Return To Abstract Listing

Gastrointestinal Imaging

E2845. Decoding of The PNPLA3 Gene Phenotype Using Noninvasive Imaging Quantitative Radiomics Approach

Khalaf A1,2,  Fuentes D1,  Morshid A1,  Hazle J1,  Hassan M1,  Elsayes K.1 1. The University of Texas MD Anderson Cancer Center, Houston, Texas; 2. The University Of Alabama at Birmingham, Birmingham, Alabama

Address correspondence to A. Khalaf (amkhalaf@uab.edu)

Objective: To establishnon-invasive quantitative imaging biomarkersthat are able to differentiate betweenthedifferent genetic variations of the Patatin-Like Phospholipase Domain Containing 3 (PNPLA3) gene and correlate these imaging features with the overall survival in patients with hepatocellular carcinoma (HCC) to uncover their prognostic power.

Materials and Methods: This IRB-approved, HIPAA compliant, retrospectivesingle-institution study included 161patients,men (n=120; mean age, 63; range, 21-88), women (n=41; mean age, 63; range, 34-82) with pathologically confirmed HCC. It included pathologically confirmed PNPLA3 patients (n=79) with heterozygous allele GG, patients (n=19) with heterozygous allele CG and patients (n=63) with heterozygous allele GC. The exclusion criteria were the presence of other types of primary liver cancer and concurrent or past history of cancer at another organ site.Image intensity, shape, and co-occurrence features were computed within the viable, necrotic, and background liver parenchyma extracted from pre-therapeutic multiphasic contrast-enhanced CT at the baseline.KruskalWallis test was conducted to compare quantitative imaging features of the different liver tissues (enhanced, non-enhanced and residual liver parenchyma) between the three genotype groups. Kaplan Meier analysis was performed to compare the overall survival (OS) of the patients according to the genetic variation of their PNPLA3 genotype.

Results: The OS of HCC patients with the GG genotype was shorter than that of HCC patients with the CC/CG genotypes. The estimated median survival was 16.8 months (95% CI, 9.9–23.7 months) for HCC cases with the GG genotype and 25.9 months (95% CI, 21.5–30.3 months) for cases with CC/CG genotypes (P = 0.1). Statistically significant differences (p<.05) within the genotypes was seen for imaging features representative of the inherent tissue Hounsfield values and heterogeneity of contrast uptake on triphasic CT.

Conclusion: This preliminary study demonstrates the feasibility of radiomics to identify a general prognostic genotype existing in HCC patients and a clear association with the clinical outcome (OS) of these patients. Since imaging is routinely done in clinical practice for this patient population, this may provide a predictive and prognostic tool that may have a clinical impact in improving decision-support in the therapeutic plan of HCC with lower expenses.Continual efforts to validate these features in matched control patient cohorts are likely to provide useful information to assist in the current patient staging methods (TNM, BCLC, CLIP).