Paper - Review

DOI: 10.1016/j.cell.2020.05.038

Summary

HCC
← Hepatocellular carcinoma
→ aggressive maligancy
→ although (early detection & surveillance) are suboptimal
∴ Profiling ← of the viral infection history

❗ The signature
→ identified cancer patient prior → to a clinical diagnosis
→ was superior to alpha-fetoprotein

Introduction

HCC
→ primary liver cancer
→ is considered → a virus-related malignancy

Viral hepatitis
→ causes inflammation & chronic liver diseases
→ may lead to 1⃣ fibrosis 2⃣ cirrhosis 3⃣ HCC

Effective strategy
← to prevent HCC
→ to eliminate (causative factors)

❓: Control of HBV infection
→ has been remarkably successful → for decades
∴ Early detection remains a key approach → to prevent HCC-inflicted mortality

Current methods
← biannual surveillance ← using ultrasound
→ yielded mixed results related 1⃣ to the effectiveness in detecting HCC 2⃣ to providing survival benefit

Viruses are known → to effect human health
← by altering host immunity
→ crucial in the pathogenesis of human chronic diseases

Viruses
→ are cleared in the host → may leave unique molecular footprints
→ may serve as → an excellent window of early onset of disease

❓: Unique viral exposure signatures
← resulting from virus-host interactions
→ could reflect a cascade of events ← which may alter the risk of developing HCC

Results

The landscape of Viral Exposure Profiles

VirScan
→ apply a (phage display library) ← covers 93,304 viral epitopes
→ to screen for previous (exposure history)

A case-control design (← of the Maryland cohort)
→ for the discovery of viral exposure profiles

Total of 30,033 (viral epitopes)
→ were significantly enriched
→ was determined ← based on both replicates

Composition ← of the viral types
← at the viral taxonomic level
→ showed yet small noticeable differences

Number ← of (detected viral infections)
→ increases initially with the sample size
→ reaches saturation → at a sample size of 200 ↑

Number of (viral species)
→ was similarly distributed
→ among 1⃣ PC (Population control) 2⃣ AR (CLD patients as at risk) 3⃣ HCC
∴ NO ❌ bias (in the landscape of overall viral exposure profiles) ← between these groups

Detected → a wide range of unique (viral epitopes) → for each (viral species)
← were recognized among different participants
∴ (B-cell antigenticity) to (same viral species) → is rather diverse
← among the participants

Compare VirScan results ← with available medical chart entries
← for 1⃣ HCV 2⃣ HBV 3⃣ HIV
∴ VirScan had 0.45, 0.47, 0.70 specificity
❗Sensitivity was donw
∵ A majority of (viral status data) ← from medical charts
→ are unknown & missing

HCC-Associated Viral Exposure Signature (VES)

A gradient boosting approach is applied
→ to search for the best-fit virus composition
→ which can discriminate HCC from PC

VES → can significantly discriminate HCC from PC
→ with AUC values 1⃣ 0.9 for training 2⃣ 0.7 for cross-validation

Total (← of HCV strains)
→ were among (the main contributing viruses) in the signature
43 viruses
→ were preferentially depleted ← in the HCC group

A significant increase ← of the VES score
← among PC & AR & HCC
∴ VES was positively linked → to hepato-carcinogenesis

Phylogenetic analysis was performed
← of 61 viral strains
→ to determine similarity → among these HCC-related viruses

These viruses can be divided → into eight main branches
← where different HCV epitopes are clustered together

NO ❌ clear enrichment ← within each branch
→ for increased & decreased viruses
∴ Varying viral epitopes (← involved in immunoreactivity) → are commonly shared among HCC

Compared AR to HCC
← using the same gradient-boosting approach
→ to avoid this confounding variable
∵ A majority of HCC patients → has evidence of CLDs

Phenotype-Genotype Association with VES

Performing a GWAS (Genome-wide association study)
→ to determine → if the (host genetic background) may be linked to VES

Performing (an association test) → for all the remaining SNPs
∵ 1⃣ Following the removal of allelic SNPs
2⃣ the ones (← which deviate away) from Hardy-Weinberg equilibrium

Determining → whether there was an association
← between an SNP & HCV infection
∵ to further assess (the quality of our GWAS data)

SNP associated ← with 375 epitopes abundances (← of HCV genotype 2 & 3)
∴ CC allele is associated with 1⃣ a decreased abundance ← of core epitopes 2⃣ an increased abundance of NS5B epitopes in HCV genome

Validation of VES in a Prospective HCC Cohort

Analyzing VirScan profiles
← in the at-risk NIDDK cohort ← for HCC
→ to further validate → the VES identified above → for its clinical utility

This cohort ← consisted of 173 CLD patients
← who were enrolled ← for a natural history study of (liver disease)
→ 44 individual developed HCC

This cohort → contained serum samples
← collected at enrollment & at various follow-up time

Observing → VES scores varied substantially
→ among HCC cases in NCI-UMD cohorts

Results ← from univariable & multivariable Cox model (survival analysis)
← on several clinico-pathologic variables
→ to clarify → the independent & additional prognostic value ← of VES

∴ Performance of VES
→ was superior to AFP
← which is a known HCC diagnostic marker ← used in clinical practice

Similar trend
← between the levels of VES & overall survival
← among 44 patients in the NIDDK cohort

Analyzing ← 104 cancer-free controls & 40 HCC cases
→ to assess the time-dependent performance of VES
→ to predict the onset of HCC

Censoring time was defined as the time difference
← between baseline and follow-up
← within the cancer-free control group

A multivariable Cox regression model
→ generated to predict the occurrence of HCC diagnosis
→ based on VES scores at base-line

Discussion

Detecting cancer ← at an early stage
→ may provide → an opportunity
← 1⃣ in achieving a cure 2⃣ improving cancer-related mortality

A conventional approach
→ is to develop biomarkers specific → for cancer cells
→ in order to determine cancer early diagnosis

A cautionary note
← using cancer gene panel in ctDNA
∵ its high (false positive rate) ← among healthy individuals

Complex etiological landscape → creates a dilemma
→ how to best design cancer-specific diagnostic panels effective
→ for early cancer detection

HCC
← is a unique malignancy
← most of the major causative etiologies

Defining biomarkers specific → for HCC cells has been challenging
∵ its complex genomic landscape ← with extensive (intratumor & intertumor) heterogeneities

How (a history of viral exposures) ← in an individual
→ is associated with its risk → to develop HCC

∴ Sensitive tool
→ which applicable to the HCC surveillance program
→ to improve early diagnosis

Advantage of a simple tool
← to profile serological samples
→ to link (an individual's history of viral infection & corresponding response to early onset of HCC)

VES
← can discriminate HCC → from at-risk & healthy individuals
→ was then validated ← using a prospective cohort of sequentially enrolled at-risk patients

1⃣ HCV appears → as a major etiological factor driving VES
2⃣ A set of viruses are enriched while many others

GWAS analysis revealed
→ several SNPs have a strong association ← with a history of viral infection

VES discovery was based ← on a case-control study design ←with 150 HCC cases
∴ Both (10-fold cross validation & 60/40 split approaches) → to test robustness of VES

VES could independently classify HCC
← in a prospective at-risk population
∴ Confirming the robustness of VES ← in HCC diagnosis

VES → as surveillance tool
→ to screen early onset of HCC

VirScan is based ← on the PhIP-seq technology
← which provides a powerful approach with 1⃣ high throughput 2⃣ low cost

HCV antigen reactivity → is largely overlapped
← with the predicted antigenicity score
← by the B-cell epitope prediction method

Current surveys
→ underestimate the prevalence of HCV infection

(Sensitivity & Specificity) ← of HBV detection ← by VirScan
→ are relatively low

---