Upper gastrointestinal (GI) and hepato-pancreatico-biliary (HPB) cancers have poor prognoses and limited therapeutic options. Basket trials are being increasingly employed in these disease groups and have led to regulatory approvals for tumour agnostic prescribing. This study aimed to explore the success of emerging basket trials including patients with upper GI and HPB cancers, assessing phase of trial, agents used, selection biomarkers and endpoints favoured. A systematic review was conducted using the combined Ovid and Medline databases, including publications released up to and including April 2022, looking for primary studies reporting the results of basket trials in which patients with upper GI and HPB cancers were eligible for enrolment. A supplementary search on ClinicalTrials.gov was conducted in July 2022. Data on trial duration, treatment used, and endpoints recorded, were collected. The literature search identified 164 studies, while ClinicalTrials.gov identified 74: 23 ineligible studies were excluded, based on irrelevance to oncology; the remaining 215 studies were assessed. Fifty-three eligible basket studies were identified; 33 (62%) industry led, 20 (38%) academic. Of the 53 trials (non-randomised), 30 (57%) were phase II, 16 (30%) combined phase I/II and 7 (13%) were phase I; conducted in multiple institutions (not mutually exclusive): USA (N=34), Asia (N=28), Europe (N=17), Australia (N=6), Africa (N=1). The commonest therapeutic agents used were immunotherapy (N=16 (30%)), followed by PARP inhibitors (N=9 (17%)), EGFR inhibitors and HER2-directed agents (N=8 (15%) each). The biomarkers for entry selection were HER2 (N=10 trials), tumour mutation burden, DNA damage repair genes and BRAF/MEK (N=5 trials each), CDKN2A, NTRK, PDL1, NF1 (N=3 each), IDH1/2 (N=2), EGFR, VEGF, FGFR, KRAS (N=1 each), other (N=6). Ten trials did not require biomarker selection for entry. Eighteen trials had either full or interim results reported: N=407 total patients had advanced, previously-treated upper GI/HPB cancers (performance status 0-2) (range 1-43 per trial). The patients had biliary tract cancer (BTC) (N=249 (61.2%)), Oesophago-gastric (OG) (N=82 (20.1%)), pancreatic (N=38 (9.3%)), small bowel (N=36 (8.8%)), and hepatocellular carcinoma (HCC) (N=2 (0.5%)). The primary endpoint for 49 of 53 trials (92%) was response rate (RR). Median RR for all upper GI and HPB cancers across all 18 reported trials was 18.2%. Upper GI cancers had a median RR of 25%, HPB cancers 17.5% and HCC 0%. Five agents (from 18 trials) (28%) have gained tumour agnostic prescribing approval in these disease groups, based on these results: Pembrolizumab, Larotrectinib, Entrectinib, Dostarlimab and Dabrafenib combined with Trametinib. The trials included in this study were early phase, conducted mainly in USA, Asia and Europe, with the most favoured therapeutic agents used being immunotherapy, likely reflecting the era of study. More patients with BTC were included, possibly due to more limited treatment options for rarer malignancies. The majority reported RR as the primary endpoint, allowing earlier read outs, with promising tumour agnostic approval in these disease groups. Evolving basket trial and statistical methodology with dedicated cohorts for these patients with emerging subgroups, will potentially lead to improved patient outcomes.

Isocitrate dehydrogenase 1 mutation ( IDH1m ) is detected in approximately 13% of patients with cholangiocarcinoma (CCA). In the global, phase 3 ClarIDHy study evaluating ivosidenib (IVO), an oral inhibitor of the IDH1m protein, and placebo (PBO) in patients with previously treated, non-resectable or metastatic IDH1m CCA, IVO significantly improved progression-free survival (PFS) compared to PBO (HR=0.37, p < 0.0001), and demonstrated a favourable safety profile. A quantitative benefit-risk assessment is a useful method to summarize the evidence and explicitly trade off the benefits and risks in comparing IVO with PBO. We present the results of a quantitative benefit-risk assessment of IVO compared to PBO from the ClarIDHy study. A panel of seven experts determined relevant key benefit and risk criteria for IVO and PBO in order to create a ‘value tree,’ and agree appropriate models to assess data from the ClarIDHy study via elicitation to minimize potential bias. The benefit criteria were determined based on efficacy endpoints and quality of life (QoL) subscales utilised in the ClarIDHy study: 3-month and 6-month PFS rate; 6-month OS rate and 6-month OS rate from the rank preserving structural failure time model (to adjust for crossover from PBO to IVO); objective response rate and QoL subscales (EORTC-QLQ-C30 Physical Functioning, Pain and Appetite Loss subscales; EORTC-QLQ-BIL21 Pain and Eating subscales). The risk criteria determined by the expert panel included common adverse events (AEs) associated with IVO: ECG QT prolongation; gastrointestinal events (diarrhoea; retching and vomiting; nausea); fatigue; AEs leading to treatment discontinuation or dose reduction. A ‘Scale Loss Score’ (SLoS) model (primary analysis) was applied to the data from ClarIDHy to estimate the probability that the benefit-risk profile of IVO was better than that of PBO. Sensitivity analyses were also applied to the data (product model; linear model and random weights analysis) to confirm or refute the results of the SLoS model. The primary analysis showed a 95.24% probability that the benefit-risk profile of IVO was better than the benefit-risk profile of PBO. Sensitivity analyses applying the SLoS model to alternative sets of benefit and risk criteria in the value tree showed consistently high probability (of >95%) for the benefit-risk profile in favour of IVO versus PBO for all endpoints evaluated. Sensitivity analyses using the linear model and the product model also showed a strong consistency of a high probability in favour of IVO compared to PBO for all endpoints evaluated (linear model: >99%; product model: >94%). Robustness of the results was demonstrated by the random weights analysis showing a similar trend in favour of IVO with all weights and the results converging quickly towards the main analysis results. A consistent positive result in support of a favourable benefit-risk profile for IVO compared to PBO was demonstrated. These results support and provide comprehensive evidence that IVO is an effective treatment with a tolerable safety profile for this aggressive, life-threatening disease, in line with the previously reported PFS, OS and safety data.

Pancreatic and biliary tract cancers (BTCs) are associated with extremely poor prognoses due to patients frequently presenting at an advanced stage, limited treatment options and lack of screening programmes. Treatment for these malignancies may involve a 5-fluorouracil (5-FU)-based chemotherapy regimen. 5-fluorouracil requires Dihydropyrimidine dehydrogenase (DPD) for catabolism. Dihydropyrimidine dehydrogenase deficiency occurs due to a Dihydropyrimidine dehydrogenase gene (DPYD) mutation and increases the risk of patients experiencing 5-FU toxicity and adverse effects. This retrospective study aimed to determine the DPD deficiency frequency in patients with pancreatic and BTCs (all stages) who received 5-FU-based treatment in a tertiary referral centre and assessed adverse events (AEs) as per the Common Terminology Criteria for Adverse Events (CTCAE) version 5. Patients commencing 5-FU-based chemotherapy for pancreatic and BTCs were identified. The institution database was used to access patient records, review evidence of DPYD mutation, type of cancer, and assessment of tolerance to treatment as per the CTCAE version 5. Patients with pancreatic or BTC who had been tested for DPD deficiency were eligible for the study. Details of chemotherapy regimens for each patient and evidence of any subsequent alteration in the dose of 5-FU-based treatment was also reviewed. Between 06/07/2020 and 19/04/2021, all (100%) of 291 patients starting a 5-FU-based chemotherapy regimen were tested for DPD deficiency. Of the 198 patients with pancreatic cancer, 83.8% of patients received treatment in the palliative setting, 8.6% adjuvant and 8.1% neoadjuvant. Of the 93 patients with BTC, 60.2% of patients received treatment in the palliative setting, 37.6% adjuvant and 2.2% neoadjuvant. Of the total cohort tested, 9.3% had a DPD deficiency; 17 of 198 patients (8.6%) with pancreatic cancer and 10 of 93 patients (10.8%) with BTCs had a DPD deficiency (all underwent a 50% reduction of 5-FU starting dose). Three DPYD gene mutations were identified in the overall cohort: c.1129-5923C>G (15 patients (5.2%)), c.2846A>T (9 patients (3.1%)), and c.1905+1G>A (3 patients (1.0%)). In the entire cohort of patients with DPD deficiency who received 5-FU at 50% dose (N = 21, (N = 10 adjuvant and N = 11 palliative)), 90% experienced AEs; 73.7% of patients experienced a grade 1 AE, 47.7% a grade 2 AE, and 36.8% a grade 3 AE. The most common grade 3 AEs were diarrhoea, neutropenia, and thrombocytopenia. No patients experienced a grade 4 AE. Furthermore, 11.1% of these 21 patients were given and tolerated a dose escalation to 75% of the standard dose, 7.4% patients required a dose decrease, and 19% required treatment-related cessation of 5-FU-based therapy. No patients tolerated a dose increase to 100%. This study provides reference DPD frequencies and associated AEs for patients with pancreas and BTC receiving 5-FU-based chemotherapy. In the entire cohort, 9.3% had a DPD deficiency; 36.8% of these patients experienced treatment-related grade 3 AEs even with the reduced 50% dose of 5-FU, with some patients requiring either a further dose reduction or cessation of 5-FU. However, a minority of patients tolerated a dose increase to 75%.

The BILCAP clinical trial established adjuvant capecitabine as the current standard of care treatment after biliary tract cancer resection. Translational work from this clinical trial involved collecting archived fixed formalin tissue from consented BILCAP patients and carrying out low-pass whole genome (lp-WGS), targeted gene (TGS) and RNA sequencing (RNAseq) for copy number (CN), mutation and gene-fusion analysis. In total, 98 patients underwent RNAseq, 95 of 98 underwent lp-WGS and 39 of 98 underwent TGS. 47 patients had intrahepatic cholangiocarcinoma, 47 patients had gallbladder cancer, 2 patients had perihilar cholangiocarcinoma, and 2 patients had distal cholangiocarcinoma. 62 (63.3%) patients were female, and 48 (49.0%) received adjuvant capecitabine. FGFR2 gene fusions were present in 24 patients (24.5%), as were fusions in NTRK1 (n=4, 4.1%), FGFR1 (n=4, 4.1%), FGFR3 (n=2, 2.0%) and FGFR4 (n=2, 2.0%). Known pathogenic mutations were seen in IDH1(n=4, 10.3%, total number of mutations=8, 20.5%), IDH2 (n=1, 2.6%, total n=6, 15.4%), and FGFR2 (n=1, 2.6%, total n=7, 17.9%). Commonly amplified (CN ≥ 4) genes included NTRK1 (n=28, 29.5%), ERBB2 (n=27, 28.4%) and MDM2 (n=20, 21.1%) with MYC (n=19, 20.0%), EGFR (n=16, 16.8%) and MET (n=15, 15.8%) also amplified. Nearly all the alterations investigated did not significantly affect recurrence risk (PFS) or overall survival (OS), including FGFR2 fusions (OS hazard ratio (HR) 1.11 p=0.762, PFS HR 1.10 p=0.763). However, the presence of amplified EGFR (CN ≥ 4) significantly decreased both OS (HR 5.40 p=0.01) and PFS (HR 3.44 p=0.04). The BILCAP cohort shows a wide variety of driver and potentially targetable mutations in unselected biliary tract cancer patients, comparable to similar datasets. Of note, patients with EGFR amplification had significantly reduced OS and PFS. This indicates that EGFR amplification may be an important indicator in determining prognosis and could provide an attractive target for future targeted anti-cancer therapy in biliary tract cancer.

Surgery with radical intent is the only potentially curative option for entero-pancreatic neuroendocrine tumors (EP-NETs) but many patients develop recurrence even after many years. The subset of patients at high risk of disease recurrence has not been clearly defined to date. The aim of this retrospective study was to define, in a series of completely resected EP-NETs, the recurrence-free survival (RFS) rate and a risk score for disease recurrence. This was a multicenter retrospective analysis of sporadic pancreatic NETs (PanNETs) or small intestine NETs (SiNETs) [G1/G2] that underwent R0/R1 surgery (years 2000–2016) with at least a 24-month follow-up. Survival analysis was performed using the Kaplan–Meier method and risk factor analysis was performed using the Cox regression model. Overall, 441 patients (224 PanNETs and 217 SiNETs) were included, with a median Ki67 of 2% in tumor tissue and 8.2% stage IV disease. Median RFS was 101 months (5-year rate 67.9%). The derived prognostic score defined by multivariable analysis included prognostic parameters, such as TNM stage, lymph node ratio, margin status, and grading. The score distinguished three risk categories with a significantly different RFS (p < 0.01). Approximately 30% of patients with EP-NETs recurred within 5 years after radical surgery. Risk factors for recurrence were disease stage, lymph node ratio, margin status, and grading. The definition of risk categories may help in selecting patients who might benefit from adjuvant treatments and more intensive follow-up programs.

Differences in pre-operative staging and surveillance after curative treatment for PDAC hamper interpretation of outcome data. This survey aimed to assess current practice and identify areas for improvement; it was circulated to members of the United Kingdom National Cancer Research Institute (NCRI) pancreatic cancer subgroup between 14/4-4/5 2021. A total of 23 responses were collected (medical oncologist 52.2%, surgeon 26.1%, radiation oncologist 13.0%, other 8.7%); the majority were Consultants (91.3%) working in tertiary care institutions (86.9%) who attended PDAC tumour boards (90.9%). For staging prior to curative surgery, all responders used computerised tomography (CT) (100%), and 61.1% used routine 18FDG positron emission tomography (PET) (16.7% used it only in specific occasions); only 38.9% used routine liver magnetic resonance imaging (MRI). In terms of surveillance following curative treatment, practice varied widely: 64.7% of responders considered imaging, tumour marker and clinical follow-up as routine practice after curative treatment, while 29.4% undertook follow-up without imaging; 5.9% did not offer any form of surveillance. Frequency of follow-up was either 6-monthly (60.0%), 3-monthly (26.7%), or variable (13.3%) and lasted for 5 years (73.3%), 2 years (6.7%), 3 years (6.7%), or other (13.3%). Surveillance imaging performed was by CT scanning in all cases (46.7% as routine, 6.7% if not done previously, 6.7% on occasions); none of the responders used FDG-PET (0%) or liver-MRI (0%). During surveillance, tumour marker (CA 19.9) was tested 6-monthly (66.7%), 3-monthly (40.0%), or annually (26.7%). Most (62.5%) stated that routine follow-up after curative treatment should be performed, but that clear evidence determining the impact on patient’s outcome was required. Pre-surgical staging with 18FDG-PET is not yet routine. Surveillance after curative treatment varies between institutions, both in terms of investigations performed (if any) and duration. Further guidance is required to establish standardised practice.

Previous studies support 18FDG-PET utility for staging of patients (pts) with non-metastatic PDAC (M0-PDAC); while its use has been adopted in some countries (i.e. NICE guidelines in the United Kingdom), its use in this setting is variable, especially in relation to non-resectable disease. This study explored the utility of performing 18FDG-PET in real-world clinical practice for pts with a working diagnosis of non-resectable M0-PDAC based on cross-sectional imaging.

Background: The role of tumour genomic profiling in the clinical management of well-differentiated neuroendocrine tumours (WdNETs) is unclear. Circulating tumour DNA (ctDNA) may be a useful surrogate for tumour tissue when the latter is insufficient for analysis. Methods: Patients diagnosed with WdNETs underwent ctDNA genomic profiling (FoundationLiquid®); non-WdNETs (paraganglioma, goblet cell or poorly-differentiated neuroendocrine carcinoma) were used for comparison. The aim was to determine the rate of: test failure (primary end-point), “pathological alterations” (PAs) (secondary end-point) and patients for whom ctDNA analysis impacted management (secondary end-point). Results: Forty-five patients were included. A total of 15 patients with WdNETs (18 ctDNA samples) were eligible: 8 females (53.3%), median age 63.2 years (range 23.5–86.8). Primary: small bowel (8; 53.3%), pancreas (5; 33.3%), gastric (1; 6.7%) and unknown primary (1; 6.7%); grade (G)1 (n = 5; 33.3%), G2 (9; 60.0%) and G3 (1; 6.7%); median Ki-67: 5% (range 1–30). A total of 30 patients with non-WdNETs (34 ctDNA samples) were included. Five WdNETs samples (27.78%) failed analysis (vs. 17.65% in non-WdNETs; p-value 0.395). Of the 13 WdNET samples with successful ctDNA analyses, PAs were detected in 6 (46.15%) (vs. 82.14% in non-WdNETs; p-value 0.018). In WdNETs, the PA rate was independent of concomitant administration anti-cancer systemic therapies (2/7; 28.57% vs. 4/6; 66.67%; p-value 0.286) at the time of the ctDNA analysis: four, one and one samples had one, two and three PAs, respectively. These were: CDKN2A mutation (mut) (one sample), CHEK2mut (one), TP53mut (one), FGFR2 amplification (one), IDH2mut (one), CTNNB1mut (one), NF1mut (one) and PALB2mut (one). None were targetable (0%) or impacted clinical management (0%). There was a lower maximum mutant allele frequency (mMAF) in WdNETs (mean 0.33) vs. non-WdNETs (mean 26.99), even though differences did not reach statistical significance (p-value 0.0584). Conclusions: Although feasible, mutation-based ctDNA analysis was of limited clinical utility for patients with advanced WdNETs. The rates of PAs and mMAFs were higher in non-WdNETs. While patients with WdNETs could still be offered genomic profiling (if available and reimbursed), it is important to manage patients’ expectations regarding the likelihood of the results impacting their treatment.

Adequate design of clinical trials using QoL-based primary-end points to assess benefit derived from supportive interventions such as exercise, nutrition or complementary therapies is challenging in PDAC due to a lack of available data describing baseline QoL and changes over time for this patient population.

The ABC-06 clinical trial stablished ASC+mFOLFOX as the standard of care treatment after Cisplatin and Gemcitabine (CisGem) for ABC. Within the ABC-06 study, patients (pts) diagnosed with ABC (cholangiocarcinoma, gallbladder or ampullary cancer) with progression after CisGem were randomised (1:1) to ASC+mFOLFOX or ASC. Tumour markers (CA19.9, CEA and CA125), were scheduled at baseline (BSL) and at every follow-up visit. This post-hoc analysis explored if changes (stable/reducing vs increasing) of Ca19.9 at week 4 from initiation of ASC+mFOLFOX was associated with radiological progression-free survival (PFS). Secondary end-points included impact of raised BSL Ca19.9 (defined as x1.5 ULN) and chemotherapy-induced changes on overall survival (OS). CEA and CA125 were also analysed. Out 162 pts randomised, BSL Ca19.9 data was available for 135 pts. Paired BSL and week-4 Ca19.9 data was available for 37 pts in the ASC+mFOLFOX arm: Ca19.9 was stable/reducing in 17 (45.9%) and increasing in 20 pts (54.1%). Stable/reducing Ca19.9 showed a numerically longer median radiological PFS (4.3 months (m) vs 3.3m) but differences did not reach statistical significance (HR 1.08 (95% CI 0.55-2.14); p=0.81). When restricted to patients with raised BSL Ca19.9 (23 pts), impact on PFS was more marked (5.7m vs 3.2m), but remained non-significant (HR 1.68 (95% CI 0.70-4.01); p=0.23). Stable/reducing Ca19.9 at week 4 did not impact significantly on OS (p=0.56 (regardless of BSL Ca19.9 level; 37 pts); p=0.84 (if raised BSL Ca19.9; 23 pts)). Raised BSL Ca19.9 was associated both with shorter unadjusted clinical median PFS (3.2m vs 5.0m; HR 1.53 (95% CI 1.05-2.23); p=0.027) and unadjusted OS (4.4m vs 6.4m; HR 1.97 (95% CI 1.33-2.93); p When Cox Regression model (120 pts) for OS exploring the prognostic roles of raised BSL Ca19.9, CEA and CA125 was adjusted for pre-defined stratification factors (platinum sensitivity, albumin, and stage) and randomised trial arm, each raised tumour marker had an independent impact (HR 1.56 (95% CI 1.03 to 2.35); p=0.03 / HR 1.60 (95% CI 1,06 to 2.43) p=0.026 / HR 1.70 (95% CI 1.13 to 2.56); p=0.011 for Ca19.9, CEA and CA125, respectively). Unadjusted median OS was 8.9m if all tumour markers at BSL were non-raised, and 7.0m, 4.0m, 3.2m in the event of having 1, 2 or 3 raised BSL tumour markers. For ABC patients treated with second-line ASC+mFOLFOX, utility of Ca19.9 measured at week-4 after chemotherapy initiation is limited; raised BSL Ca19.9, CEA and CA125 have independent prognostic roles and future studies may need to consider these (individually or pooled) as stratification factors.

Biliary tract cancers (BTC) are a heterogeneous group of malignancies, including gallbladder cancer (GBC). Treatment for GBC is based on studies recruiting patients with all BTC primary sites rather than GBC alone. GBC represents a different molecular entity to other BTCs, which may impact the response to treatment. The benefit of chemotherapeutic regimens specifically for the treatment of patients with GBC, including the current standard first- and second-line regimens for BTC, is poorly understood. This study explored the benefit derived from palliative cytotoxic chemotherapeutic regimens in GBC. A systematic review and meta-analysis were designed and registered with the PROSPERO database prior to commencement (CRD42019155745). A systematic search was conducted on MEDLINE; key bibliographies were reviewed and selected annual conferences were used to identify articles. Eligible studies reported data on patients with advanced GBC treated with systemic chemotherapeutic regimens. Phase II and III trials, case series, and cohort studies were eligible; phase I studies and small cohorts (< 10 in the first line, < 5 in second/third line) were excluded. Data were pooled using random effects models. 3,035 studies were identified; 58 studies with 66 study arms (n= 1,986 patients with GBC) were eligible for meta-analysis. In patients with GBC, estimated pooled radiological overall response rates (ORR), mean progression-free survival (PFS) and overall survival (OS) were 23.2% [95% confidence interval (CI), 20.0-26.5], 4.8 months (95% CI, 4.3-5.2) and 8.3 months (95% CI, 7.6-8.9), respectively. In patients with GBC, the use of ≥3 chemotherapy agents in combination was associated with increased ORR [35.8% (95% CI, 25.4-46.8)], mean PFS [5.9 months (95% CI 5.2-6.7)] and OS [9.9 months (95% CI, 8.5-11.3). There was a significant improvement in ORR, disease-free survival (DFS), mean PFS and OS with increasing numbers of chemotherapeutic agents (all spearman rho coefficients=1; P< 0.001). Patients with GBC had a lower ORR than non-GBC BTC [odds ratio (OR) 0.65 (95% CI, 0.50-0.84)]; specifically, patients with GBC have lower ORR when compared with the individual subgroups of BTC: cholangiocarcinoma (not otherwise specified) [OR 0.63 (95% CI, 0.48-0.83)], intrahepatic cholangiocarcinoma [OR 0.51 (95% CI 0.32-0.83)] and extrahepatic cholangiocarcinoma [OR 0.64 (95% CI, 0.40-1.00)]. Patients with GBC respond differently to systemic therapy compared with other BTC primary sites; they achieve lower ORR. In GBC, increasing numbers of chemotherapy agents are associated with better outcomes. Although differences may be due to selection bias, intensification of chemotherapy in fitter patients may be investigated either in a GBC-specific trial or (with a planned statistical and reporting approach) within a large randomised BTC trial.

TPS592 Background: For advanced CCA, standard of care 1L systemic treatment is GEM + CIS. Genetic alterations in intrahepatic CCA provide potential therapeutic targets. Fibroblast growth factor receptor (FGFR) 2 gene rearrangements driving CCA tumorigenesis were identified almost exclusively in intrahepatic CCA patients (pts) (incidence, 10–16%). In phase 2, PEM (INCB054828), a selective, potent, oral FGFR1–3 inhibitor elicited an objective response rate (ORR) of 35.5% and median progression-free survival (PFS) of 6.9 months (mo) in previously treated, locally advanced or metastatic CCA with FGFR2 rearrangements (NCT02924376). FIGHT-302, a randomized, open-label, phase 3 study will evaluate efficacy and safety of 1L PEM vs GEM + CIS in unresectable/metastatic CCA with FGFR2 fusions or rearrangements (NCT03656536). Methods: Eligible pts are adults with confirmed unresectable/metastatic CCA; no prior systemic therapy for advanced disease < 6 mo before enrollment; radiographically measurable/evaluable disease (per RECIST v1.1); ECOG PS ≤1; documented FGFR2 fusions or rearrangements. Exclusions include clinically significant corneal or retinal disorder; history of calcium and phosphate homeostasis disorder or systemic mineral imbalance with ectopic soft tissue calcification; untreated CNS metastases or history of uncontrolled seizures. Pts will be randomized (1:1; stratified by region and tumor burden) to PEM 13.5 mg QD on a 21-day (d) cycle or GEM (1000 mg/m2) + CIS (25 mg/m2) on D1 and D8 of 21-d cycles (max 8). Crossover to PEM allowed after confirmed progression. PEM titration to 18 mg from cycle 2 allowed for pts without hyperphosphatemia (serum phosphate > 5.5 mg/dL) and Grade ≥2 treatment-related adverse events during cycle 1. Hyperphosphatemia will be managed with diet modifications, phosphate binders, diuretics, or dose adjustments. Treatment will continue until progression or unacceptable toxicity. Primary endpoint is PFS (by independent review). Secondary endpoints are ORR, overall survival, duration of response, disease control rate, safety, and quality of life. Four pts (target N = 432) are enrolled as of Sep 25, 2019. Clinical trial information: NCT03656536.

150 Background: PDAC is an aggressive cancer with median overall survival (OS) ~6 months (m). Methods: Pts with advanced PDAC due to start first-line chemo (Chx) completed 1 study survey and 2 quality of life (QoL) questionnaires (QLQ-C30 and PAN26) on 3 time-points: baseline (T1), before (T2) and after (T3) their 1st on-treatment CT scan, PIs filled in paired surveys T1/2/3. Results: 71 Pts and 12 PIs were recruited. 4% Pts died/deteriorated between consent and T1, 28% between T1 - T2 (~2.3m later), 13% between T2 - T3 (~1.3m), only 55% reached T3. 49% Pts preferred to share decision making with their PI, 31% Pts preferred to make the final decision about treatment after considering PI opinion. 86% Pts had personal goals to reach (12% PIs knew of these). 70% Pts were aware that Chx was unlikely to cure cancer however, they had much higher expectations than PIs (%, Table). Choosing between treatment options Pts prioritised: 54% OS, 26% balance between side-effects (SEs) and OS, 15% could not choose and 5% symptom control. These did not match with PIs (p<0.001). Pts who prioritised OS had higher symptom burden (p=0.03). OS was different in these Pts: who prioritised symptom control- OS 2.8m from T1, OS priority- 6.4m, could not choose- 8.7m and balance priority- 9.2m (p=0.01). At T1 Pts had low QoL (57/100, higher better). ‘Future worries’ and ‘planning of activities’ were scored the worst (58 and 56/100, lower better). Clinically significant (+/- 10p) worsening between T1/2/3 were: PF (10p), nausea/vomiting (NV, 12p), body image (18p p=0.01), taste (25p p=0.002). Financially 29% Pts were “a little/ lot” out of pocket at T1, 41% at T2 (p=0.036), 42% at T3 (p=0.034). Acceptability of SEs showed that least acceptable SEs (ranked 1-10, 10-least) were the ones Pts were already struggling with: NV (ranked 9/10, 46p/100 on QLQ), diarrhoea (8/10, 45/100); fatigue (less acceptable T3 (p=0.05), 46/100); appetite loss (7-8/10, 47/100). In contrast, most Pts (97%) were willing to take few/medium and around 50% large amount of SEs as a trade-off for extra time, whilst 9% PIs thought that Pts would accept excess SEs (p<0.001). Conclusions: Pts worry about future, have high symptom burden and high expectations of Chx. Pts want to be involved in decisions, but the contrasting views between Pts and PI show differing aims. [Table: see text]

Background: Hepatic resection offers a chance at cure in patients with colorectal liver metastases (CLM). Patient selection to attain R0 liver resection is a key pre-treatment decision in the HPB MDT but there is wide variation, and while Gd-EOB-DTPA MR imaging is recommended to determine CLM extent, it fails to capture other factors influencing decision such as functional capacity of the future liver remnant (FLR). This paper summaries a study protocol to evaluate the potential added value of two novel assessment tools to assess technical resectability - the LiMAx test (determining hepatic functional capacity) and a new Perspectum Diagnostics Imaging Suite (PDIS: providing additional MR-based volume and functional properties of the FLR). Methods and Results: This project will launch in May 2019 and run until December 2020, utilising a modified Delphi framework to harness international expertise in optimally framing a subsequent large-scale trial question. The study will progress via five workstreams: In order to achieve maximal international HPB community participation and increase the validity of the Results, this study will be presented where possible at key international meetings. Conslusion: This will be the first systematic study to establish a core set of factors influencing decision-making in CLM technical resectability. WS5 will quickly deliver a trial design driven by clinical equipoise obtained from HPB MDT members.

Searchable abstracts of presentations at key conferences in endocrinology ISSN 1470-3947 (print) | ISSN 1479-6848 (online)