LPAR3, NMUR2, EGF, KIT, KRAS, CXCL8, IL-6, GCGR, FGA, and FGB were observed to be altered. (F) Pathway analysis showed that the complement and coagulation cascades, cytokine–cytokine receptor interaction, and multiple metabolic pathways were altered in high FAIPTAT patients compared with low FAIPTAT patients. (G–H) The PPI networks were drawn using Cytoscape, key genes such as ITGAM, FGA, and KNG1 were observed for the complement and coagulation cascades, while alterations of IL-6, TNF, CXCL8, IL-19, CXCL2, IL-1B, etc. were observed for the cytokine pathway.

• 

  Figure 4. The immune microenvironment of tumor tissue of the imaging genomics cohort (FUSCC) and the FAI cohort B (TCGA). Immune deconvolution showed good consistency between the both cohort. (A) Tumors of high FAIPTAT patients showed enhanced infiltration of Treg, NK, CD56bright cells, Tem cells, and macrophages, and a reduced infiltration of pDC, Tcm cells, Th17 cells, and mast cells, as well as a lower angiogenesis level compared with low FAIPTAT patients. (B) Immune deconvolution showed an enhanced infiltration of macrophages, pDC cells, Th1 cells, and DC cells, and a reduced infiltration of Th17 cells, Tcm cells, Treg cells, CD8+ T cells, and NK cells, as well as a lower angiogenesis level in peri-tumor fat. (C) Hypoxia score and immune infiltration score were enhanced in fat of high FAIPTAT patients.

• 

  Figure 5. Dynamic changes of FAIPTAT after targeted therapy. (A) Scattergram showed that FAIPTAT changes synchronously with tumor size after targeted therapy. (B) Waterfall plot showed that patients with high FAIPTAT had short PFS during targeted therapy.