Clinical outcomes of abiraterone acetate (AA) or enzalutamide (E) as first-line therapy (Rx) for men aged ≥75 with metastatic castration-resistant prostate cancer (mCRPC) according to previous use of docetaxel (D) for metastatic castration-sensitive prostate cancer (mCSPC) in a multicenter international registry: A SPARTACUSS – Meet-URO 26 study

107

Background: The optimal management of mCRPC in men aged ≥75 is challenging, and there is a paucity of clinical data in the literature. Although AA and E are commonly used as 1st line Rx for mCRPC, it is unclear whether use of upfront D for mCSPC may impact their clinical efficacy or safety in this elderly population. Methods: Patients aged ≥75 who started AA or E as 1st line Rx for mCRPC within January 2015 - April 2019 were identified from the IRB approved databases of 10 institutions in Europe, South and North America. Demographic and clinicopathological data were collected from available medical records, including Gleason, prior local therapy, newly diagnosed metastatic disease, disease volume, ECOG, PSA and sites of metastases. Patients were classified by use of upfront D for mCSPC. The primary endpoints were overall survival (OS) from AA/E onset and OS from ADT start and safety of AA/E. The endpoints distributions including median (95% CI) were estimated by Kaplan-Meier method. Results: Of the 337 patients selected, 24 (7.1%) received ADT+D and 313 (92.9%) ADT alone for mCSPC. Patients with ADT+D tended to be younger (78 vs 81, p=0.022) and, albeit not statistically significant, had higher rates of Gleason score >8 (81.0% vs 62.6%, p=0.10), newly diagnosed (83.3% vs 65.6%, p=0.08) and high volume disease (45.8% vs 34.6%, p=0.28), compared to those with ADT alone. Median follow-up was 18.8 months. No significant difference of OS from ADT start or from AA/E onset was observed between the 2 cohorts (see table). Despite OS from ADT start being longer in those having ADT+D, OS from AA/E start was approximately 2 years in both cohorts. Rates of adverse events (AEs) of any grade (58.3% vs 52.1%, p=0.67) and grade ≥3 (12.5% vs 15.7%, p=1.0) did not significantly differ between the 2 cohorts. Conclusions: While limited by small sample size for ADT+D and retrospective study design, patients aged ≥75 having AA/E as 1st line mCRPC Rx showed similar survival outcomes and tolerability regardless of previous use of D for mCSPC. [Table: see text]

Prognostic biomarkers and clinical outcomes in neuroendocrine prostate cancer (NEPC)

209

Background: NEPC includes both pure small cell carcinoma and mixed tumors with varying degrees of adenocarcinoma and neuroendocrine histology. It arises de novo or is treatment associated (TA) post androgen deprivation therapy. Clinical outcome data and prognostic biomarkers are limited and were thus explored. Methods: Patients with high grade prostate cancer and morphologic and/or immunohistochemical (IHC) NEPC features were included in this retrospective multicentre study. Clinical stage, Gleason score, and serum biomarkers were recorded. Kaplan-Meier method and log-rank test calculated and compared overall survival (OS) from time of NEPC diagnosis.Cox proportional hazards regression assessed prognostic impact of serum biomarkers at diagnosis and de novo vs TA status, adjusting for clinical stage and castration resistance. Results: 135 NEPC cases were identified. 124 (92%) were mixed tumors. 56 (41%) arose de novo. 79 (59%) were TA. 77% of those with a Gleason score (N=85/110) were grade group 5. Median PSA pre-NEPC biopsy was 11.6 ng/mL. At NEPC diagnosis, 19 (14%) had localized disease (median OS 123.0 mo); 33 (24%) non-metastatic castrate-sensitive disease (median OS 42.3 mo); 6 (4%) non-metastatic castrate-resistant disease (median OS 14.3 mo); 35 (26%) metastatic castrate-sensitive disease (median OS 17.6 mo); and 42 (31%) metastatic castrate-resistant disease (median OS 9.6 mo). Median OS for those with visceral metastases was 8.6 mo (95% CI 6.0 – 14.6), compared to patients with non-visceral metastases (11.1 mo; 95% CI 13.7 – 21.5) and no metastases (42.3 mo; 95% CI 47 – 89). Anemia (adjusted HR 1.66; 95% CI 1.05 - 2.16, p = 0.031) and NLR >3 (adjusted HR 1.51; 95% CI 1.01 - 2.52, p = 0.045) were associated with increased risk of death. De novo disease, elevated LDH, serum PSA, and Gleason score were not prognostic. Conclusions: This study identifies NEPC clinical outcomes by stage, with survival poorer than expected in pure prostate adenocarcinoma. Anemia and elevated NLR >3 are prognostic biomarkers that may help risk stratify and guide treatment intensification, including platinum-based chemotherapy. Further biomarker characterization of NEPC through IHC-staining pattern and genomic analysis is currently underway by this group.

Efficacy and safety of docetaxel (D) vs androgen-receptor signaling inhibitors (ARSi) as second-line therapy (Rx) after progression on alternative ARSi as first-line Rx for patients who are elderly (≥75 years old) with metastatic castration-resistant prostate cancer (mCRPC) in a multicenter international database: A SPARTACUSS–Meet-URO 26 study

166

Background: About 2/3 of all prostate cancer (PCa) deaths occur in patients aged ≥75, who are frequently diagnosed with advanced PCa. ARSi abiraterone acetate (AA) and enzalutamide (E) are the most common 1st line Rx for patients with mCRPC. Yet, the optimal treatment sequence for the elderly ≥75 after ARSi failure is still unclear. Methods: Using available medical records, patients aged ≥75 who started ARSi as 1st line Rx for mCRPC within January 2015 - April 2019 and, upon progression, 2nd line alternative ARSi or D were identified from the IRB approved hospital registries of 10 centers in Europe, North and South America. Patients were categorized by type of 2nd line Rx for mCRPC into cohorts AA/E and D. Primary endpoints were overall survival (OS) from 1st line AA/E start, OS and radiographic progression-free survival (rPFS) from 2nd line Rx start, and safety. The Kaplan Meier method was used to calculate endpoint distributions and medians (95% CI). Results: Of the 122 patients identified, 57 (46.7%) had AA/E and 65 (53.3%) D, as 2nd line Rx for mCRPC. Median follow-up was 26.3 months (95% CI, 23.1-27.9 months). Cohort AA/E tended to be older (81 vs 78 years; p=0.001) and with high-volume disease (45.5% vs 25.0%; p=0.022) compared to cohort D. No significant difference in OS from 1st line ARSi onset and OS or rPFS from 2nd line Rx start was found between the 2 cohorts. Cohort AA/E had longer rPFS than cohort D, albeit not significant (18.5 vs 12.0 months; p=0.13). Rates of adverse events (AEs) of any grade (42.1 vs 53.8; p=0.21) and AEs of grade ≥3 (19.3% vs 18.5%; p=1.0) did not show significant differences between the 2 cohorts. Conclusions: Within the limitations of small cohorts and retrospective design, treatment sequences with 2nd line AA/E or D after failure of 1st line alternative ARSi for mCRPC showed similar efficacy and safety in the elderly ≥75 years old.

Efficacy and Safety of First-line Systemic Therapy for Metastatic Renal Cell Carcinoma: A Systematic Review and Network Meta-analysis

CONTEXT

Considerable advances have been made in the first-line treatment of metastatic renal cell carcinoma (mRCC), with immunotherapy-based combinations including immunotherapy-tyrosine kinase inhibitors (IO-TKIs) and dual immunotherapy (IO-IO) favored. A lack of head-to-head clinical trials comparing these treatments means that there is uncertainty regarding their use in clinical practice.

OBJECTIVE

To compare and rank the efficacy and safety of first-line systemic treatments for mRCC with a focus on IO-based combinations.

EVIDENCE ACQUISITION

MEDLINE (Ovid), EMBASE, Cochrane Library, Web of Science, and abstracts of recent major scientific meetings were searched to identify the most up-to-date phase 3 randomized controlled trials (RCTs) of first-line IO-based combinations for mRCC up to June 2021. A systematic review and network meta-analysis were completed using the Bayesian framework. Primary endpoints included overall survival (OS) and progression-free survival (PFS). Secondary endpoints included the objective response rate (ORR), complete response (CR), grade 3-4 treatment-related adverse events (TRAEs), treatment-related drug discontinuation (TRDD), and health-related quality of life (HRQoL). The analysis was performed for the intention-to-treat (ITT) population as well as by clinical risk group.

EVIDENCE SYNTHESIS

A total of six phase 3 RCTs were included involving a total of 5121 patients. Nivolumab plus cabozantinib (NIVO-CABO) had the highest likelihood of an OS benefit in the ITT population (surface under the cumulative ranking curve 82%). Avelumab plus axitinib (AVEL-AXI) had the highest likelihood of an OS benefit for patients with favorable risk (65%). Pembrolizumab plus AXI (PEMBRO-AXI) had the highest likelihood of an OS benefit for patients with intermediate risk (78%). PEMBRO plus lenvatinib (PEMBRO-LENV) had the highest likelihood of an OS benefit for patients with poor risk (89%). PEMBRO-LENV was associated with a superior PFS benefit across all risk groups (89-98%). Maximal ORR was achieved with PEMBRO-LENV (97%). The highest likelihood for CR was attained with NIVO plus ipilimumab (NIVO-IPI; 85%) and PEMBRO-LENV (83%). The highest grade 3-4 TRAE rate occurred with PEMBRO-LENV (95%) and NIVO-CABO (83%), but the latter was associated with the lowest TRDD rate (2%). By contrast, NIVO-IPI had the lowest grade 3-4 TRAE rate (6%) and the highest likelihood of TRDD (100%). HRQoL consistently favored NIVO-CABO (66-75%), PEMBRO-LENV (44-85%), and NIVO-IPI (65-93%) in comparison to the other treatments.

CONCLUSIONS

IO-TKI drug combinations are associated with consistent improvements in clinically relevant outcomes for all mRCC risk groups. This benefit may be at the cost of higher TRAE rates; however, lower TRDD rates suggest a manageable side-effect profile. Longer follow-up is required to determine if the benefits of IO-TKIs will be sustained and if they should be favored in the first-line treatment of mRCC.

PATIENT SUMMARY

Combination treatments based on immunotherapy agents continue to show meaningful benefits in the first-line treatment of metastatic kidney cancer. Our review and network meta-analysis shows that immunotherapy combined with another class of agents called tyrosine kinase inhibitors is promising. However, longer follow-up is needed for this treatment strategy to clarify if the benefits are long-lasting.

Patterns of care for patients with non-metastatic castration-resistant prostate cancer: Population-based study in Ontario, Canada

53

Background: To describe patterns of practice of PSA testing and imaging for Ontario men receiving continuous androgen deprivation therapy (ADT) for the treatment of non-metastatic castration-resistant prostate cancer (nmCRPC). Methods: This was a retrospective, longitudinal, population-based study of administrative health data from 2008 to 2019. Men > 65 years old receiving continuous ADT with documented CRPC were included. An administrative proxy definition was applied to capture patients with nmCRPC patients and excluded those with metastatic disease. Patients were indexed upon progression to CRPC and were followed until death or end of study period to assess frequency of monitoring with PSA tests and conventional imaging. A 2-year look-back window was used to assess patterns of care leading up to CRPC, as well as baseline covariates. Results: At a median follow-up of 40 months, 944 patients with CRPC were identified. Their median time from initiation of ADT to CRPC was 26 months, 61% of patients had their PSA measured twice or fewer in the year prior to index and 71% patients did not receive any imaging in the year following progression to CRPC. Almost all patients (98%, n = 921/944) in the study progressed to high-risk CRPC (HR-CRPC) during the study period, of which more than half received fewer than 3 PSA tests in the year prior to progression to HR-CRPC, and 31% received no imaging in the subsequent year. Conclusions: PSA testing and imaging studies are under-utilized in a real-world setting for the management of nmCRPC, including those at high-risk of developing metastatic disease. Infrequent monitoring impedes proper risk stratification, disease staging, detection of treatment failure and/or metastases, likely delaying necessary treatment intensification with life-prolonging therapies. Adherence to guideline recommendations and the importance of timely staging should be reinforced to optimize patients’ outcome.

Analysis of the role of PI3K-AKT and DNA damage repair (DDR) genomic biomarkers as predictors of clinical outcomes in nonmetastatic castration-resistant prostate cancer (nmCRPC)

175

Background: Clinically relevant outcomes in nmCRPC treated with androgen receptor-axis-targeted therapies (ARAT) may be inferior in patients with tumors harboring mutations bypassing androgen receptor signalling. This final update of a retrospective, multicenter analysis explores the association between genomic mutations in the PI3K-AKT and DDR signalling pathways with ARAT treatment outcomes in nmCRPC patients. Methods: Relevant clinical endpoint were collected for high-risk nmCRPC patients treated with an ARAT at APCaRI affiliated cancer centers, including median metastasis-free survival (MFS), overall survival (OS), PSA decline ≥ 50% (PSA50), and second progression free survival (PFS2). Archival tumor tissue was accessed for next generation gene sequencing, examining for genomic alterations in 500 genes, including those involved in the DDR and the PI3K-AKT signalling pathways. Comparison of outcomes of patients with DDR and PI3K-AKT pathway mutations was conducted using Cox proportional hazards regression using wildtype cases as the reference group, adjusting for PSA doubling time and pelvic lymphadenopathy. Results: Of the 37 patients included, 30 (82%) received apalutamide, 5 (13%) received darolutamide, and 2 (6%) received enzalutamide. 10 patients (27%) had PI3K-AKT pathway mutations (4 PTEN, 3 PIK3Ca, 2 PIK3C2G, 1 PIK3C2b), 8 patients (22%) had DDR gene mutations (3 ATM, 2 CHEK1, 1 BRCA2, 1 CDK12, 1 CHEK2, 1 FANCD2, 1 FANCL), and 1 patient (3%) had 2 MLH1 mutations (microsatellite instability). Of those who had subsequent treatment, 1 received enzalutamide and 5 received abiraterone. Patients with PI3K-AKT pathway mutations had significantly shorter MFS (4.8 mo; HR 4.2; 95% CI 1.2 – 15.0; p = 0.025). Those with DDR mutations had a trend towards shorter MFS (23.3 mo HR 3.7; 95% CI 0.71 – 13.4; p = 0.134). OS data remains immature. 4 (11%) patients did not achieve PSA50, including a patient with 2 MLH1 mutations. Conclusions: This final analysis demonstrates that nmCRPC with PI3K and DDR signalling pathway mutations have poor clinical outcomes when treated with ARAT, likely secondary to decreased reliance on the androgen receptor signalling pathway. These results highlight the potential value of exploring targeted therapies, such as PARP or AKT inhibitors in patients with these mutations.

Determining the cancer diagnostic interval using administrative data in a cohort of patients with pancreatic cancer

336

Background: Pancreatic cancer is a leading cause of cancer death, largely due to vague presenting symptoms and late stage at diagnosis. Population-based administrative data can be a valuable resource for studying the diagnostic interval. The objective of this study was to determine the first encounter in the diagnostic interval and to calculate that interval in a cohort of patients with pancreatic cancer using an empirical approach. Methods: This is a retrospective, cohort study of patients diagnosed with pancreatic ductal adenocarcinoma (PDAC) from 2007 – 2015 in Alberta, Canada. We used the Alberta Cancer Registry (ACR), physician billing claims, hospital discharge and emergency room visits to identify health encounters that occurred more frequently in the 3 months prior to diagnosis compared to those in the 3-24 months prior to diagnosis. We used statistical control charts to define the lookback period for each encounter category and identify the earliest encounter that represented the start of the diagnostic interval (index contact date). The end of the interval was the diagnosis date. Quantile regression was used to determine factors associated with the diagnostic interval. Results: We identified 3142 patients with PDAC. Median age of diagnosis was 71 (IQR 61-80). We identified an index contact date in 96.5% of the patients. The median length of the diagnostic interval was 76 days (IQR 21-191; 90th percentile 276 days). A higher Elixhauser comorbidity score (+18.57 days/ 1 point increase, 95% CI 16.07-21.07, p < 0.001) and stage 3 disease (+22.55 days, 95% CI 5.02-40.08, p = 0.01) was associated with a longer diagnostic interval. Conclusions: In this cohort of patients with pancreatic cancer, there was a wide range in the diagnostic interval with 10% of patients having a diagnostic interval approaching one year. Diagnostic interval research using administrative databases can understand variations in diagnosis times, can inform early detection efforts and can improve quality of care.

Association between oncology drug review times and public funding recommendations

99

Background: New oncology drugs undergo detailed review of clinical, economic, and patient data. Thoroughly assessing these data can require lengthy review processes, in the absence of accelerated approval pathways. The aim of this study was to assess how cancer drug review times impact public funding recommendations. Methods: Drugs reviewed by Canada’s health technology assessment body, the pan-Canadian Oncology Drug Review (pCODR), from April 2012 to November 2020 were included in this study. Data was collected including Health Canada approval date, initial and final funding recommendations, treatment intent, drug class, clinical indication (tumour type) and incremental cost-effectiveness ratios (ICER). Univariable and multivariable analyses were used to determine the association between funding recommendations and review times. Results: Of the 227 applications submitted to pCODR, 168 had received positive funding recommendations. Amongst the total drug applications, 24 (14.3%) drugs were intended for the treatment of thoracic cancers, 19 (11.3%) for gastrointestinal cancers, 17 (10.1%) for genitourinary cancers, 17 (10.1%) for breast cancer, and 91 (54.2%) for other tumour sites. Median time from pCODR submission to final recommendation was longer for drugs indicated for the treatment of lung and breast cancer compared to those indicated for treatment of other tumours (223 vs. 212 vs. 203 days, respectively; Kruskal-Wallis p = 0.0322). Drugs with longer review times were more likely to receive a negative pCODR recommendation, even when adjusting for ICER (157 vs 298 days, Wilcoxon p-value = 0.0003). There was no association between positive or negative funding recommendation and tumour type. Conclusions: Oncology drugs with longer review times are less likely to receive recommendation for public funding in Canada. Addressing factors contributing to variance in review times and standardizing the review process can ensure equitable access to cancer drugs.

Determining the cancer diagnostic interval using administrative data in a cohort of patients with pancreatic ductal adenocarcinoma (PDAC)

e13551

Background: PDAC is a leading cause of cancer death that is often diagnosed at an advanced stage. Population-based administrative data can be a valuable resource for studying the diagnostic interval, defined as the time from the first related healthcare encounter to cancer diagnosis. The objective of this study was to determine the diagnostic interval in a cohort of patients with PDAC using an empirical approach. Methods: This is a retrospective, cohort study of patients diagnosed with PDAC from 2007 – 2015 in Alberta, Canada. We used the Alberta Cancer Registry, physician billing claims, hospital discharge and emergency room visits to identify and categorize cancer-related healthcare encounters before diagnosis. We used statistical control charts to define the lookback period for each encounter category and used these lookback periods to identify the earliest encounter that represented the start of the diagnostic interval (index contact date). The end of the interval was the diagnosis date. Quantile regression was used to determine factors associated with the diagnostic interval. Results: We identified 3,142 patients with PDAC. Median age of diagnosis was 71 (IQR 61-80). We identified an index contact date and thus a diagnostic interval in 96.5% of patients. The median diagnostic interval length was 76 days (IQR 21-191; 90th percentile 276 days). A higher Elixhauser comorbidity score (+18.57 days/ 1 point increase, 95% CI 16.07-21.07, p<0.001) and stage 3 disease compared to stage 2 disease (+22.55 days, 95% CI 5.02-40.08, p=0.01) were associated with a longer diagnostic interval. Conclusions: In this cohort of patients with PDAC, there was a wide range in the diagnostic interval with 10% of patients having a diagnostic interval of approximately 9 months. Diagnostic interval research using administrative databases can understand variations in diagnosis times and can inform early detection efforts by identifying where and in whom delays may occur.

Impact of lenvatinib (LEN) dose-intensity and starting dose on survival among patients with advanced hepatocellular carcinoma (HCC): Results from a Canadian multicenter database (HCC CHORD)

e16142

Background: The REFLECT trial established LEN as a first-line treatment option for HCC. However, decreased LEN exposure is common due to adverse events leading to dose reductions and treatment discontinuations. The aim of this study was to evaluate whether starting dose or dose-intensity of LEN affects survival. To our knowledge, this is the first study to examine dosing of LEN and survival in HCC patients treated outside of Asia. Methods: From July 2018 to December 2019, HCC patients treated with first-line LEN from 10 different Canadian cancer centers were included. Overall survival (OS), progression-free survival (PFS), disease control rate (DCR) and objective response rate (ORR) were retrospectively analyzed and compared across different mean dose-intensities (> 66.7% vs <=66.7%) and starting dose groups (Full vs reduced). Survival outcomes were assessed with Kaplan-Meier curves and Cox proportional hazards models. DCR and ORR were determined radiographically according to the treating physician´s assessment in clinical notes and not RECIST 1.1 or mRECIST. Results: A total of 173 patients were included. Median age was 67 years, 77% were men and 23% East Asian. The most frequent causes of liver disease were hepatitis C (38%) and B (20%). 56% of patients received localized treatment prior to LEN. Of those, 24% had TACE, 6% TARE and 8% had liver transplant. Before starting LEN 31% of patients were ECOG 0 and 57% were ECOG 1. Most patients were Child-Pugh A (81%) and BCLC stage C (73%). Main portal vein invasion was present in 15% of the patients. Median follow-up was 4.5 months. LEN was started at full dose in 54% of patients and 60% had a mean dose intensity greater than 66.7%. ORR, PFS and OS results and their comparison between the different starting dose and dose-intensities are shown in the table. In a multivariate model that adjusted for age, gender, stage, ECOG, Child-Pugh, BCLC, cirrhosis, liver etiology disease (hepatitis B, C and non-viral), presence of tumor thrombus, prior transplant and localized treatment, dose intensity (>66.7 vs <=66.7% [HR 0.70, 95% CI 0.42-1.18; p=0.18]) was not a predictor of survival. Conclusions: In HCC patients starting LEN at a reduced dose does not appear to compromise survival. LEN dose-intensity of > 66.7% was associated with improved survival, but not after controlling for potential confounders.[Table: see text]

Real-World Outcomes of Oxaliplatin-Based Chemotherapy on R0 Resected Colonic Liver Metastasis

INTRODUCTION

In resected colonic liver metastasis (CLM), randomized studies of oxaliplatin-based chemotherapy have demonstrated improvements in disease-free survival (DFS), but not overall survival (OS). Additionally, oxaliplatin regimens have not been compared to non-oxaliplatin chemotherapy. Despite limited evidence, perioperative chemotherapy is often used in the management of CLM. The primary aim of this study was to assess the impact of oxaliplatin chemotherapy regimens on OS in patients who have undergone resection of CLM in a real-world setting.

PATIENTS AND METHODS

Patients who underwent resection of CLM in the provinces of Alberta and British Columbia, Canada, were identified from 1996 to 2016. Perioperative (pre- and/or post-) systemic therapy was categorized as oxaliplatin or non-oxaliplatin-based chemotherapy or no chemotherapy. The primary and secondary outcomes were OS and DFS, respectively.

RESULTS

We identified 511 patients who underwent R0 resection of CLM. A significant difference in median OS was identified among the oxaliplatin, non-oxaliplatin, and no-chemotherapy groups of 100, 60, and 59 months, respectively (P = .009). In multivariate analysis, patients who received oxaliplatin regimens had a lower risk of death (hazard ratio, 0.68; 95% confidence interval, 0.51-0.92; P = .012), whereas the non-oxaliplatin chemotherapy group did not (hazard ratio, 0.88; 95% confidence interval, 0.65-1.20; P = .422) compared with no chemotherapy.

CONCLUSIONS

In this multicenter, retrospective, population-based study, perioperative oxaliplatin-based chemotherapy was associated with improved OS in conjunction with R0 resection of CLM. Further studies should evaluate the optimal duration and sequencing of perioperative chemotherapy in relation to curative-intent surgical resection of CLM.

Retrospective cohort analysis of real-world clinical outcomes in nonmetastatic (M0) castration-resistant prostate cancer (CRPC) treated with novel androgen receptor pathway inhibitors (ARPI)

51

Background: The landscape of M0 CRPC has changed with the recent demonstration of metastasis-free survival (MFS) and overall survival (OS) improvements with the use of ARPIs in clinical trial settings. However, the extrapolation of this data to clinical practice is limited by strict exclusion criteria in these trials, including prior or concurrent malignancy, cardiovascular disease, or hypertension. The purpose of this study was to assess real-world outcomes in patients with M0 CRPC treated with ARPIs compared to historical controls. Methods: We designed a retrospective cohort study with the inclusion of patients in Alberta, Canada diagnosed with M0 CRPC between 2001-2020. Via chart review, we identified baseline characteristics, potential confounders, treatment details, and clinical outcomes. The primary outcome of interest was MFS. Secondary outcomes included: second progression-free survival (PFS2) and OS. Median survival times were measured using the Kaplan-Meier method and the log-rank test was used for comparison of outcomes based on ARPI exposure. Cox proportional hazard regression models were used to calculate hazard ratios (HR) accounting for impact of PSA doubling time (PSADT), use of osteoclast inhibiting agents, and presence of pelvic lymphadenopathy. Results: We identified 211 patients across multiple centres in Alberta with M0 CRPC, with 54 having received apalutamide (40/54), enzalutamide (7/54), or darolutamide (7/54). Median age at M0 CRPC diagnosis was 74 years; median PSADT was 4.4 months; and 19% of patients (40/211) had pelvic lymphadenopathy at diagnosis. Median MFS in patients treated with ARPIs was 47.5 months compared to 20.6 months in those not treated with ARPIs (HR, 0.23; 95% confidence interval [CI], 0.11-0.49; p < 0.001). Median PFS2 in ARPI treated patients was 66.3 months compared with 35.6 months (HR, 0.40; 95% CI, 0.18-0.87; p = 0.022). Median OS for patients treated with ARPI was not reached. Conclusions: Given the older age of men with advanced prostate cancer, real-world outcomes that include patients with comorbidities are important adjuncts to the interpretation of clinical trials exploring the benefit of novel therapeutics. Here, we demonstrate that in a real-world, unselected population of men with M0 CRPC, apalutamide, enzalutamide, and darolutamide seem to confer similar MFS and PFS2 benefits to those demonstrated in the SPARTAN, PROSPER, and ARAMIS studies. Real-world OS data remain immature and will be an important addition to these findings.

Sequence decision making for cabazitaxel (Cbz) versus abiraterone (Abt) or enzalutamide (Enz) post-docetaxel (Dtx) in a publicly funded health care system

57

Background: The TROPIC trial demonstrated an overall survival (OS) benefit of Cbz after Dtx in metastatic castrate-resistant prostate cancer (mCRPC). However, the novel anti-androgens (NAA) Abi and Enz have demonstrated similar improvements post-Dtx. The recent CARD trial suggests Cbz may provide the greatest OS benefit in selected patients who were rapid progressors ( < 12 months, RP) on first NAA, however Cbz use and efficacy in the real-world is uncertain. We sought to quantify the real-world use of Cbz and evaluate outcomes post-Dtx. Methods: mCRPC patients who received Dtx at the two tertiary referral centres in the Canadian province of Alberta from October 2012 (Cbz funding approval) to December 31st 2017 were assessed. We examined Cbz eligibility per TROPIC and CARD trial criteria, tracked therapies received, and documented objective and subjective reasoning for therapeutic decisions. OS was measured using the Kaplan-Meier method and the log-rank test was used to compare outcomes. The Chi-Square test was used to compare relative therapy utilization. Results: 463 mCRPC patients received Dtx over the study period, including 83 (18%) for castrate sensitive disease. At Dtx progression, 262 patients (56%) were eligible for Cbz per TROPIC trial criteria, while only 162 (62%) of those were RP on first NAA. Post-Dtx OS was lower among TROPIC-eligible patients receiving Cbz compared to those receiving Abi or Enz (9.1 vs 14.2 months, p = 0.001). This OS difference was not demonstrated among RP patients (11.2 vs 12 months, p = 0.664). The most common reasons for TROPIC ineligibility were Dtx intolerance (13%), serious comorbidities (12%), unacceptable blood counts (11%), performance status (9%) or, for CARD ineligible patients, no progression within 12 months on first NAA (38%). The most common agent immediately post-Dtx was Abi (n = 180, 39%), followed by Enz (n = 129, 28%). Significantly fewer patients (n = 56, 12%) received Cbz immediately post-Dtx (p = 0.001), and 149 (32%) received Cbz overall. First line post-Dtx, 286 patients (62%) did not have a documented discussion about Cbz, and in 172 cases (38%) consideration of Cbz was never documented. Patient choice against Cbz chemotherapy was recorded in 15% of discussions. Conclusions: In a real-world cohort of mCRPC patients, Cbz was a significantly less common choice than Abi or Enz after progression on Dtx. In a majority of these cases, no first line discussion of Cbz was documented, and in documented discussions, patient choice was the driving factor in a minority. OS post-Dtx in patients who met TROPIC trial criteria was lower for those receiving Cbz, noting that, unlike in TROPIC, these patients also received NAAs. This OS difference was not seen in those who also progressed rapidly on first NAA. These data suggest ongoing hesitation towards Cbz use in mCRPC and support careful selection of patients who may obtain benefit.

Clinical outcomes of patients with metastatic castration-resistant prostate cancer (mCRPC) receiving radium-223 (Ra-223) early versus late in the treatment sequence

136

Background: Ra-223 is the only targeted alpha therapy shown to prolong overall survival (OS) in men with mCRPC. The purpose of this real-world study is to evaluate clinical outcomes of patients (pts) when Ra-223 is used early (second-line) or late (third or later lines) for mCRPC. Methods: We used administrative databases in Ontario (2012-2017) to estimate OS from start of second-line life prolonging therapy (LPT), event free survival (EFS; time from start of second-line LPT to start of fourth-line or death) using counting process models, comparing 2 cohorts of patients: Early Ra-223 (2nd line) vs Late Ra-223 (3rd or later line). All models are adjusted for relevant fixed and time-varying covariates, including age, prostate specific antigen, hemoglobin, Charlson Comorbidity Index (CCI), use of bone health agents, prior systemic treatments, Gleason score, TNM score, Androgen Deprivation Therapy (ADT) and standardized pain score. Results: Data from 598 men in Ontario with mCRPC who received at least 2 lines of LPT, including Ra-223 in second line or later were analyzed (Early Ra-223=253; 42.3%; Late Ra-223=345; 57.7%). The mean age (standard deviation) at the start of first-line LPT was 72.2 years (8.8). Patients in the early Ra-223 cohort had a longer time from diagnosis of prostate cancer to receiving first-line LPT, a longer time from start of first-line to start of second-line LPT, and were less likely to receive docetaxel. The median number of Ra-223 cycles was 5 (range 3-6) and 4 (range 3-6) and the mean number lines of total LPT was 2.7 ± 0.9 (2.0 - 7.0) and 3.8 ± 0.9 (3.0 - 7.0) in the early and late Ra-223 cohorts, respectively. OS was better in the Early Ra-223 cohort compared to the Late Ra-223 cohort (Hazard Ratio [HR] 0.79; 95% Confidence Interval [CI] 0.66-0.95). EFS was better in the Early Ra-223 cohort compared to the Late Ra-223 cohort (HR 0.71; 95% CI 0.58-0.86). Time to first hospitalization and time to first emergency department visit was longer in the early Ra-223 cohort. Conclusions: Real-world data from Ontario, suggests that patients who received Ra-223 in second-line versus third-line or later had better outcomes. Patients who received Ra-223 early received less chemotherapy, but had better survival. The selection of patients who may benefit the most from Ra-223, and the optimal timing of the Ra-223 in the sequence of treatments, are being evaluated in a larger Canadian study including data from four Canadian provinces (REACTIVATE NCT04281147 ).

Efficacy and safety of lenvatinib in the real-world treatment of hepatocellular carcinoma: Results from a Canadian multicenter database (HCC CHORD)

275

Background: The REFLECT trial establishedlenvatinib (LEN) as a first-line treatment option for hepatocellular carcinoma (HCC). Compared to sorafenib (S), LEN has a higher objective response rate (ORR) and progression-free survival (PFS) with a slightly different toxicity profile. The aim of this study was to gather data regarding the efficacy and safety of LEN when used in the real-world treatment of HCC. To our knowledge, this is the first study to examine LEN use in HCC patients treated outside of Asia. Methods: HCC patients treated with LEN from 10 cancer centers in the Canadian provinces of British Columbia, Alberta, Ontario and Nova Scotia between July 2018 to July 2020 were included. Overall survival (OS), PFS, disease control rate (DCR) and ORR were retrospectively analyzed and compared across first- and second-to-fourth line use of LEN. ORR was determined radiographically according to the treating physician´s opinion in clinical notes and not RECIST 1.1 or mRECIST. Toxicities were also examined. Results: A total of 220 patients were included in this analysis. Median age was 67 years, 80% were men and 25.5% East Asian. The most frequent causes of liver disease were hepatitis C (37%) and B (26%). 62% of patients received any localized treatment before LEN, of those 26% had TACE, 15% TARE and 7.7% had liver transplant. Before starting LEN 29% of patients were ECOG 0 and 59% were ECOG 1. Most patients were Child-Pugh A (81%) and BCLC stage C (75.5%). Main portal vein invasion was present in 14% of the patients. Median follow-up was 4.5 months. A total of 173 patients (79%) received LEN as first line therapy and 47 patients (21%) were treated in second-to-fourth line. Of patients receiving LEN in first line, 22 (13%) started treatment with S, but switched to LEN before progression due to poor tolerance of S. ORR, DCR, PFS and OS are shown in the table. Toxicities occurred in 86% of patients and led to dose reductions in 76 (35%) patients and drug discontinuation in 53 (24%) patients. The most common side effects were fatigue (59%), hypertension (41%), decreased appetite (25%) and diarrhea (22%). Conclusions: Outcomes of HCC patients treated in Canada with LEN in the first line are comparable to those demonstrated in the REFLECT trial, despite the inclusion of Child-Pugh B and ECOG >1 patients. LEN use in second or later lines also showed similar outcomes, although more conclusions are difficult to draw due to the small numbers. LEN appears to be effective and safe in real world practice outside of Asia in first- and second-to-fourth line treatment of HCC. [Table: see text]

Systemic therapy for nonmetastatic castrate-resistant prostate cancer (M0 CRPC): A systematic review and network meta-analysis (NMA)

113

Background: The treatment landscape for M0 CRPC has changed following the demonstrated efficacy of new agents in recent randomized control trials (RCT). However, the comparative effectiveness of these novel agents is unknown. This NMA indirectly compared the efficacy and safety of available therapies for M0 CRPC. Methods: A literature search of MEDLINE (Ovid), EBM Reviews, HealthSTAR, PubMed, PubMed Central, CINAHL, and TRIP Database was performed. Studies were screened by two independent reviewers. Hazard ratios (HR) and confidence intervals were extracted for the primary outcome metastasis-free survival (MFS) and the secondary outcomes overall survival (OS) and grade 3 or higher adverse events (AE). Bone MFS was used as a surrogate for MFS when MFS was not reported. Risk of bias was assessed using the Cochrane Collaboration tool. A Bayesian NMA was performed using a fixed-effects model. Results: Four RCT were analyzed (n=5549). Each trial compared either apalutamide (APA), enzalutamide (ENZA), darolutamide (DARO), or denosumab (DENO) plus androgen deprivation therapy (ADT) to placebo plus ADT. Risk of bias was low. For MFS, APA and ENZA had similar efficacy (Table), and Surface Under the Cumulative Ranking Analysis demonstrated a 59% probability that APA was preferred for MFS, followed by ENZA (41%). There was a trend for improved OS for APA, DARO and ENZA, but no meaningful differences between these agents. APA, ENZA, and DARO had a similar risk of AEs and all had a greater risk of AEs compared to DENO. Conclusions: APA and ENZA appear to be the most efficacious treatments for MFS in M0 CRPC, though more data for OS is required. Compared to DARO, APA and ENZA’s demonstrated efficacy is not at the expense of added toxicity.[Table: see text]

Metastasis-free survival as a predictor of overall survival in non-metastatic castration-resistant prostate cancer

201

Background: Recent trials have shown that apalutamide and enzalutamide can improve metastasis free survival (MFS) in advanced non-metastatic (M0) castrate-resistant prostate cancer (CRPC). MFS is a novel clinical endpoint, demonstrated to be a strong predictor of overall survival (OS) for localized prostate cancer, yet it is unknown if this is also true for M0 CRPC. Our aim was to determine how strongly MFS in M0 CRPC correlates with OS in a real world population. Secondary analyses evaluated whether a rapid PSA-doubling time (PSADT), of ≤10 months, impacts outcomes. Methods: We performed an analysis of patients diagnosed with advanced prostate cancer, followed at the Tom Baker Cancer Centre, in Calgary, Alberta from 2001-2017. Patients were excluded if they did not develop M0 CRPC. MFS and OS were measured using the Kaplan-Meier method and the log-rank test was used to compare outcomes based on PSADT. Correlation between OS and MFS was determined using Pearson Correlation and Kendall’s Tau-B. Results: A total of 1310 patients were identified with advanced prostate cancer, of which 87 developed M0 CRPC. The median age of diagnosis of M0 CRPC was 72 years, with a median Gleason score of 7.0, initial PSA of 10.4, and PSADT of 5.1 months. Only 6 patients were treated with second-generation anti-androgens or chemotherapy. Median MFS and OS after M0 CRPC diagnosis were 44.1 and 83.7 months, respectively. Pearson Correlation between MFS and OS was strong with a coefficient of 0.850 (p < 0.001); with non-parametric Kendall’s Tau, correlation was also strong with a coefficient of 0.632 (p < 0.001). PSADT ≤10 months was identified in 70 patients, and associated with a significantly shorter MFS, compared to a PSADT > 10 months (40.2 vs. 90.4 months; p = 0.001), as well as shorter OS (76.2 vs. 104.3 months; p = 0.008). Conclusions: MFS for M0 CRPC is strongly correlated with OS in a real world population. PSADT of ≤10 months seems to be a useful prognostic tool in estimating MFS and OS in patients with M0 CRPC. MFS was better than expected even in patients with a PSADT of ≤10 months, which may due to our adherence to the biochemical definition of castration-resistant disease, as well as lack of standard imaging intervals in the real world.

Treatment of hepatocellular carcinoma (HCC) after sorafenib (S) over the last 10 years

438

Background: Until recently there were no standard treatments for HCC patients after S. This study characterizes subsequent treatments (STx) received by HCC patients over the past 10 years and assesses their impact on survival. Methods: HCC patients treated with S between 01/2008 – 06/2017 in British Columbia, Alberta, and two cancer centers in Toronto, Ontario, Canada (Princess Margaret and Sunnybrook Cancer Centre) were included. Clinical, pathologic, laboratory, treatment, and outcome data were collected. The Kaplan-Meier method was used to assess overall survival (OS) based on STx, and stratified according to a better prognostic group (BPG), defined as ECOG 0-1 and CP-A, and worse prognostic group (WPG), defined as ECOG≥2 or CP-B/C. Results: A total of 730 patients were identified. 177 (24.2%) received STx (table). Patients who received STx had longer median OS (mOS) than those who had no further treatment (12.1 vs. 3.3 months; p < 0.001). For patients treated with localized, systemic, or palliative radiation treatment, mOS was 16.8, 10.5 and 8.6 months, respectively (p < 0.001). After S, there were 206 (30.7%) patients in the BPG and 464 (69.3%) in the WPG. BPG patients were more likely to receive STx compared to WPG patients (60.5% vs. 39.5%, p < 0.001). BPG patients who received STx had better mOS than those who did not (15.9 vs. 7.0 months; p < 0.001). WPG patients also had better mOS if they received STx compared to those who did not (6.0 vs. 2.6 months; p < 0.001). Conclusions: Only a small proportion of HCC patients received subsequent treatment after sorafenib. This is likely due to poor performance status, liver dysfunction, or lack of treatment options. Patients who received subsequent treatment had improved mOS, regardless of whether they were in the better or worse prognostic group. [Table: see text]

Real world eligibility for cabozantinib (C), regorafenib (Reg), and ramucirumab (Ram) in hepatocellular carcinoma (HCC) patients after sorafenib (S)

422

Background: The CELESTIAL, RESORCE, and REACH-2 trials showed survival benefit of C, Reg, and Ram, respectively, when given after S to HCC patients. However, strict eligibility criteria (SEC) may limit generalizability. In clinical practice, modified eligibility criteria (MEC) may be used to offer treatments to select patients with slightly worse performance status (ECOG 2) or limited liver dysfunction (Child-Pugh (CP) B7). This study evaluated which patients in the real world would be eligible for these new treatments using SEC and MEC, and their prognostic impact. Methods: HCC patients who received S between 01/2008-06/2017 in British Columbia, Alberta, Princess Margaret Cancer Centre, and Sunnybrook Cancer Centre in Canada were included. Clinical, pathologic, laboratory and outcome data were collected. Patients were classified as eligible or ineligible based on available CELESTIAL, RESORCE, REACH-2 clinical trial SEC or MEC. Median overall survival (mOS) for these groups was assessed using the Kaplan-Meier method. Results: A total of 730 patients were identified. Using SEC, only 13.1% of patients would be eligible for C, Reg, or Ram (table). Expanding eligibility to include patients who meet MEC increased the proportion of eligible patients to 31.7%. Patients who met SEC had longer mOS compared to those who were ineligible. The most common reasons for not meeting SEC across all 3 trials were ECOG ≥ 2 (61.7%) and CP ≥ B (63.9%). Higher ineligibility for Reg or Ram was likely driven by strict trial-specific criteria, with 28.0% of patients ineligible for Reg due to S intolerance and 58.9% ineligible for Ram due to AFP < 400. Conclusions: Only a small proportion of real-world HCC patients would be eligible for C, Reg, or Ram based on SEC. More than twice as many patients would likely receive treatment if MEC were applied. If MEC are adopted, ongoing real-world evidence generation will be important to evaluate outcomes in these unstudied patient groups. [Table: see text]

A real-world, population-based study of patterns of referral, treatment, and outcomes for advanced pancreatic cancer

Background

To describe patterns of referral, consultation, and treatment of advanced pancreatic cancer patients in a population‐based health care system and to evaluate the impact of these factors on outcomes.

Methods

This is a retrospective analysis of population‐based cancer data from the province of Alberta, Canada. We analyzed patients diagnosed with either locally advanced or metastatic pancreatic adenocarcinoma from 2009 to 2016 and evaluated their patterns of referral to a cancer center, consultation with oncology, and treatment with active anticancer therapies. Logistic regression models were constructed to determine the factors associated with referral, late oncology assessment, and late receipt of treatment.

Results

We identified 1621 pancreatic cancer patients. Median age was 70 years, 50% were men, and 51% had a Charlson index of 2+. Within this cohort, only 884 (54%) patients were referred to one of the provincial cancer centers. Adjusting for confounders in logistic regression models, older age and worse comorbidity scores were associated with nonreferral (both P < 0.01). In multivariable analysis among treated patients, the following factors were associated with improved overall survival, including younger age, earlier stage, and better comorbidity scores (all P < 0.01). Neither referral to consultation times nor consultation to treatment times correlated with outcomes. Importantly, nonreferred patients were more likely to use acute care services, including longer total duration of hospitalizations and more frequent visits with physician specialists.

Conclusion

A significant proportion of patients with advanced pancreatic cancer were never referred to a cancer center. Nonreferred patients were more likely to utilize specific health care resources.

Report from the 19th annual Western Canadian Gastrointestinal Cancer Consensus Conference; Winnipeg, Manitoba; 29-30 September 2017

The 19th annual Western Canadian Gastrointestinal Cancer Consensus Conference (wcgccc) was held in Winnipeg, Manitoba, 29-30 September 2017. The wcgccc is an interactive multidisciplinary conference attended by health care professionals from across Western Canada (British Columbia, Alberta, Saskatchewan, and Manitoba) who are involved in the care of patients with gastrointestinal cancer. Surgical, medical, and radiation oncologists; pathologists; radiologists; and allied health care professionals participated in presentation and discussion sessions for the purpose of developing the recommendations presented here. This consensus statement addresses current issues in the management of colorectal cancer.

Prognosis of patients with hepatocellular carcinoma treated with sorafenib: a comparison of five models in a large Canadian database

Several systems (tumor-node-metastasis [TNM], Barcelona Clinic Liver Cancer [BCLC], Okuda, Cancer of the Liver Italian Program [CLIP], and albumin-bilirubin grade [ALBI]) were developed to estimate the prognosis of patients with hepatocellular carcinoma (HCC) mostly prior to the prevalent use of sorafenib. We aimed to compare the prognostic and discriminatory power of these models in predicting survival for HCC patients treated with sorafenib and to identify independent prognostic factors for survival in this population. Patients who received sorafenib for the treatment of HCC between 1 January 2008 and 30 June 2015 in the provinces of British Columbia and Alberta, and two large cancer centers in Toronto, Ontario, were included. Survival was assessed using the Kaplan-Meier method. Multivariate Cox regression was used to identify predictors of survival. The models were compared with respect to homogeneity, discriminatory ability, monotonicity of gradients, time-dependent area under the curve, and Akaike information criterion. A total of 681 patients were included. 80% were males, 86% had Child-Pugh class A, and 37% of patients were East Asians. The most common etiology for liver disease was hepatitis B (34%) and C (31%). In all model comparisons, CLIP performed better while BCLC and TNM7 performed less favorably but the differences were small. The utility of each system in allocating patients into different prognostic groups varied, for example, TNM poorly differentiated patients in advanced stages (8.7 months (m) (95% CI 6.5-11.5) versus 8.4 m (95% CI 7.0-9.6) for stages III and IV, respectively) while ALBI had excellent discrimination of early grades (15.6 m [95% CI 13.0-18.4] versus 8.3 m [95% CI 7.0-9.2] for grades 1 and 2, respectively). On multivariate analysis, hepatitis C, alcoholism, and prior hepatic resection were independently prognostic of better survival (P < 0.01). In conclusion, none of the prognostic systems was optimal in predicting survival in sorafenib-treated patients with HCC. Etiology of liver disease should be considered in future models and clinical trial designs.

Differential utilization of preventive care among colon cancer survivors compared to non-cancer patients

577

Background: With advances in diagnosis and treatment, many cancer patients survive more than 5 years. The care of these cancer survivors (CS) represent an area of unmet need. We aim to characterize the patterns of preventive care in colon CS compared to non-cancer controls (NCC) and identify areas of deficiencies within the context of a universal health care system. Methods: Adult patients with non-metastatic colon cancer treated at the BC Cancer Agency between 2000-2012 were included. An age and gender matched cohort constructed from the provincial database served as NCC. Areas of preventive care examined include vaccinations, cancer, osteoporosis and cardiovascular diseases (CVD) screening. Multivariate regressions were done to test for associations between CS and preventive care. Results: In total, 9381 colon CS and 47187 NCC, matched at a ratio of 1:5, were analyzed. Among CS, median age of diagnosis was 68, 58% were male and 47% had stage 3 disease. The median overall survivals were 12/10/8 years for stages 1/2/3 disease respectively. 61% of these survivors died from colon cancer, 12% from other cancers and 25% from non-cancer causes. Deaths from colon cancer are more common within 5 years of diagnosis, particularly stage 3 disease. CS were more likely to receive any preventive care. In CS compared to NCC, 90% vs 85%, 47% vs 39% and 53% vs 46% of eligible patients had CVD screening, cancer screening and other preventive care respectively. This remained significant in multivariate analyses (Table). Patients who were female, had higher income and resided in urban areas were more likely to receive screening. Among CS, patients > 65 years (OR1.2, p = 0.04 95%CI 1.0-1.4), females (OR 1.5, p < 0.01 95%CI 1.3-1.8) and stages 1 or 2 disease (OR 1.3, p < 0.01 95%CI 1.1-1.5) had higher uptake of screening. Conclusions: Many colon cancer patients are long term survivors. CS are more likely to receive screening than NCC but uptake is suboptimal in certain areas. Targeted education towards certain sub-groups such as males, ≤65 years, low income and rural area patients may improve long term health outcomes. [Table: see text]

A real-world comparison of multi-modality therapies in locally advanced gastro-esophageal junction (GEJ) cancers

96

Background: Trials show that addition of systemic therapy and/or radiation to surgery improves survival in GEJ cancers. However, the different regimens have not been directly compared. We examined population-based outcomes of 3 treatments: 1) neoadjuvant carboplatin and paclitaxel plus radiation (CROSS); 2) perioperative epirubicin, cisplatin, and fluoropyrimidine (MAGIC); and 3) cisplatin and fluoropyrimidine with radiation (CisFP). Methods: We reviewed patients diagnosed with GEJ cancer from 2005 to 2015 who received CROSS, MAGIC, or CisFP at 2 tertiary, 4 regional, and 11 community cancer centers in Alberta, Canada. Survival was assessed with Kaplan-Meier curves and compared with the log-rank test. A Cox proportional hazards model was constructed to evaluate the impact of treatment on overall survival (OS). Results: 331 patients were identified. Median age was 63 (IQR 56-69) years and 86% were men. CROSS was used in 217 (65%) cases followed by CisFP in 72 (22%) and MAGIC in 42 (13%). Age, sex, and stage were not associated with treatment selection (all p > 0.05), but a higher proportion of CROSS and CisFP patients had adenocarcinoma (86% and 85%, respectively) compared to MAGIC patients (41%) ( p < 0.01). CROSS and MAGIC correlated with higher surgical resection rates when compared to CisFP (82% vs. 79% vs. 50%, respectively, p < 0.01). Median OS favored CROSS and MAGIC rather than CisFP, but this was not statistically significant (29 vs. 34 vs. 20 months, respectively, p= 0.17). Adjusting for confounders, OS remained similar for MAGIC (HR 0.8, 95%CI 0.5-1.3, p= 0.36) and CisFP (HR 0.7, 95%CI 0.5-1.1, p= 0.10) when compared to CROSS. In addition, age > 65, advanced stage, and lack of surgical resection were associated with increased risk of death (HR 1.5, 95%CI 1.1-2.0, p= 0.02, HR 2.2, 95%CI 1.2-3.9, p< 0.01 and HR 4.1, 95%CI 2.8-5.9, p< 0.01, respectively). Conclusions: OS was similar across all 3 regimens, but outcomes were inferior to those seen in original trials. This observation suggests that GEJ patients in routine practice are different from study participants or that treatment selection may be driven by factors other than trial eligibility criteria.

Relationship between ethnicity and overall survival (OS) in patients with advanced hepatocellular carcinoma (HCC) treated with sorafenib (S)

248

Background: The SHARP and Asia Pacific (AP) trials showed that S improves OS compared to placebo in advanced HCC. However, OS was worse in the AP trial which included predominantly Asian patients. It is unclear if ethnicity, or perhaps Hepatitis B (HBV) infection, is a poor prognostic factor for these patients. The purpose of this study was to determine whether ethnicity affects OS in patients with advanced HCC being treated with S. Methods: All patients treated with S for HCC in Alberta, Canada from 01/2008 to 07/2016 were included. Patient demographics and clinical/pathological variables were retrospectively collected. Patients were dichotomized by ethnicity as either East Asian or not according to a validated list of surnames. Survival outcomes were assessed with Kaplan-Meier curves and compared with the log-rank test. A Cox-proportional hazard model was constructed with ethnicity and relevant clinical/pathological characteristics to assess their impact on survival. Results: A total of 175 patients were included. Mean age was 64 years. 78% were men, 28% were East Asian, and 79% were Child-Pugh A at initiation of S. The most common etiologies of underlying liver disease were Hepatitis C (HCV) (31%), HBV (29%) and alcohol (21%). 42% had distant metastatic disease. The majority of patients had an ECOG performance status of 0 (26%) or 1 (64%). Median OS was 9.0 months in Asians and 9.5 months in non-Asians (p = 0.68). On multivariate analysis, ethnicity (HR 0.76, 95% CI 0.39 – 1.32, p = 0.33) was not a significant prognostic factor for OS. However, lack of distant metastases (HR 0.57 95% CI 0.40 - 0.82, p < 0.01), initial AFP < 400 (HR 0.54 95%CI 0.38 - 0.78, p < 0.01) and 3+ localized treatments (HR 0.60 95% CI 0.39 - 0.92, p = 0.02) were associated with better OS. Further, HBV was associated with inferior OS when compared to HCV (HR 2.12, 95% CI 1.08 - 4.17, p = 0.03). Conclusions: When treated with S in one Canadian province, ethnicity does not appear to be a prognostic factor for OS. However, HCV infection, lack of distant metastases, initial AFP < 400, and 3+ previous localized treatments were significant prognosticators of OS. We are validating these findings in a larger multi-centre Canadian dataset.

A real world multicenter study of first (1L) and second (2L) line treatment patterns and outcomes in advanced pancreatic cancer (APC)

476

Background: FOLFIRINOX (FFX), gemcitabine plus nab-paclitaxel (GN), and gemcitabine (gem) are 3 publicly funded and available treatment options for locally advanced (LAPC) and metastatic pancreatic cancer (MPC) in Canada since 2014. Without head-to-head trials that directly compare all 3 regimens, treatment selection and outcomes in 1L and 2L remain poorly characterized in routine clinical practice. Methods: Data from 4 tertiary, 8 regional, and 28 community hospitals in Canada were pooled. LAPC and MPC patients diagnosed from 2014 onwards and who received at least 1 line of systemic therapy were included. Analyses were conducted to identify predictors of treatment choice and to determine the relationship between treatment patterns and overall survival (OS) from APC diagnosis to death. Results: We identified 279 eligible patients. Median age was 64 (IQR 56-69) years, 55% were men, and 46% were ECOG ≥2. There were 27% LAPC and 73% MPC. In the 1L setting, FFX and GN were given in 44% and 41% of patients, respectively, and gem in 15%. GN was the preferred multi-agent therapy in worse ECOG patients (66% in ECOG 2+ vs 21% in ECOG 0, p = .001) and in more recently diagnosed cases (63% in 2016 vs 25% in 2014, p = .001). 1L treatment selection was not influenced by other baseline characteristics, such as age, sex, tumor location, or LAPC vs MPC status (all p > 0.05). A total of 91 patients proceeded to subsequent therapies, of whom 55 (60%), 27 (30%), and 9 (10%) had received 1L FFX, GN, and gem, respectively. In the 2L setting, GN after 1L FFX (41/55; 75%) and fluoropyrimidine (FP) after 1L GN (21/27; 78%) were the most common sequential approaches. Patients who underwent 2L therapy had better OS than those who did not (13 vs 7 months, p = .001). After adjusting for confounders, receipt of 1L FFX plus 2L GN or 1L GN plus 2L FP resulted in improved OS when compared to other treatment sequences (HR 0.43, 95%CI 0.28-0.67, p = 0.001 and HR 0.57, 95%CI 0.39-0.83, p = 0.004, respectively). Conclusions: One third of APC patients receive 2L therapy, highlighting the feasibility of 2L trials. Use of 1L multi-agent therapy followed by 2L non-cross-resistant regimens represents a reasonable treatment strategy for APC in the real world.

Patterns of futile care for comorbidities in colorectal cancer patients near the end of life

e18242

Background: Futile care includes interventions that no longer provide benefit, may cause significant harm, and/or lack sufficient utility to justify the required resources. Cancer care such as chemotherapy use has been shown to be overly aggressive near the end of life. We hypothesized that futile care for non-cancer comorbidities in terminal colorectal cancer (CRC) patients may also be prevalent and associated with specific clinical factors. Methods: We constructed a large retrospective cohort of decedents aged ≥18 years who were diagnosed with CRC during 2008-2012 and who died by the end of 2013. We analyzed population-based data through linked provincial files with information on cancer management, pharmacy records, hospital discharges, and vital statistics. We defined endpoints for futile care as ≥1 hospitalization and/or ≥1 prescription for statins, dihydropyridine calcium channel blockers (CCBs), or acid suppressants within 2 months preceding death. Different time points were used in sensitivity analyses. Results: We included 2,530 patients. Median age at CRC diagnosis and death was 70 (IQR 61-79) and 72 (IQR 62-80) years, respectively. Among them, 59% were men, 66% had colon cancer, and 87% were diagnosed with advance disease. Median time from diagnosis to death was 452 (IQR 201-800) days. In terms of futile care, 4.7% had hospitalizations, 8.7% received statins, 6.8% received CCBs, and 38.2% received acid suppressants within 2 months prior to death. In multivariate analyses, there were no clear associations between demographics and hospitalizations. With respect to medication use, advanced age was correlated with increased use of statins (OR 2.235, 95%CI 1.469-3.401, p < 0.001) and CCBs (OR 2.039, 95%CI 1.304-3.190, p = 0.002), but inversely associated with use of acid suppressants (OR 0.750, 95%CI 0.598-0.941, p = 0.013). Men were also more likely to receive statins (OR 1.653, 1.099-2.488, p = 0.016), but less likely to receive acid suppressants (OR 0.764, 0.608-0.960, p = 0.021). Conclusions: Even near death, a fair number of decedents with CRC continued to receive medications for comorbidities that were unlikely to provide clinically meaningful benefits. Futile care was more prevalent in men and in the elderly.

Comparison of prognostic models for hepatocellular carcinoma (HCC) in patients treated with the sorafenib: Results from a Canadian multi-center HCC database

e15653

Background: Several staging systems and models (TNM, BCLC, Okuda, CLIP and ALBI) have been developed to estimate the prognosis of patients with HCC. Most of these were developed prior to the prevalent use of sorafenib. The purpose of this study was to compare the prognostic and discriminatory power of these models in predicting survival for HCC patients treated with sorafenib. Methods: Patients who received sorafenib for the treatment of HCC between January 1, 2008 and June 30, 2015 in the provinces of British Columbia and Alberta, as well as Princess Margaret Cancer Centre and Sunnybrook Odette Cancer Centre in Toronto, Ontario were included. Survival outcomes for each model were assessed with Kaplan-Meier (KM) curves and compared with the log-rank test. Time dependent area under the curve (t-AUC) was used to test the discriminatory power of each model (higher t-AUC = more discriminatory power). Akaike information criterion (AIC), a measure of goodness-of fit of models while penalizing overly complex models, was used to compare the models (lower AIC = better model). Results: A total of 681 patients were included in this analysis. Median age was 64 years (range 8-91). Majority were males (80%), had a Child-Pugh score A (86%), ECOG performance status 0 (30%) and 1 (60%). 37% of patients were of East Asian ethnicity. Most common etiology for liver disease was hepatitis B (33%) and C (29%). At start of sorafenib, most patients were BCLC stage C (92%) and TNM stage IV (61%). The median overall survival for the entire cohort was 9.2 months (95% CI 8-10.4). See table below for t-AUC and AIC results. Conclusions: According to ourlarge multi-center study, CLIP appears to be the most informative in predicting survival in HCC patients treated with sorafenib. Prospective studies are needed to determine its role in patient selection for clinical trials and in guiding treatment decisions. The TNM and BCLC staging systems were the least useful in predicting survival in this population. [Table: see text]

The impact of peri-operative chemotherapy for patients with lymph node-positive urothelial cancer

388

Background: Patients with locally advanced urothelial cancer with regional lymph node involvement (LN+) have a poor prognosis. Surgical management of these patients is controversial and practice patterns vary. We evaluated the outcomes of patients with LN+ disease treated with pre-operative chemotherapy and cystectomy, cystectomy and post-operative chemotherapy, and chemotherapy alone. Methods: Patients with urothelial cancer with TxN1-3M0 disease treated with chemotherapy in Alberta from 2005 to 2015 were evaluated. Progression-free survival (PFS) and overall survival (OS) were evaluated using Kaplan-Meier analysis. Cox regression analysis was performed to evaluate the impact of age, gender, T stage, and N stage on survival. Results: 184 patients with LN+ disease treated with chemotherapy were evaluable for outcomes; 42 underwent pre-operative chemotherapy (Group A), 92 underwent post-operative chemotherapy (Group B), and 50 received chemotherapy alone (Group C). The median age at diagnosis was 65 years (range 31-89) and most patients (83%) were male. The median follow-up time was 23.2 months. A higher T stage was seen in patients in Group A, while patients in Group C had a higher N stage. The median number of chemotherapy cycles delivered was equal in all arms at 4. Patients in Group A or B had significantly better PFS and OS compared with patients in Group C (Table). When adjusting for age, gender, T stage, and N stage, patients in Group C had significantly lower OS compared with those patients in Group A (HR 1.87, 95% CI 1.09 – 3.18, p=0.02). Conclusions: In this real-world analysis of patients with LN+ urothelial cancer, patient outcomes were improved with surgical resection of disease in combination with pre-operative chemotherapy. After chemotherapy in fit patients with LN+ disease, surgical management is a reasonable consideration. [Table: see text]

Cost-effectiveness of surveillance after curative resection (CR) of metastatic colorectal cancer (CRC)

526

Background: Surveillance after CR of stage I-III CRC is recommended by most major oncology organizations to detect asymptomatic recurrences. Such recurrences are more likely to benefit from early interventions such as CR of metastases. Only the NCCN recommends a surveillance schedule after CR of metastases that includes CEA testing, imaging and clinical evaluation every 3-6 months for 2 years, and then every 6-12 months in years 3 to 5. Periodic endoscopy is also recommended. It is unclear if there is cost-effective surveillance strategy for metastatic CRC after CR. Methods: A Monte Carlo micro-simulation model was constructed using a 1-month cycle length and 10 year time horizon. Surveillance strategies were compared based on NCCN guidelines, with testing every 3 months (3M) or 6 months (6M), as well as two alternate strategies of testing every 12 months (12M) or no surveillance (None) for 5 years. Recurrence, repeat CR rates, and survival outcomes were modeled from population-based outcomes of 257 patients who had CR of mCRC in British Columbia, Canada. Asymptomatic recurrences were more likely to undergo CR, compared to symptomatic ones. Additional costs, utilities, and probabilities were derived from the literature. Costs are in 2015 CAD and utilities in Quality-adjusted life years (QALY), and both discounted at 3% and half-cycle corrected. Analyses were performed using TreeAge Pro with 1000 trials and 1000 distribution samplings. Results: The incremental cost-effectiveness ratio (ICER) and net monetary benefit (NMB) are listed in the Table. Increasing the frequency of surveillance tests does lead to modest gains in QALY, however, the cost of surveillance and subsequent treatment is high. Using a willingness to pay threshold of 150 000 CAD, the 6M strategy would be favored. Conclusions: In the Canadian context, the optimal surveillance strategy after CR of mCRC matches with the 6M strategy recommended by the NCCN. An additional Canadian data set will be used to externally validate the model outcomes. [Table: see text]

Impact of primary tumor sidedness on survival after resection of colorectal liver metastases (CRLM)

694

Background: The impact of primary tumour sidedness has recently been demonstrated in patients with metastatic colorectal cancer (mCRC). Differences in right (R) versus left (L) sided mCRC may be due to differences in consensus molecular subtyping. Clinically predictive mutations in ras, ( kras, nras and braf) may also help drive some of the differences in outcome. However, patients with mCRC who undergo surgical resection of CRLM often have a good prognosis. The aim of this study was to assess the impact of tumour sidedness on OS after resection of CRLM. Methods: Patients who underwent resection of CRLM in the province of Alberta, Canada were identified from 2004-2016. Tumour sidedness was determined by chart review, with R from the cecum to transverse and L from splenic flexure to sigmoid. Where available, ras mutational status was collected. OS was measured from the time of CRLM resection to death or last follow-up using the Kaplan-meier method. R and L were compared using the log-rank test and a Cox regression model. Results: 471 patients were identified who underwent resection of CRLM for mCRC, including 204 R and 267 L. Median age was 65, 63% male, with 54% synchronous metastatic disease, and 67% with a Charleson comorbidity index of 0. All ras wildtype was present in 22% of cases, any ras mutation was detected in 21% and 57% were unknown at the time of analysis. The median OS of R was 45 months, compared to 72 months for L, log-rank p = 0.01. After adjusting for potential confounders with a Cox-proportional hazard model, R compared to L remained significant, with a HR of 1.4 (95% CI 1.0-1.9, p = 0.02). ras mutational status was also significant for ras mutant, HR 2.4 (95% CI 1.7-3.3) and ras unknown HR 2.2 (95% CI 1.5-3.1) compared to ras wildtype p < 0.01. Conclusions: Primary tumour sidedness continues to have an impact on OS in mCRC, even when disease is managed surgically with resection of CRLM. Though limited by numbers, the impact remained significant even after controlling for potential confounders, including ras mutational status. Additional ras testing is underway. Further molecular classification may provide a biologic rationale for the observed differences.

Outcomes of perioperative systemic therapy (ST) in patients with R0 resection of metastatic colorectal cancer (mCRC)

496

Background: Adjuvant fluoropyrimidine (FP) +/- oxaliplatin (OX) ST improves overall survival (OS) following curative resection of stage II or III CRC, while other regimens do not. Utility of pseudo-adjuvant ST in mCRC patients who achieved R0 resection of their metastases remains controversial. We aim to describe population-based outcomes based on choice of ST. Methods: Patients diagnosed with mCRC from 2003 to 2010 and referred to any 1 of 5 cancer centers in British Columbia, Canada were reviewed. We categorized patients who underwent a successful R0 resection of their metastases into 3 groups based on receipt of peri-operative ST: 1) FP alone; 2) OX-based; and 3) non-standard or no ST. We compared OS using multivariate Cox regression models that adjusted for potential confounders. Results: We identified and reviewed 1,641 patients with mCRC among whom 225 achieved R0 resection of their metastases. In this cohort, median age was 63 years (Interquartile range (IQR) 55-70), 118 (52%) were men, 196 (87%) reported ECOG 0/1, 149 (66%) had a colonic primary, and 103 (46%) presented with de novo metastatic disease. The site of metastatic resection was hepatic in 144 (64%), pulmonary in 34 (15%), locoregional in 11(5%), and other in 36 (16%). A total of 122 (54%) received standard ST. Regimens included 28 (12%) FP alone, 94 (42%) OX-based, 25 (11%) irinotecan and 38 (17%) including bevacizumab. The median duration of ST was 10 cycles (IQR 7-12), with 56% and 44% delivered pre- and post-operatively, respectively. In multivariate analyses, liver or lung involvement (HR 0.60, 95% CI 0.40-0.88, p=0.01) predicted for improved OS when compared to metastases affecting other organs or sites. Receipt of OX-based ST was the only regimen associated with better OS, while the rest were not (Table). Conclusions: Findings from this population-based cohort of mCRC suggest that use of a defined course of OX-based ST for pseudo-adjuvant intent around the time of R0 resection of liver or lung metastases is correlated with improved OS. [Table: see text]

Local tumor control and survival outcomes of percutaneous radiofrequency ablation plus post-ablation chemotherapy for lung tumors in nonsurgical patients: A meta-analysis

9550

Background: Local recurrence is a frequent outcome post radiofrequency ablation (RFA) for local control of lung tumors. We sought to examine local tumor control and survival benefits of RFA plus post-ablation chemotherapy versus RFA alone for management of lung tumors in non-surgical patients. Methods: Search strategy: MEDLINE, the Cochrane Library, and EMBASE databases from January 2000 to December 2012. Inclusion criteria: RFA +/- post-ablation chemotherapy in non-surgical patients with solid lung tumors. Exclusion criteria: Post-RFA radiation therapy, biologics, brachytherapy, or other ablation modalities. Outcomes: Local tumor progression (LTP), overall survival (OS), and disease-free survival (DFS) at 12 month follow-up. Statistical analysis: Fixed effect analyses, bias assessment, and sensitivity analyses (BioStat Inc., NJ, USA). Results: RFA plus post-ablation chemotherapy group: 11 clinical studies, 684 patients (mean age 64 years [range 50 to 74]; 434 men, ECOG ≤2), ablation of 1,314 lung tumors, with a 4:1 ratio being <3cm versus ≥3cm in diameter, and a 1:4 ratio being primary versus metastatic. RFA alone group: 38 clinical studies, 1,874 patients (mean age 65 years [range 49 to 75]; 1,041 men, ECOG ≤2), ablation of 2,604 lung tumors, with a 2.1:1 ratio being <3cm versus ≥3cm in diameter, and a 1:1 ratio being primary versus metastatic. RFA plus post-ablation chemotherapy versus RFA alone: LTP of 15% over median follow-up of 31 months [range 12 to 59]) versus 19% over median follow-up of 21 months [range 12 to 29]); OR 0.73 (95% CI: 0.61-0.86, p<0.05) at 12 month follow-up. OS was 89% versus 78%, respectively, at 12 month follow-up; OR 1.52 (95% CI: 1.16-2.00, p=0.003). DFS was 90% versus 82%, respectively, at 12 month follow-up; OR 3.18 (95% CI: 2.04-4.96, p<0.05). Sensitivity analyses were robust, publication bias relatively narrow, and heterogeneity within acceptable limits; Q statistic<21; p>0.13 for all outcomes. Conclusions: This meta-analysis reveals that RFA plus post-ablation chemotherapy of lung tumors yields improved outcomes in terms of LTP, OS, and DFS compared with RFA alone.