Hee Sang Hwang1, Jihyun An2, Hyo Jeong Kang1,3, Bora Oh4, Yoo Jin Oh4, Ji-Hye Oh4, Wonkyung Kim4, Chang Ohk Sung5,6, Ju Hyun Shim7,8, Eunsil Yu9,10. 1. Department of Pathology, Asan Medical Center, University of Ulsan College of Medicine, Seoul, Republic of Korea. 2. Department of Gastroenterology and Hepatology, Hanyang University College of Medicine, Guri, Gyeonggi, Republic of Korea. 3. Asan Liver Center, Asan Medical Center, University of Ulsan College of Medicine, Seoul, Republic of Korea. 4. Asan Institute for Life Science, Asan Medical Center, Seoul, Republic of Korea. 5. Department of Pathology, Asan Medical Center, University of Ulsan College of Medicine, Seoul, Republic of Korea. tooearly@hanmail.net. 6. Center for Cancer Genome Discovery, Asan Institute for Life Science, Asan Medical Center, University of Ulsan College of Medicine, Seoul, Republic of Korea. tooearly@hanmail.net. 7. Asan Liver Center, Asan Medical Center, University of Ulsan College of Medicine, Seoul, Republic of Korea. s5854@amc.seoul.kr. 8. Department of Gastroenterology, Asan Medical Center, University of Ulsan College of Medicine, Seoul, Republic of Korea. s5854@amc.seoul.kr. 9. Department of Pathology, Asan Medical Center, University of Ulsan College of Medicine, Seoul, Republic of Korea. d890075@gmail.com. 10. Asan Liver Center, Asan Medical Center, University of Ulsan College of Medicine, Seoul, Republic of Korea. d890075@gmail.com.
Abstract
BACKGROUND: Although hepatocellular carcinomas (HCCs) often recur in patients undergoing hepatectomy, there are no reliable biomarkers of this undesirable event. Recent RNA-based efforts have developed valuable genetic indices prognostic of cancer outcomes. We aimed to identify molecular predictors of early recurrence after resection of HCC, and reveal the genomolecular structure of the resected tumors. METHOD: Based on the transcriptomic and genomic datasets of 206 HCC samples surgically resected in the Asan Medical Center (AMC), we performed a differential gene expression analysis to identify quantitative markers associated with early recurrence and used the unsupervised clustering method to classify genomolecular subtypes. RESULTS: Differential gene expression profiling revealed that S100P was the highest-ranked overexpressed gene in HCCs that recurred within 2 years of surgery. This trend was reproduced in immunohistochemical studies of the original cohort and an independent AMC cohort. S100P expression also independently predicted HCC-specific mortality post-resection (adjusted hazard ratio 1.09, 95% confidence interval 1.01-1.19; p = 0.042). Validation in a Chinese cohort and in in vitro experiments confirmed the prognostic value of S100P in HCC. We further identified five discrete molecular subtypes of HCC; a subtype with stem cell features ('AMC-C4') was associated with the worst prognosis, both in our series and another two Asian datasets, and S100P was most strongly upregulated in that subtype. CONCLUSION: We identified a promising prognostic biomolecule, S100P, associated with early recurrence after HCC resection, and established the genomolecular architecture of tumors affecting clinical outcomes, particularly in Asian patients. These new insights into molecular mediators should contribute to effective care for affected patients.
BACKGROUND: Although hepatocellular carcinomas (HCCs) often recur in patients undergoing hepatectomy, there are no reliable biomarkers of this undesirable event. Recent RNA-based efforts have developed valuable genetic indices prognostic of cancer outcomes. We aimed to identify molecular predictors of early recurrence after resection of HCC, and reveal the genomolecular structure of the resected tumors. METHOD: Based on the transcriptomic and genomic datasets of 206 HCC samples surgically resected in the Asan Medical Center (AMC), we performed a differential gene expression analysis to identify quantitative markers associated with early recurrence and used the unsupervised clustering method to classify genomolecular subtypes. RESULTS: Differential gene expression profiling revealed that S100P was the highest-ranked overexpressed gene in HCCs that recurred within 2 years of surgery. This trend was reproduced in immunohistochemical studies of the original cohort and an independent AMC cohort. S100P expression also independently predicted HCC-specific mortality post-resection (adjusted hazard ratio 1.09, 95% confidence interval 1.01-1.19; p = 0.042). Validation in a Chinese cohort and in in vitro experiments confirmed the prognostic value of S100P in HCC. We further identified five discrete molecular subtypes of HCC; a subtype with stem cell features ('AMC-C4') was associated with the worst prognosis, both in our series and another two Asian datasets, and S100P was most strongly upregulated in that subtype. CONCLUSION: We identified a promising prognostic biomolecule, S100P, associated with early recurrence after HCC resection, and established the genomolecular architecture of tumors affecting clinical outcomes, particularly in Asian patients. These new insights into molecular mediators should contribute to effective care for affected patients.