Search

1000 Background: In the DESTINY-Breast03 (NCT03529110) primary analysis (data cutoff [DCO], May 21, 2021), T-DXd showed superiority over T-DM1 in pts with HER2+ mBC, with a significant improvement of progression-free survival by blinded independent central review (HR, 0.284; 95% CI, 0.217-0.373; P < 0.001), and a safety profile consistent with prior studies. This analysis provides updated safety data with longer follow-up. Methods: Pts were randomized 1:1 to T-DXd or T-DM1. Prespecified safety analysis of treatment-emergent adverse events (TEAEs) was conducted; endpoints included time to event, duration of event, and resolution. Results: At DCO (September 7, 2021), 116 (45.1%) pts vs 39 (14.9%) pts remained on treatment in the T-DXd vs T-DM1 arms; median treatment duration was 16.1 mo (range, 0.7-33.0) for T-DXd vs 6.9 mo (range, 0.7-28.5) for T-DM1. Any-grade (G), G≥3, and serious AE (SAE) rates were similar for T-DXd vs T-DM1 (99.6% vs 95.4%; 53.3% vs 49.8%; and 21.0% vs 19.2%), while exposure-adjusted incidence rates (EAIRs; per pt-year) for G≥3 and SAEs were lower for T-DXd vs T-DM1 (0.42 vs 0.70 and 0.17 vs 0.27). Median time to TEAE associated with drug discontinuation or dose reduction was longer with T-DXd vs T-DM1 (224.0 vs 147.0 d and 96.0 v 19.0 d, respectively). Most TEAEs in ≥20% of pts were hematologic or gastrointestinal. Median time to first onset of select any-G TEAEs was 70.0 vs 42.0 d for anemia, 196.0 vs 168.0 d for lymphopenia, 132.0 vs 8.0 d for thrombocytopenia, 22.0 vs 24.0 d for fatigue, 74.5 vs 92.0 d for leukopenia, and 64.0 vs 105.0 d for neutropenia, with T-DXd vs T-DM1, respectively. In both arms, most nausea and vomiting events were G1/2; while G≥3 events with T-DXd vs T-DM1 were 6.6% vs 0.4% for nausea and 1.6% vs 0.8% for vomiting, respectively. Rates of nausea, vomiting, and alopecia were highest in cycle 1 and lower in subsequent cycles for T-DXd. Rates of hematologic events were generally lower in earlier cycles vs cycle ≥8 in both arms. Rates of adjudicated, drug-related ILD/pneumonitis were 10.9% (1 G2 event since previous DCO) with T-DXd vs 1.9% with T-DM1, with no G4/5 events. Median time to first adjudicated, drug-related ILD/pneumonitis event was 5.9 vs 9.5 mo for T-DXd vs T-DM1, respectively; at DCO, most events resolved (57.1% vs 80.0%), and follow-up is ongoing. Conclusions: In this updated safety analysis, T-DXd demonstrated a tolerable safety profile consistent with prior studies. Despite longer treatment duration with T-DXd, EAIRs of G≥3 and SAEs were lower for T-DXd vs T-DM1. Rates of ILD/pneumonitis for T-DXd were similar to those in the previous DCO. Nausea, vomiting, and alopecia rates decreased over time. This longer safety update reinforces the consistent safety profile of T-DXd, supporting the clinical benefit of T-DXd over T-DM1 in patients with HER2+ mBC. Clinical trial information: NCT03529110.

e24121 Background: The diagnosis of cancer can motivate patients to change their physical activity habits. No data has reported level of physical activity before and after breast cancer diagnosis in Chinese women. Methods: In an on-going prospective cohort study which involved 1462 Chinese women with early-stage breast cancer, a validated modified Chinese Baecke questionnaire was used to assess physical activity at baseline, 18-, 36- and 60-month after diagnosis. At baseline, patients recalled their habitual physical activity in the preceding 12 months before cancer diagnosis. At 18-, 36- and 60-month follow-up, patients reported their habitual physical activity over the previous 12 months. The level of physical activity at post-diagnosis was defined as the average value assessed at 18-, 36- and 60-month follow-up. Results: Breast cancer patients significantly increased level of physical activity, with median value of 0.6, 5.3, 4.4 and 3.9 MET-hours/week at baseline, 18-, 36- and 60-month follow-up. The average level of physical activity at post-diagnosis was also significantly higher than that at pre-diagnosis ( P < 0.001), with median value of 5.8 MET-hours/week. However, there was no significant difference between any two follow-ups at post-diagnosis. The proportions of participant who met the exercise recommendation (according to WCRF/AICR, 10 MET-hours/week) were low at pre- and post-diagnosis, being 20.7% and 35.1%, respectively. Compared to pre-diagnosis, most of the patients improved or had no change on level of recreational physical activity at post-diagnosis, with the respective proportion being 48.2% and 43.8%. Multivariate analysis showed that higher increase in physical activity after cancer diagnosis was observed among breast cancer patients who were married or cohabitation, unemployed (compared to full time) and had no comorbidity (compared to patients who had one comorbidity). Conclusions: Chinese breast cancer patients reported significant and long-term changes in physical activity after cancer diagnosis, which was in line with current recommendation. However, the proportion of patients who met the exercise recommendation for cancer survivors was still low. Empowering patients on the importance of durable high level of physical activity in breast cancer survivorship is warranted.

TPS474 Background: Gastric carcinoma (GC) is the third-leading cause of death in China (291,000 deaths in 2015). Current therapies include surgery, chemotherapy, radiotherapy and targeted therapy, which prolong PFS and OS to 6 months and 8-14 months, respectively. Tumor Treating Fields (TTFields) are a non-invasive, regional antimitotic treatment modality approved by the FDA for glioblastoma and malignant pleural mesothelioma. TTFields at specific frequency (100- 500 kHz) are delivered via transducer arrays placed on the skin of the upper abdomen, back, right and left hypochondriac regions where the primary tumor lesion is located. TTFields were effective in preclinical models of gastric cancer and there are several ongoing Phase 3 trials of TTFields in multiple solid tumors. In this phase 2, single arm, open-label, multi-center study, we will investigate for the first time the efficacy and safety of TTFields concomitant with XELOX (oxaliplatin/capecitabine) as the first-line treatment of GC. Methods: Patients (N = 50) with histologically confirmed unresectable, locally advanced or metastatic Gastroesophageal Junction (GEJ) or Gastric Adenocarcinoma (GC), aged ≥ 18 years, ECOG PS 0-1, who had no previous systemic treatment for the recurrent or metastatic disease will be enrolled. Patients will receive TTFields (150 kHz via the NovoTTF-100L (P) medical device for average monthly use of 18 hrs/day) plus XELOX chemotherapy (Oxaliplatin: 130 mg/m2 on day 1 every 3 weeks; Capecitabine: 1000 mg/m2, PO, BID on day 1-14 every 3 weeks). For HER-2 positive patients, trastuzumab is allowed. The primary endpoint is investigator-assessed Objective Response Rate (ORR) per RECIST 1.1. Secondary endpoints are time to tumor progression (TTP), progression-free survival (PFS), overall survival (OS), and 12-month OS rate. Adverse events (AEs) will be graded for severity according to CTCAE 5.0. Based on the historical ORR data in first-line chemotherapy in GC, we assumed that ORR will be higher than 45% with TTFields concomitant with chemotherapy. At least 45 patients need to be enrolled to ensure the lower boundary is 30% of the 95% CI. Estimating a patient drop-out rate of 10%, 50 patients will be actually enrolled.

Purpose Hepatitis B virus (HBV) infection is an important etiology for hepatocellular carcinoma (HCC). We aim to develop a simple clinical score in predicting the risk of HCC among HBV carriers. Patients and Methods We first evaluated 1,005 patients and found that the following five factors independently predicted HCC development: age, albumin, bilirubin, HBV DNA, and cirrhosis. These variables were used to construct a prediction score ranging from 0 to 44.5. The score was validated in another prospective cohort of 424 patients. Results During a median follow-up of 10 years, 105 patients (10.%) in the training cohort and 45 patients (10.6%) in the validation cohort developed HCC. Cutoff values of 5 and 20 best discriminated HCC risk. By applying the cutoff value of 5, the score excluded future HCC development with high accuracy (negative predictive value = 97.8% and 97.3% in the training and validation cohorts, respectively). In the validation cohort, the 5-year HCC-free survival rates were 98.3%, 90.5%, and 78.9% in the low-, medium-, and high-risk groups, respectively. The hazard ratios for HCC in the medium- and high-risk groups were 12.8 and 14.6, respectively. Conclusion A simple prediction score constructed from routine clinical and laboratory parameters is accurate in predicting HCC development in HBV carriers. Future prospective validation is warranted.

Purpose Reactivation of hepatitis B virus (HBV) infection is a well-recognized complication in cancer patients with chronic HBV (hepatitis B surface antigen [HBsAg] positive) undergoing cytotoxic chemotherapy. In patients who have resolved HBV (HBsAg negative and antibody to hepatitis B core antigen [anti-HBc] ± antibody to hepatitis B surface antigen [anti-HBs] positive), such incidence has been much less common until recent use of rituximab. In this study on HBsAg-negative/anti-HBc–positive lymphoma patients, the objectives were to determine the HBV reactivation rate in patients treated with rituximab-containing chemotherapy and to compare it with the rate in patients treated without rituximab. Patients and Methods Between January 2003 and December 2006, all patients diagnosed with CD20+ diffuse large B-cell lymphoma (DLBCL) had HBsAg determined before anticancer therapy. They were treated with either cyclophosphamide, doxorubicin, vincristine, and prednisone (CHOP) alone or rituximab plus CHOP (R-CHOP). HBsAg-negative patients had anti-HBc determined; serum was stored for anti-HBs and HBV DNA. All patients were observed for HBV reactivation, which was defined as detectable HBV DNA with ALT elevation during and for 6 months after anticancer therapy. Results Among 104 CD20+ DLBCL patients, 80 were HBsAg negative. Of the latter, 46 patients (44.2%) were HBsAg negative/anti-HBc positive; 25 of these patients were treated with CHOP, and none had HBV reactivation. In contrast, among the 21 patients treated with R-CHOP, five developed HBV reactivation, including one patient who died of hepatic failure (P = .0148). Exploratory analysis identified male sex, absence of anti-HBs, and use of rituximab to be predictive of HBV reactivation. Conclusion Among HBsAg-negative/anti-HBc–positive DLBCL patients treated with R-CHOP, 25% developed HBV reactivation. Close monitoring until at least 6 months after anticancer therapy is required, with an alternative approach of prophylactic antiviral therapy to prevent this potentially fatal condition.

Purpose To study the incidence and treatment outcomes of hepatocellular carcinoma (HCC) detected in an intensive surveillance program (ISP) of hepatitis B virus (HBV) carriers. Patients and Methods We screened 1,018 HBV carriers by serum alpha-fetoprotein (AFP) measurement and abdominal ultrasonography (AUS). Patients with an abnormal AFP level or AUS result were enrolled in an ISP that included Lipiodol computed tomography followed by AFP measurement/AUS every 3 months for 2 years and then every 6 months thereafter. The rest were on routine surveillance for 2 years. Results A total of 9,849 serum AFP measurements and 3,053 AUSs were performed. After a median follow-up of 4.12 years, we diagnosed 24 HCCs among 78 patients with abnormal screening test results at enrollment (group A); 23 HCCs among 93 patients with only abnormal surveillance test results during follow-up (group B); and nine HCCs among 847 patients with 2 years of normal surveillance (group C). Annual incidence of HCC in the ISP was 760.2 (95% CI, 538.4 to 1,073.7) per 100,000. Mean tumor sizes were 3.02, 2.91, and 4.82 cm in groups A, B, and C, respectively (P = .01). Tumor resection rate of the ISP was 36.2%, although another 29.8% of the patients were eligible for locoregional ablative therapy. Conclusion This study illustrated that a high incidence of relatively small HCCs may be detected by using intensive surveillance of high-risk HBV carriers. However, the surgical resection rate was low, and we were not able to demonstrate clinical benefit with the early detection. Future surveillance studies should consider incorporation of therapy aimed at long-term control of small-sized tumors.

e15679 Background: Various serum inflammatory factors are implicated in pathogenesis and prognostication of HCC but none of them is incorporated into a HCC prognostic model. We aim to build an inflammation-integrated prognostic model without any radiological feature. Methods: A prospective-retrospective cohort of patients with HCC was accrued from 2001 to 2013, and was randomly spitted into training and validation cohorts. Cox proportional hazards model was used for univariable and multivariable survival analyses. The risk score was constructed by factors and their coefficients identified by multivariable Cox model. Results: The entire cohort was composed of 1315 patients (BCLC 0/A/B/C/D 73/533/237/425/47; Surgery/Locoablation/TACE/Systemic agent/BSC 538/128/318/122/209). In the training cohort (n=657), a novel model, PANALA score, was generated by using6 independent prognostic factors: 1.488*(Performance status >0) – 0.057*Albumin + 0.098*Neutrophil – 50.4*Alkaline phosphatase-1 – 0.269*Lymphocyte + 0.11*ln(Alpha fetoprotein). PANALA outperformed other staging systems (Table) and allowed the prediction of survival of an individual patient. In the validation cohort (n=658), the prognostic performance of PANALA was still the most superior one. The goodness of fit of PANALA score in prediction of survival was evaluated by dividing patients into 4 equal quartiles by PANALA. In the quartile 1, overall survival between actual observation and prediction was highly comparable (6m survival: 100% vs. 95%, 3y survival: 80% vs. 79%, 5y survival: 72% vs. 74%; P=0.71). Discrepancies between observed and predicted survival were also insignificant in other quartiles. Moreover, PANALA was able to stratify each BCLC stage into 3 prognostically different subgroups. Conclusions: PANALA is a novel and validated model with high prognostic performance. [Table: see text]

4013 Background: Many pts with advanced HCC progress on SOC therapy. Nivo is a fully human anti–PD-1 IgG4 mAb that demonstrated durable responses (20% ORR with a median DOR of 9.9 mo; 9-mo OS rate was 74%) in pts with advanced HCC in the dose-expansion (EXP) phase of the CheckMate 040 study (NCT01658878; Melero et al. 2017). Here we present survival and durability of response data in both sor-naive and -experienced pts with advanced HCC in CheckMate 040. Methods: Pts naive to or previously treated with sor received nivo in phase 1/2 dose-escalation (ESC; 0.1–10 mg/kg) and -EXP (3 mg/kg) cohorts Q2W regardless of PD-L1 status. Primary endpoints were safety/tolerability (ESC) and ORR (EXP; ORR by investigator [INV] and blinded independent central review [BICR]) using RECIST v1.1. Secondary endpoints included DOR, DCR, and OS. Biomarkers were assessed using pre-treatment tumor samples. Results: Overall, pts (N=262) had a median follow-up of 12.9 mo, and 98% had Child-Pugh scores 5–6. In sor-naive pts (n=80), the ORR (INV) was 23%, with 44% of responses (8/18) ongoing (Table). The DCR was 63%; 40% of pts had stable disease ≥6 mo. In sor-experienced pts (n=182; 91% progressed on sor), the ORRs (INV) were 16%–19%. Overall, responses occurred regardless of etiology or tumor cell PD-L1 expression. Nivo had a manageable safety profile consistent with that reported in other tumor types. Updated data with additional 4 mo of follow-up will be presented. Conclusions: Nivo demonstrated durable responses with long-term survival and favorable safety in both sor-naive and -experienced pts with advanced HCC. Clinical trial information: NCT01658878. [Table: see text]

226 Background: HCC diagnosed at advanced stages has a poor prognosis. Patients who progress on sorafenib have few options. Nivolumab, a fully human IgG4 monoclonal antibody inhibitor of programmed death-1 (PD-1), has demonstrated clinical and survival benefit in a number of tumor types. Here we report updated interim analyses of safety, efficacy, and exploratory biomarkers in patients with advanced HCC treated with nivolumab in the CheckMate 040 study (NCT01658878). Methods: Patients enrolled had advanced HCC with or without hepatitis C or B virus (HCV or HBV) infection. Prior sorafenib was allowed. Phase 1 dose-escalation evaluated nivolumab (0.1–10 mg/kg) Q2W. Phase 2 dose-expansion was initiated in 4 cohorts: sorafenib naïve/intolerant, sorafenib progressors, HCV infected, and HBV infected. All cohorts received nivolumab 3 mg/kg Q2W. The primary endpoint in the dose-escalation phase was safety/tolerability, and the primary endpoint in the dose-expansion phase was objective response rate (ORR) by RECIST v1.1 (central review). Secondary endpoints included duration of response (DOR), disease control rate (DCR), and overall survival (OS). Biomarkers within pre-treatment tumors were assessed. Results: Across dose escalation and expansion phases, 262 patients have been treated. Grade 3/4 treatment-related adverse events occurred in 20%. No maximum tolerated dose was reached during dose escalation (n = 48). The ORR (investigator-assessed) was 20% (95% CI 15–26) in 214 patients treated in the dose expansion phase with a median DoR of 9.9 months; DCR was 64% (95% CI 58–71). Responses were observed across etiologies and regardless of tumor PD-L1 expression. ORRs of 23% (95% CI 13–36) and 21% (95% CI 11–34) were observed in the uninfected sorafenib-naive and -treated patients, respectively. The 9-month overall survival rate in the expansion phase was 74% (95% CI 67–79). Association between immune-cell biomarkers and clinical outcomes will be presented. Conclusions: In this heavily pretreated population, responses to nivolumab were durable with encouraging overall survival. Safety was manageable and consistent with that observed in other solid tumors with no new safety signals. Clinical trial information: NCT01658878.

4012Background: For pts with advanced HCC on sorafenib (sor), overall survival (OS) is 11 mo; median OS with best supportive care (BSC) post-sor failure is 7–8 mo.Safety and preliminary antitumor e...

4094 Background: Survival in HCC depends on tumour-related features and liver dysfunction, the latter currently characterised by Child-Pugh score (C-P.S). Since some constituents of C-P.S are inter...

181 Background: There have been few international studies that systematically evaluate the impact of geographical region on the prognosis of HCC. Methods: Prospective cohorts of patients (pt) were accrued in UK (n=567; 2006-2011) and HK (n=517; 2007-2012). Clinical data were documented at baseline; treatments were decided by multidisciplinary teams at both centers. Results: The median follow-up time of the UK and HK cohort was 27.9 and 29.8 months, respectively. The median overall survival (OS) of the whole UK and HK cohort was 22.9 months and 8.6 months. In the HK cohort, 425 underwent palliative treatment (TACE, systemic agents or supportive care) and 92 had curative treatment (surgery and locoablation); in the UK cohort, 338 had palliative treatment and 228 underwent curative treatment. In the curative group, the median OS was 54.8 months in the UK cohort but has not yet been achieved in the HK cohort. For the palliative group, the median OS was 12.7 and 5.5 months in the UK and HK cohort respectively. In the palliative group, geographical difference was an independent prognostic factor (HK vs. UK, HR=2.0; p

e14528 Background: Transarterial injection of LEM produces long-lasting embolization of both arteriolar feeders and peribiliary plexus of HCC leading to infarction of tumor. Our group previously reported the efficacy and safety of LEM for HCC in a phase II clinical trial (Yu et al. J Vasc Interv Radiol. 2008). We initiated a randomized study to compare LEM vs. TACE in HCC. (NCT00467974). Methods: Major eligibility criteria: histologically confirmed HCC; unresectable disease without distant metastases; size ≤15cm; absence of portal vein thrombosis. Consented patients (Pts) were randomized into LEM vs. TACE stratified by tumor size and number. Pts were prospectively followed up and assessed with computed tomography (CT) every 8 weeks. The primary endpoint is overall survival (OS) and secondary endpoints include progression free survival (PFS), RECIST response rate (RR) and toxicity. An interim analysis is planned after recruitment of 70 pts. Results: From 07 to 10, total 70 pts were enrolled. 2 pts withdrew consent and were excluded from the analysis. There were no differences in baseline characteristics between LEM and TACE. Male:Female=53:15; Mean Age=64; Tumor size≤7: >7cm=45:23; Child A:B=55:13. The median follow-up was 2.3 years. The OS of LEM was 25.32 months(m) vs. TACE 22.56m (p=0.38). PFS of LEM was 9.00m vs. TACE 7.8m (p=0.50). Partial response of LEM vs. TACE was 30% vs. 18% (p=0.52); disease control rate of LEM vs. TACE was 80% vs. 70% (p=0.39). Commonest G3 or above toxicities include ALT elevation (LEM 22.2% vs. TACE 9.4%); hyperbilirubinemia (LEM 19.4% vs. 9.4%); pain (LEM 0% vs. TACE 4%). Except pain, there were no differences of toxicities between LEM and TACE after procedure (p=0.04). Conclusions: At interim analysis, LEM showed comparable efficacy and toxicity profile to TACE. A trend of better OS, PFS and RR was observed in LEM arm but statistically insignificant. Further accrual and follow-up is ongoing.

4103 Background: Sorafenib (S) is the only approved systemic treatment for unresectable HCC. Nevertheless, there remains an unmet medical need for more effective treatment options for this disease. BAY 86-9766 (B) is an oral, allosteric inhibitor of MEK, a key component of the MAP kinase pathway. This study evaluated the efficacy and safety of a combination therapy with B plus S in patients (pts) with HCC. Methods: This is a single arm, open-label, phase 2 study. Eligible were pts with unresectable HCC, Child-Pugh Class A, performance status (PS) 0-1, and no prior systemic anticancer therapy for HCC. Pts started Cycle 1 (21 days) with B 50 mg bid orally plus S 600 mg daily (200 mg AM, 400 mg PM) orally. If there was no hand-foot skin reaction, fatigue, or gastrointestinal toxicity ≥ grade 2, S was escalated to 400 mg bid from Cycle 2 on. Treatment continued until progression or withdrawal criteria were met. Tumor assessment was performed every 6 weeks during treatment. Safety was evaluated every week for the first 6 weeks and every 3 weeks thereafter. Results: Seventy pts from Asia started study treatment. Pts were predominantly male (86%); median age was 56 years; 54% had PS of 0 and 46% PS of 1. The vast majority had liver cirrhosis (83%) and infection with HBV (76%) or HCV (17%). Sixty-five were evaluable for efficacy per protocol. Three pts (5%) had confirmed partial response and 25 pts (38%) had prolonged stable disease (≥10 weeks), with a disease control rate of 43%. Median time-to-progression was 4.1 months. Survival data are not mature, yet. The most frequent drug-related adverse events (AEs) were rash (60%), diarrhea (59%), AST elevation (43%), vomiting (30%), nausea (29%), ALT elevation (26%), and anorexia (26%). There were 4 Grade 5 related AEs (hepatic failure, sepsis/hepatic encephalopathy, tumor lysis syndrome, and unknown cause, respectively). Dose modifications due to AEs were necessary in almost all pts. The median daily dose was 64.2 mg for B and 443.3 mg for S, respectively. Conclusions: B in combination with S showed antitumor activity in pts with HCC. However, frequent dose modifications due to AEs might have limited the treatment effect of this combination.

4058 Background: Methylation of RASSF1A is a frequent molecular event along hepatocarcinogenesis. Our group previously developed a sensitive and specific assay in the quantitation of circulating me...

205 Background: Serum interleukin-10 (IL-10) is associated with active hepatitis in patients (pts) with hepatitis B viral (HBV) infection. In HBV-related HCC, elevation of serum IL-10 is frequently observed but its significance on the outcome of HCC is unclear. Methods: A prospective cohort of newly diagnosed and inoperable HCC was recruited from a multidisciplinary clinic from Prince of Wales Hospital from 2006 to 2008. The baseline demographics, tumor characteristics/stage, laboratory parameters, virologic factors (HBV DNA, antiviral therapy) and first-line treatment modality were documented at the time of diagnosis. Serum IL-10 was measured by enzyme-linked immunosorbent essay. Univariate and multivariate analyses were conducted. Overall survival (OS) was the primary endpoint. Results: Total 180 new cases of inoperable HCC were evaluated. The median follow-up time was 15.5 months. Median age was 60.5 years. Most (159 pts; 88.3%) were males. 81.1% of them were positive for HBsAg. Total 120 (66.7%) had radiologic evidence of cirrhosis. 20 (11.1%), 85 (47.2%) and 75 (41.7%) received locoablative, trans- arterial/systemic therapy and supportive care respectively. The mean level of serum IL-10 was 18.1pg/ml (range: 2.8-11.7). 114 (63.3%) had log IL-10 higher than 1.0 pg/ml. Pts with log IL-10 >1.0 pg/ml had significantly worse OS than those with log IL-10≤ 1.0 pg/ml (14.8 vs. 4.5 months; HR 2.39; p1.0pg/ml had higher ALT and HBV DNA, lower albumin, higher chance of ascites, worse Child-Pugh stage and worse tumor stage (Table). Conclusions: Serum IL-10 is an independent prognostic factor for inoperable HCC. Pts with high IL-10 level have poorer liver reserves and worse tumor staging. [Table: see text] No significant financial relationships to disclose.

4028 Background: The role of antiviral therapy (AT) on the outcome of established HBV-related HCC is unclear. Based on two independent prospective cohorts on surveillance of HCC (Mok et al. J Clin ...

4584 Background: Surveillance for HCC in hepatitis B virus (HBV) carriers aims to improve survival by detection of early resectable tumor. In 2005, our group reported low surgical resection rates despite detection of small sized tumors in a prospective surveillance cohort of 1018 HBV carriers (Mok TS et al. J Clin Oncol 05). In current study, we aim to identify predictive factors for treatment outcomes in HCC detected from our surveillance program. Methods: The prospective cohort was recruited at the Prince of Wales Hospital between Oct 1997 and Nov 2000. We updated the database in Dec 2008 and performed univariate and multivariate analysis on clinical (age, sex, cirrhosis, ascites, anti-viral therapy, bilirubin, ALT, albumin, INR), tumorous (size, number, resection) and virologic factors (HBV DNA, Genotype) for prediction of outcome in HCC patients (pts). Results: In the prospective cohort, total 923 HBV carriers were updated (95 lost to follow-up). After median follow-up of 9.95 years, we confirmed diagnosis of 105 HCC. Median age = 51 (range: 40–69); M:F = 82:23; Child's A and B cirrhosis = 38:67. Fifty seven pts (54.3%) had solitary tumor but only 34 (32.4%) are amenable to resection. Absence of cirrhosis (p=0.0072) and normal albumin level (p=0.0379) are predictors of surgical resection while tumor size and number are not. The median survival of all HCC pts was 2.26 years. Anti-viral therapy during the surveillance period is a strong predictor of survival (3.74 vs. 1.63 years; p=0.0115). In multivariate analysis, both anti- viral therapy (HR=0.35; 95%CI: 0.19–0.73; p=0.0041) and normal bilirubin level (HR=0.30; 95% CI: 0.25–0.80; p=0.0069) were predictive of improved survival in pts with HCC. The benefits of anti-viral therapy applied to both surgical and non-surgical candidates (p=0.17). Conclusions: Anti-viral therapy is a potent predictor of survival of HBV related HCC. Liver function is more important than tumor characteristics in determining the outcome of HCC detected in the surveillance program. No significant financial relationships to disclose.

4602 Background: Although alpha-feto protein (AFP) is frequently measured during treatment of hepatocellular carcinoma (HCC), the trend of AFP in predicting response and survival of HCC patients is...

4591 Background: Prognosis of patients with HBV-related chronic liver disease and hepatocellular carcinoma (HCC) depends on both extent of tumor involvement and functional hepatic reserve. The Chin...

20549 Background: A Phase III randomized, double-blind placebo controlled trial was done to determine if there is a favourable QOL outcome for patients receiving aprepitant for the prevention of chemotherapy-induced nausea and vomiting (adriamycin 60 mg/m2 + cyclophosphamide 600 mg/m2). Methods: 80 breast cancer patients were entered with 40 in the study treatment arm (aprepitant 125mg; ondansetron 8mg x 2 doses; dexamethasone 20mg on day 1 and aprepitant 80mg QD day 2–3) and 40 patients in the control arm (ondansetron 8mg x 2 doses; dexamethasone 12mg on day 1, and ondansetron 8mg every 12 hours on days 2–3). The QLQ-C30 and FLIE were completed prior to chemotherapy and at the end of day 5. A change score from baseline to day 5 was used in the analysis. Results: The average nausea scores captured by 100 mm visual analog scale between placebo and aprepitant arms were not significant (16.5 vs. 12.8, p=0.554). However, in cycle 1 the QLQ-C30 NV subscale (40.3 vs 20.4, p=0.0066) and appetite loss (57.4 vs 33...

9048 Background: Lymphedema is a relatively common occurrence after treatment for breast cancer. Some of the causes are well known such as axillary surgery and axillary radiation therapy. However, many of the potential modifiable factors such as weight-bearing exercises and airline travel have been inadequately studied. Methods: A matched case-control study was designed to evaluate potential factors associated with lymphedema for breast cancer patients. Subjects were 202 women undergoing a unilateral axillary dissection for breast cancer, consisting of 101 cases with lymphedema and 101 controls without lymphedema who were matched cases in terms of surgery date, having received axillary radiotherapy or not and stage of cancer. Potential risk factors were collected by using clinical data and questionnaire. Two-arm circumferences were measured to determine presence and severity of lymphedema. A multiple logistic regression was used to obtain the adjust odds ratios for potential risk factors. Results: Adjusted odds ratio (OR) for lymphedema were separately 3.80 (95% CI 1.84- 7.87) for previous inflammation-infection, and 1.06 (95% CI 1.02–1.10) for an increase in one year of age at axillary dissection. Adjusted ORs for moderate to severe degree of lymphedema were separately 4.49 (95% CI 2.16–9.30) for previous inflammation-infection, 2.97 (95% CI 1.46- 6.03) for operation on dominant arm, 1.11 (95% CI 1.01–1.21) for an increase in 1kg/m2 body mass index at recruitment, and 1.05 (95% CI 1.01–1.10) for an increase in one year of age at recruitment time. Leisure activities requiring walking once a week to twice a week was a protective factor with adjusted ORs of 0.23 (95% CI 0.08–0.66) for developing lymphedema and 0.19 (95% CI 0.05–0.69) for developing moderate to severe lymphedema. Hypertension and tumor location at upper-outer or outer quadrant in breast were not important factors. Conclusions: Previous inflammation-infection and older age at axillary dissection are risk factors for initiating lymphedema. Previous inflammation-infection, operation on the side of dominant hand, obesity and aging are risk factors for aggravating lymphedema. No significant financial relationships to disclose.

15081 Background: Hepatocellular carcinoma (HCC) is a common cause of cancer morbidity and mortality. It is a highly aggressive tumor with 90% presenting with unresectable disease, resulting in a median survival of 3–6 months. Inhibitors of histone deacetylase (HDAC) have been demonstrated in HCC cell lines and xenografts to induce apoptosis and tumor regression, and have anti-proliferative, anti-metastatic and anti-invasive effects. Belinostat (N-hydroxy-3-[phenylsulphamoylphenyl] acryl amide) is a novel, low molecular weight, HDAC inhibitor. The Cancer Therapeutic Research Group (CTRG) is conducting a phase I/II study of belinostat as the first line therapy for pts with unresectable HCC. The phase I study aims to determine dose limiting toxicity (DLT) and maximum tolerated dose (MTD). Methods: Patient eligibilities include histologically or cytologically confirmed unresectable HCC, ECOG = 2, adequate hematologic, renal and hepatic functions. The dose of belinostat was started at 600 mg/m2/day i.v. on day 1–5 every 3 weeks; with increment of 300 mg/m2/day up to 1200 mg/m2/day. Dose limiting toxicities (DLT) are defined as grade 4 hematological toxicity or grade 3 or 4 nonhematological toxicity during cycle 1 (according to NCI CTC v3), or treatment delay >2 weeks. The MTD is defined as the dose below which ≥ 2 of 3 or ≥ 2 of 6 patients experiencing DLT. After determination of the MTD, 3 additional patients will be treated at this dose to further define toxicity. Results: From June-December 2006, 12 pts were entered; 3 were treated at level I (600 mg/m2/day), 3 at level 2 (900 mg/m2/day), and 6 at level 3 (1200 mg/m2/day). There were no DLTs on level 1, 2 or 3. Therefore, the MTD of belinostat was not reached. Overall toxicities were WBC gr. 0/1/2/3/4: 12/0/0/0/0, ANC gr. 0/1/2/3/4: 12/0/0/0/0; Hb gr. 0/1/2/3/4: 4/7/0/1/0; platelet gr. 0/1/2/3/4: 8/4/0/0/0; lethargy gr. 0/1/2/3: 12/0/0/0/0, phlebitis gr. 0/1/2/3: 10/0/2/0; Cough gr. 0/1/2/3: 9/0/3/0; Nausea gr. 0/1/2/3: 7/3/2/0 and vomiting gr. 0/1/2/3: 9/3/0/0. Conclusions: At the maximum dose of 1200 mg/m2/day (level 3), MTD has not been reached. Belinostat is very well tolerated. Phase II study will be started at dose level 1200 mg/m2/day. Sponsor: The Division of Cancer Treatment and Diagnosis, National Cancer Institute, USA. No significant financial relationships to disclose.

14105 Background: Cytotoxic chemotherapy has been widely used in patients (pts) with advanced or metastatic gastric cancer, but the reported high response rates in phase II settings have not been confirmed by randomized trials. Moreover, the therapeutic impact on survival has been modest. The purpose of this study was to evaluate Irofulven, a DNA interacting acylfulvene analog, as first line therapy for pts with recurrent or metastatic gastric cancer. Methods: The main eligibility criteria were: Histologically or cytologically confirmed disease; age > 18 yrs; ECOG ≤ 2; measurable disease; no prior chemotherapy for recurrent or metastatic disease; life expectancy > 3 months. Irofulven was given at 0.45 mg/kg i.v. over 30-min infusion (max. 50 mg), on Day 1 and 8, every 3 weeks. The primary endpoint was treatment response and toxicity; the secondary end-point was survival. A 2-stage phase II design was used. In the first stage, the target enrollment was 20, if ≤ 2 responses were observed, the study would be stopped and the treatment concluded to be ineffective; if ≥ 6 responses were observed, the study drug would be concluded to be active. Otherwise, another 15 pts would be entered into the second stage. Survival was constructed using the Kaplan-Meier method. All patients entering the trial were included in the survival analysis. Results: 23 pts were entered into the first stage. The median no. of cycles per pt was 2 (range: 1–6). 2 pts (9%) had ≥ 1 week delay in administration of subsequent cycle of chemotherapy. For the day 8 chemotherapy, 7 pts (30%) required dose reductions; 5 (22%) had dose omitted. 22% and 17% pts developed grade 3/4 anemia and neutropenia respectively. There was no grade 4 thrombocytopenia or neutropenic fever. Of the 20 evaluable pts, no responses was observed, 3 pts had stable disease after 2 cycles of treatment which was not confirmed by a further assessment. Median overall survival was 6.05 months (95% CI 4.55–9.39). Conclusions: Irofulven was tolerated at the described dose but showed no evidence of antitumor activity in patients with advanced gastric cancer. Acknowledgement: The study was sponsored by the Division of Cancer Treatment and Diagnosis, National Cancer Institute, U.S.A, and its collaborator MGI Pharma, Inc. No significant financial relationships to disclose.

14054 Background: Transarterial ablation with lipiodol-ethanol mixture (LEM) induced potent dual embolization of arterioles and portal venules in preclinical studies, and extensive tumor necrosis in clinical studies. It is potentially an effective treatment and a better alternative to chemoembolization for unresectable hepatocellular carcinoma (HCC). The authors aim to study the safety and efficacy of this treatment. Methods: Unresectable HCC diagnosed by histology or viral serology + AFP > 500 ug/l + typical imaging; bilirubin ≤ 30 umol/l; hypervascular tumor on CT / angiogram, no vascular invasion or extrahepatic spread. LEM with Lipiodol and ethanol mixed in ratio of 2:1 by volume was infused transarterially into tumor vasculature by superselective catheterization of tumor feeders at a subsegmental level under fluoroscopy until flow stagnation. Adverse events (AE) graded by NCI-CTC V2.0. Tumor response monitored by CT scan q3m. LEM repeated for residual enhancing lesion or intrahepatic progression on CT scan. Primary endpoint was response rate. Secondary endpoints: toxicity and survival. Results: From 7/2001 to 4/2005, 77 pts enrolled. Age 28–85 (median 64). M:F=60:17. ECOG 0=68; 1=9. Cirrhosis 92%. HBV 77%, HCV 14%, alcohol 4%. Child A 88%, B 12%. Okuda A 73%, B 27%. Histology confirmed in 78%. 39% had previous treatment. Median tumor size 4.5cm (range 1.5–15cm); Number of lesions: One (58), Two (14), Three (5). Best tumor response (WHO criteria): CR 0%, PR 43%, MR 36%, SD 17%, PD 4%. Patients received 1 to 8 LEM treatments (mean 2.2). Mean hospital stay 2.7 days (range 2 -15). Significant AEs are shown in the table. Fever in 15% for mean duration of 3.1 days. After a median follow-up of 1.65 yr, 23 pts died. One year survival 77%. 3 pts had tumor resection after downstage by LEM. Conclusions: Transarterial ablation with LEM is a well tolerated and effective local regional treatment of unresectable HCC. The high tumor response and 1-year survival rate warrants further investigation in comparative study. [Table: see text] No significant financial relationships to disclose.

20026 Background: In order to obtain a good prognostic model for HCC, clinical data alone may not be adequate. DNA microarray technology has enabled quantification of thousands of genes in a single assay but it has the inborn drawback of high background noise. In order to deal with the problem of information overflow, we developed multivariate dependencies approach using CGH to guide the initial modeling process. This approach takes into account the multivariate nature and potential interaction among the genes during the prognostic modeling process. Methods: The study includes 165 patients with CGH data containing 858 regions/bands. The clinical outcome is survival at 1.72 years. Potential prognostic factors include CGH regions, albumin, ALT, bilirubin, AFP, ascites, alkaline phosphatase (ALP), tumor size, and encephalopathy. We used a blocked wavelet-shrinkage principal component analysis (BWSPCA) to reduce the dimension of CGH data with respect to clinical outcome and followed by logistic regression. We compared the BWSPCA with PCA alone, supervised PCA (Bair et al. 2004), and supervised BWSPCA. Results: Among the 165 patients, 133 (80%) were male, average age of 54.8 years, 78 (47%) stage I-II and 75 (45%) ECOG 0–2. PCA alone and supervised PCA models failed to identify significant CGH regions. BWSPCA model includes tumor size (p = 0.005), albumin (p = 0.047), ALP (p = 0.025), Chr.6q25.1 (p = 0.031), Chr.12q24.32 (p = 0.012), Chr.Xq28 (p = 0.035). Supervised BWSPCA approach includes tumor size (p = 0.016), ALP (p = 0.002), Chr.12q24.11 (p = 0.047). The area under the receiver operating curve (ROC) for the BWSPCA model was 0.78 with a sensitivity and specificity about 0.8 and 0.6 respectively. Conclusions: PCA alone is not effective in identifying CGH regions as prognostic factors. Supervised learning approach did not improve the results. The BWSPCA method identified a number of significant CGH regions associated with survival outcome for HCC patients. These results would be verified in future study. This method will be extended and applied to develop prognostic models using both CGH and DNA microarray data. Acknowledgement: This study is funded by the Research Grant Council of Hong Kong #CUHK4469/03M. No significant financial relationships to disclose.

4205 Background: Phase II clinical trials aim to assess anti-tumor activity of investigational therapies in order to detect promising agents for future phase III trials. A review of previous phase II studies on single agent doxorubincin (Dox) for HCC shows that response rates (RR) vary widely from 0% to 44%. A randomized phase II design is preferred but the decision theoretical approach by Simon picking the winner may not be appropriate in this circumstance. We proposed a new randomized phase II design using control as an adjustment for the efficacy of study treatment. Method: A hypothesis similar to a single-arm phase II design will be stated, e.g. H0: p = p0 vs H1: p = p1. Patients (pts) randomized to control arm will receive standard treatment. The RR of the control arm (p0) with s0 responders from n0 pts will be used in a conditional model to adjust for the RR in the study arm (p1). The probability of the study arm with s1 responders from n1 pts given the null hypothesis, i.e. P(s1| n1, p0) has binomi...

5523 Background/Objective: NPC is highly radiosensitive and chemosensitive. 3 cycles of 3-weekly chemotherapy concurrent with RT followed by 3 cycles of adjuvant chemotherapy has improved survival over RT alone in a previous randomized study. The regimen was associated with significant toxicities. This randomized phase III study compared concurrent weekly cisplatin-RT with RT alone in patients with locoregionally advanced NPC. Methods: Patients with Ho's N2 or N3 stage or N1 stage with nodal size ≥ 4 cm were randomized to receive cisplatin 40 mg/m2 weekly up to 8 weeks concurrently with RT (CRT) or RT alone. The primary endpoints are progression-free survival (PFS) and overall survival (OS). Results: 350 eligible patients were randomized. Baseline patient characteristics, treatment toxicities and PFS analysis at median follow up of 2.71 years have been published previously. At a median follow up of 5.5 years where 58 and 73 deaths were found out of 174 CRT and 176 RT patients respectively (95th and 5th pe...