Hong-Dar Isaac Wu

Prognostic Implications of Epidermal Growth Factor Receptor and KRAS Gene Mutations and Epidermal Growth Factor Receptor Gene Copy Numbers in Patients with Surgically Resectable Non-small Cell Lung Cancer in Taiwan

Introduction: The prognostic role of epidermal growth factor receptor (EGFR) mutations in patients with surgically resectable non-small cell lung cancer (NSCLC) without EGFR tyrosine kinase inhibitor treatment has not been well established, because the reports are still few. Materials and Methods: We analyzed the survival data of 164 patients with surgically resectable (stages I to IIIA) NSCLC of two year groups (1996–1998 and 2002–2004), and compared with EGFR mutations, KRAS mutations, and EGFR gene copy numbers. Results: Comparing the survival of wild-type patients and patients having L858R mutations or exon 19 deletion, the median survival was much longer for patient with EGFR mutations (54.7 months) than wild type (34.9 months). The difference was not statistically significant by univariate analysis (p = 0.1981) but had borderline significance by multivariate analyses (p = 0.0506). In addition, the 3-year survival rates of patients with EGFR mutations were also significantly higher than wild type (p = 0.0232). After exclusion of 18 patients treated by EGFR-tyrosine kinase inhibitor for tumor recurrence, the trends were still the same. Patients with KRAS mutations had shorter median survival (21 months) than wild type (44.4 months). Patients with EGFR polysomy (≧copies) also had longer median survival (56.2 months) than wild type (53.4 months). But the survival differences of these two genetic markers were all not significant statistically. Conclusion: It is intriguing that patients with NSCLC with EGFR mutations had better survival than wild type. Such a tumor biology may confound the survival data in a study without the stratification by EGFR mutation.

Clinical Implications of High MET Gene Dosage in Non-Small Cell Lung Cancer Patients without Previous Tyrosine Kinase Inhibitor Treatment

Introduction: Recently, two studies revealed that MET amplification was associated with secondary epidermal growth factor receptor (EGFR) tyrosine kinase inhibitor (TKI) resistance in non-small cell lung cancer (NSCLC) patients. But it remains uncertain whether MET amplification could be related to primary TKI resistance in NSCLC because of limited data. Materials and Methods: MET gene dosage of the tumor tissues from 208 NSCLC patients was investigated by real time quantitative polymerase chain reaction and compared with molecular and clinical features, including EGFR mutations, KRAS mutations, EGFR gene copy numbers, and patient survivals. Three copies were used as the cutoff. Among them, 25 patients were also evaluable for EGFR TKI responsiveness. Results: The proportion of high MET gene dosage was 10.58% (22/208) with higher incidence in squamous cell carcinoma (11.86%) and smokers (16.18%), although the differences with adenocarcinoma and nonsmokers were nonsignificant. Coexisting EGFR mutations were identified, and the incidence (8.54%) was similar to wild type (12.0%). High MET gene dosage was significantly associated with higher tumor stage (stage I + II versus stage III + IV; p = 0.0254) and prior chemotherapy for stage III + IV adenocarcinoma patients (35.71% versus 7.41%; p = 0.0145) but not correlated with primary TKI resistance. Among the 155 surgically resectable patients (stage I to IIIA), high MET gene dosage was significantly associated with shorter median survival (21.0 months versus 47.1 months; p = 0.042) by univariate analysis. Conclusions: High MET gene dosage was not related to primary TKI resistance and the incidence was increased after chemotherapy, suggesting high MET gene dosage may also be related to chemotherapy resistance.

The Hepatitis Viral Status in Patients With Hepatocellular Carcinoma: a Study of 3843 Patients From Taiwan Liver Cancer Network.

AbstractHepatocellular carcinoma (HCC) is the leading cancer death in Taiwan. Chronic viral hepatitis infections have long been considered as the most important risk factors for HCC in Taiwan. The previously published reports were either carried out by individual investigators with small patient numbers or by large endemic studies with limited viral marker data. Through collaboration with 5 medical centers across Taiwan, Taiwan liver cancer network (TLCN) was established in 2005. All participating centers followed a standard protocol to recruit liver cancer patients along with their biosamples and clinical data. In addition, detailed viral marker analysis for hepatitis B virus (HBV) and hepatitis C virus (HCV) were also performed. This study included 3843 HCC patients with available blood samples in TLCN (recruited from November 2005 to April 2011). There were 2153 (56.02%) patients associated with HBV (HBV group); 969 (25.21%) with HCV (HCV group); 310 (8.07%) with both HBV and HCV (HBV+HCV group); and 411 (10.69%) were negative for both HBV and HCV (non-B non-C group). Two hundred two of the 2463 HBV patients (8.20%) were HBsAg(-), but HBV DNA (+). The age, gender, cirrhosis, viral titers, and viral genotypes were all significantly different between the above 4 groups of patients. The median age of the HBV group was the youngest, and the cirrhotic rate was lowest in the non-B non-C group (only 25%). This is the largest detailed viral hepatitis marker study for HCC patients in the English literatures. Our study provided novel data on the interaction of HBV and HCV in the HCC patients and also confirmed that the HCC database of TLCN is highly representative for Taiwan and an important resource for HCC research.

Identification of Transforming Hepatitis B Virus S Gene Nonsense Mutations Derived from Freely Replicative Viruses in Hepatocellular Carcinoma

Background & Aims The correlation between chronic hepatitis B virus (HBV) infection and hepatocellular carcinoma (HCC) has been well-established. But the roles of viral factor remain uncertain. Only HBV X gene and nonsense mutations of S gene (C-terminal truncation of HBV surface protein) have been demonstrated to have transforming activity. Whether they play a significant role in hepatocarcinogenesis is still uncertain. Methods Twenty-five HBV-related HCC patients were positive for hepatitis B core antigen (HBcAg) in the cancerous parts of their HCC liver tissues by immunohistochemistry studies, and had available tissue for whole HBV genome sequence analysis. The results were compared with 25 gender and age-matched HBcAg negative HCCs. Plasmids encoding HBV S gene nonsense mutations identified from HBcAg (+) HCC tissue were constructed to investigate their cell proliferation, transformation activity and the oncogenic potentials by xenograft study and in vivo migration assay. Results HBcAg (+) HCC patients were significantly associated with cirrhosis and small tumor size (≦2 cm) when compared with HBcAg (−) HCC patients. Southern blot analyses revealed freely replicative forms of HBV in the cancerous parts of HBcAg(+) HCC. Three nonsense mutations of S gene (sL95*, sW182*, and sL216*) were identified in the HBcAg(+) HCC tumor tissues. sW182* and sL216* were recurrently found in the 25 HBcAg (−) HCC tumor tissue, too. Functional studies of the above 3 non-sense mutations all demonstrated higher cell proliferation activities and transformation abilities than wild type S, especially sW182*. Tumorigenicity analysis by xenograft experiments and in vitro migration assay showed potent oncogenic activity of sW182* mutant. Conclusions This study has demonstrated potent oncogenic activity of nonsense mutations of HBV S gene, suggesting they may play an important role in hepatocarcinogenesis.

Validation of A Heteroscedastic Hazards Regression Model

A Cox-type regression model accommodating heteroscedasticity, with a power factor of the baseline cumulative hazard, is investigated for analyzing data with crossing hazards behavior. Since the approach of partial likelihood cannot eliminate the baseline hazard, an overidentified estimating equation (OEE) approach is introduced in the estimation procedure. Its by-product, a model checking statistic, is presented to test for the overall adequacy of the heteroscedastic model. Further, under the heteroscedastic model setting, we propose two statistics to test the proportional hazards assumption. Implementation of this model is illustrated in a data analysis of a cancer clinical trial.

Carboxypeptidase E is a prediction marker for tumor recurrence in early-stage hepatocellular carcinoma

Tumor recurrence and metastasis are the major causes of death for hepatocellular carcinoma (HCC) patients who are able to receive curative resection. Identifying the predicting biomarkers for tumor recurrence would improve their survival. RNA extracted from fresh frozen tumors and adjacent non-tumor liver tissues of 120 HCC patients were obtained from Taiwan Liver Cancer Network (TLCN) in year 2010 for determination of the carboxypeptidase E (CPE) expression level (including its splicing mutant CPE-ΔN) in the tumor tissue (T) and paired non-tumor liver tissue (N) by real-time quantitative polymerase chain reaction. All patients were male, had chronic hepatitis B virus infection, were in the early pathology stage, and received curative resection. The T/N ratio of the CPE expression level was correlated with the updated survival data from TLCN in 2015. The CPE expression level in the 120 HCC patients was divided into three groups according to the T/N ratio: <1, ≥1 and ≤2, and >2, respectively. By multivariate analyses, the recurrence-free survival (RFS) was only significantly associated with the pathology stage and the CPE expression level. For overall survival (OS), only the CPE expression level was the significant prognostic factor. The CPE expression level was also significantly correlated with the tumor recurrence for both stage I (p = 0.0106) and stage II patients (p = 0.0006). The CPE mRNA expression level in HCC can be a useful biomarker for predicting tumor recurrence in HCC patients who are in the early pathology stage and able to receive curative resection.

Effect of Ignoring Heterogeneity in Hazards Regression

We studied a heteroscedastic hazards regression (HHR) model in order to evaluate bias when intrinsic heterogeneity is ignored. We used the proportional hazards model and the associated estimate to show that bias resulting from neglecting the heterogeneity over different ‘treatment’ groups or other covariates is an analytically calculated value. We then compared the value with those obtained from simulation results. Further, considering that a ‘stratified’ proportional hazards (SPH) model could possibly be used as an alternative, we made similar calculations to clarify when the SPH model will succeed or fail in order to account for heterogeneity when the underlying model is the HHR model.

Overexpression of MutL homolog 1 and MutS homolog 2 proteins have reversed prognostic implications for stage I–II colon cancer patients

Background The outcome of colon cancer patients without lymph node metastasis is heterogeneous. Searching for new prognostic markers is warranted. Methods One hundred twenty stage I–II colon cancer patients who received complete surgical excision during 1995–2004 were selected for this biomarker study. Immunohistochemical method was used to assess p53, epidermal growth factor receptor, MLH1, and MSH2 status. KRAS mutation was examined by direct sequencing. Results Thirty three patients (27.5%) developed metachronous metastasis during follow up. By multivariate analysis, only female gender (p = 0.03), high serum carcinoembryonic antigen (CEA) level (≧5 ng/ml) (p = 0.04), and MLH1 overexpression (p = 0.003) were associated with the metastasis group. The 5-year-survival rate were also significantly lower for female gender (71.7% versus 88.9%, p = 0.025), high CEA level (64.9% versus 92.4%, p < 0.001), and MLH1 overexpression (77.5% versus 94.4%, p = 0.039). In contrast, MSH2 overexpression was associated with better survival, 95.1% versus 75.5% (p = 0.024). Conclusions The reversed prognostic implications in the overexpression of MLH1 and MSH2 for stage I–II colon cancer patients is a novel finding and worthy of further confirmation.

The Cox model and its applications

Introduction: Several Classical Data Examples for Survival Analysis.- Elements of Survival Analysis.- The Cox Proportional Hazards Model.- The AFT, GPH, LT, Frailty, and GLPH Models.- Cross-effect Models of Survival Functions.- The Simple Cross-effect Model.- Goodness-of-Fit for the Cox Model.- Remarks on Computations in Parametric and Semiparametric Estimation.- Cox Model for Degradation and Failure Time Data.- References.- Index.

Early radiographic response to epidermal growth factor receptor-tyrosine kinase inhibitor in non-small cell lung cancer patients with epidermal growth factor receptor mutations: A prospective study

Background: The time schedules for response evaluation of epidermal growth factor receptor-tyrosine kinase Inhibitor (EGFR-TKI) in non-small cell lung cancer (NSCLC) patients are still ill-defined. Methods: Stage IIIB/IV patients with histologically proven NSCLC were enrolled in this study if the tumor cells bore EGFR mutations other than T790M. Eligible patients were treated with either 250 mg of gefitinib or 150 mg of erlotinib once daily. The early response rate [computed tomography (CT) scan on Day 14], definitive response rate determined on Day 56, progression-free survival (PFS), overall survival (OS), and toxicity profile were assessed prospectively. Results: Thirty-nine patients were enrolled in this study. A total of 29 patients (29/39, 74.4%) achieved partial response (PR). Twenty-one patients (21/39, 53.8%) had early radiological response on Day 14. The early radiological response rate in patients with PR was 72.4% (21/29). Only eight patients without a PR on early CT still ended with PR. Among the 29 patients with PR, the PFS (8.1 months) and OS (18.3 months) of the 21 patients with early CT response were shorter than those of the 8 patients without early CT response (11.9 and 24.0 months for PFS and OS, respectively). But the survival differences were statistically non-significant. Conclusions: A very high percentage (72.4%, 21/29) of NSCLC patients with EGFR mutations with PR demonstrates early radiological response to EGFR-TKIs, which would advocate early radiological examination for EGFR-TKI therapy in NSCLC patients.