Search

PURPOSE To evaluate early circulating tumor DNA (ctDNA) kinetics using a tumor-naïve assay and correlate it with clinical outcomes in early phase immunotherapy (IO) trials. METHODS Plasma samples were analyzed using a 425-gene next-generation sequencing panel at baseline and before cycle 2 (3-4 weeks) in patients with advanced solid tumors treated with investigational IO agents. Variant allele frequency (VAF) for mutations in each gene, mean VAF (mVAF) from all mutations, and change in mVAF between both time points were calculated. Hyperprogression (HyperPD) was measured using Matos and Caramella criteria. RESULTS A total of 162 plasma samples were collected from 81 patients with 27 different tumor types. Patients were treated in 37 different IO phase I/II trials, 72% of which involved a PD-1/PD-L1 inhibitor. ctDNA was detected in 122 plasma samples (75.3%). A decrease in mVAF from baseline to precycle 2 was observed in 24 patients (37.5%) and was associated with longer progression-free survival (hazard ratio [HR], 0.43; 95% CI, 0.24 to 0.77; P < .01) and overall survival (HR, 0.54; 95% CI, 0.3 to 0.96; P = .03) compared with an increase. These differences were more marked if there was a >50% decrease in mVAF for both progression-free survival (HR, 0.29; 95% CI, 0.13 to 0.62; P < .001) and overall survival (HR, 0.23; 95% CI, 0.09 to 0.6; P = .001). No differences in mVAF changes were observed between the HyperPD and progressive disease patients. CONCLUSION A decrease in ctDNA within 4 weeks of treatment was associated with treatment outcomes in patients in early phase IO trials. Tumor-naïve ctDNA assays may be useful for identifying early treatment benefits in phase I/II IO trials.

PURPOSE Antibiotic exposure before immune checkpoint inhibitor (ICI) treatment can negatively affect outcomes through alteration in the gut microbiome, but large-scale evaluations are lacking. We performed a population-level retrospective cohort study to evaluate the impact of antibiotic exposure before starting ICI on overall survival (OS). PATIENT AND METHODS Patients with cancer, age 65 years or older, who initiated treatment with ICIs between June 2012 and October 2018 in Ontario, Canada, were identified using systemic therapy administration data. The cohort was deterministically linked to other health care databases to obtain covariates and antibiotic prescription claim data at both 1 year and 60 days before ICI therapy. Multivariable Cox models evaluated the association between exposure and OS. RESULTS Among the 2,737 patients with cancer who received ICIs, 59% and 19% of patients received antibiotics 1 year and 60 days before ICI therapy, respectively. Median OS was 306 days. Any antibiotic exposure within 1 year before ICI was associated with worse OS (adjusted hazard ratio [aHR], 1.12; 95% CI, 1.12 to 1.23; P = .03). In antibiotic class analysis, exposure to fluoroquinolones within 1 year (aHR, 1.26; 95% CI, 1.13 to 1.40; P < .001) or 60 days before ICI (aHR, 1.20; 95% CI, 0.99 to 1.45; P = .06) was associated with worse OS, with a dose effect seen on the basis of total weeks of exposure over 1 year (aHR, 1.07 per week; 95% CI, 1.03 to 1.11; P < .001) and 60 days (aHR, 1.12 per week; 95% CI, 1.03 to 1.23; P = .01). CONCLUSION In this population-level study, exposure to antibiotics and specifically fluoroquinolones before ICI therapy was observed to be associated with worse OS among older adults with cancer. Interventions aimed at altering the gut microbiome to boost immunogenicity may help improve outcomes for patients receiving ICIs with prior antibiotic exposure.

PURPOSE Testicular cancer survivors may experience mental illness as a consequence of their cancer diagnosis and treatment. METHODS All incident cases of testicular cancer treated with orchiectomy in Ontario, Canada (2000-2010), were identified using the Ontario Cancer Registry. Cases were matched to controls in a 1:5 ratio on age and geography. Population-level databases were used to identify mental health service use episodes; outpatient use included visits to a general practitioner for a mental health concern or any visit to a psychiatrist. Negative binomial regression modeling was used to estimate the rate of mental health service use in the pretreatment (2 years prior until 1 month before orchiectomy), peritreatment (1 month before until 1 month after orchiectomy), and post-treatment periods (1 month after orchiectomy until end of follow-up). Rate ratios (RR) comparing cases with controls in the peri- and post-treatment periods were adjusted for baseline mental health service use. RESULTS Two thousand six hundred nineteen cases of testicular cancer were matched to 13,095 controls. There was no baseline difference in the rate of mental health service use. Cases were significantly more likely than controls to have an outpatient visit for a mental health concern in the peritreatment (adjusted RR [aRR], 2.45; 95% CI, 2.06 to 2.92) and post-treatment periods (aRR, 1.30; 95% CI, 1.12 to 1.52). The difference in mental health service use persisted over a median follow-up of 12 years. In the postorchiectomy period, cases with baseline mental health service use were those most likely to use mental health services (aRR, 5.64; 95% CI, 4.64 to 6.85). CONCLUSION Testicular cancer survivors use mental health services more often than healthy controls. Survivorship care plans that address the long-term mental healthcare needs of this population are needed.

115 Background: To mitigate the risks of chemotherapy associated neutropenia, during the COVID-19 pandemic, all genitourinary (GU) cancer patients treated with chemotherapy at the Princess Margaret Cancer Centre (PMCC) were offered primary prophylaxis with GCSF. We hypothesize that this reduced rates of febrile neutropenia, hospitalizations, healthcare costs and improved overall outcomes, compared to GU cancer patients treated with chemotherapy without GCSF in the 2 years prior to the pandemic. Methods: We performed a retrospective review of GU cancer patients, receiving curative or palliative intent chemotherapy, with or without primary GCSF prophylaxis between January 2018 and June 2022. GCSF was given either as a single dose or as consecutive doses post chemotherapy. Main outcomes were incidence of febrile neutropenia, hospitalization, health care expenditures as well as disease specific outcomes. Results: Overall, 248 patients with prostate cancer (44%), urothelial cancers (33%) germ cell (21%), and rare GU cancers (4%) were identified. Median age was 70 (range 19-91), 92% were male, 65% were ECOG 0/1. Treatment intent was neoadjuvant (13%), adjuvant (20%), or palliative (67%). Main regimens used were docetaxel, cabazitaxel, carboplatin, cisplatin/etoposide, gemcitabine/cisplatin and BEP. Median follow-up was 10.5 months (0.23-52.3 months). A total of 206/248 received primary GCSF prophylaxis. During chemotherapy, the median white blood cell levels were higher in the GCSF group compared to the non-GCSF group (14.1*10*9/L vs 2.90*10*9/L, p

PURPOSE Active surveillance (AS) for testicular nonseminomatous germ cell tumors (NSGCT) is widely used. Although there is no consensus for optimal treatment at relapse on surveillance, globally patients typically receive chemotherapy. We describe treatment of relapses in our non–risk-adapted NSGCT AS cohort and highlight selective use of primary retroperitoneal lymph node dissection (RPLND). METHODS From December 1980 to December 2015, 580 patients with clinical stage I NSGCT were treated with AS, and 162 subsequently relapsed. First-line treatment was based on relapse site and extent. Logistic regression was used to explore factors associated with need for multimodal therapy on AS relapse. RESULTS Median time to relapse was 7.4 months. The majority of relapses were confined to the retroperitoneum (66%). After relapse, first-line treatment was chemotherapy for 95 (58.6%) and RPLND for 62 (38.3%), and five patients (3.1%) underwent other therapy. In 103 (65.6%), only one modality of treatment was required: chemotherapy only in 58 of 95 (61%) and RPLND only in 45 of 62 (73%). Factors associated with multimodal relapse therapy were larger node size (odds ratio, 2.68; P = .045) in patients undergoing chemotherapy and elevated tumor markers (odds ratio, 6.05; P = .008) in patients undergoing RPLND. When RPLND was performed with normal markers, 82% required no further treatment. Second relapse occurred in 30 of 162 patients (18.5%). With median follow-up of 7.6 years, there were five deaths (3.1% of AS relapses, but 0.8% of whole AS cohort) from NSGCT or treatment complications. CONCLUSION The retroperitoneum is the most common site of relapse in clinical stage I NSGCT on AS. Most are cured by single-modality treatment. RPLND should be considered for relapsed patients, especially those with disease limited to the retroperitoneum and normal markers, as an option to avoid chemotherapy.

2589 Background: The molecular networks comprising circadian rhythm are expressed in immune cells where they affect immune-related processes. Within the emerging field of chrono-immunotherapy, it has been proposed that immunotherapies should be applied at certain times of day to optimize efficacy. Two recent reports have suggested that earlier administration of immune checkpoint inhibitors (ICIs) in non-small cell lung cancer and melanoma may offer improved survival outcomes (Karaboué et al. ASCO 2021, PMID 34780711). Whether this observation applies to other tumour types, or occurs in geographic regions with differing seasonality, is not known. Methods: We retrospectively analyzed the time of administration data from the INSPIRE single-centre, phase II, multi-cohort study of pembrolizumab (200 mg IV over 1 hour, q3w to maximum 35 cycles) in patients with advanced solid tumours (NCT02644369). Kaplan-Meier methods were used to estimate PFS and OS. Cox proportional hazards models were fitted to assess the association between time of administration and PFS or OS, adjusting for cohort. Fisher’s exact test was used to test for association with immune-related adverse events (irAEs). Results: A total of 106 patients (19 head and neck squamous cell, 22 triple negative breast, 21 epithelial ovarian, 12 melanoma, 32 other solid tumours) were accrued between March 21, 2016, and May 9, 2018. Median time of follow-up was 11.5 months. Start of infusion times were obtained for 806 total doses. The median time of administration was 15h06 for the first dose and 15h11 for all doses (range 09h19 – 18h34). No differences in PFS or OS were observed between patients who received their first dose before or after noon, or before or after 15h06. Furthermore, no differences in PFS or OS were observed between patients who received ≥ 50% of their doses before or after noon, or before or after 15h11. There was also no difference in PFS or OS between patients who did or did not have a significant proportion of doses (≥ 20%) after 16h30 (“evening” in previous reports). There were no differences in the frequency of grade ≥ 2 irAEs amongst the various groups. No differences in efficacy were found when individual cohorts were evaluated separately. Finally, no differences in PFS or OS were observed when participants were grouped by season of first dose. Conclusions: Despite previous reports of improved survival with earlier ICI dosing, we did not identify any association of time of pembrolizumab administration with clinical outcomes. Our analysis is limited by small sample size and patient heterogeneity which may hinder identification of smaller associations. It is also unclear whether lower response rates in a pan-cancer population (relative to prior reports in lung cancer and melanoma) might impact correlation analysis. Further studies will be necessary to interrogate this phenomenon and ensure that ICI are optimally applied.

2513 Background: BCA101 is a first-in-class bifunctional fusion protein consisting of an anti-EGFR monoclonal antibody (mAb) and TGFβ receptor 2 extracellular domain (TGFβRII-ECD). Herein, we report the safety, pharmacokinetic (PK), pharmacodynamic (PD), and preliminary efficacy data of BCA101 as monotherapy and in combination with pembrolizumab among patients (pts) with advanced solid tumors refractory to standard therapies. Methods: Pts received BCA101 as a single agent (SA) or in combination with pembrolizumab at escalating doses in a parallel 3+3 design starting at 64 mg intravenously (IV) weekly (qw); and at 240 mg IV qw with pembrolizumab 200 mg IV q3w. Primary endpoint: safety and tolerability (CTCAE v5.0); dose limiting toxicity (DLT) period: 21 days. Secondary endpoints: overall response rate (ORR), PK/PD profile, progression-free survival (PFS), and changes in plasma and intra-tumoral TGFβ signaling assessed by SMAD2 phosphorylation. Results: As of 08-Feb-2022, 60 pts have received BCA101 (part A). Forty-five pts (colorectal, n=14; pancreatic, n=7; head and neck squamous cell carcinoma [HNSCC], n=6) received SA BCA101 at doses up to 1500 mg IV weekly. Fifteen subjects (SCC of the anal canal [SCAC], n=8; HNSCC, n=7) received BCA101 doses ranging from 240 to 1500 mg IV qw in combination with pembrolizumab. Maximum tolerated dose has not been reached. Common adverse events (AEs) attributed to BCA101 include rash (70%), fatigue (23%), pruritis and epistaxis (17% each); all grade (G)2 or less. One DLT was observed at the 1250 mg SA dose (G3 anemia, hematuria). No drug-related G4 AEs or deaths were observed. At data cutoff, best response in the SA arm was stable disease (SD) in 15/39 (39%) evaluable pts. In combination, partial response (PR) was observed in 3/11 (27%) evaluable pts (2 in SCAC, 1 in HNSCC) and a disease control rate (DCR) of 9/11 (82%). Two of 3 responders have been on study >4 months; including 1 confirmed PR in a HNSCC pt refractory to anti-PD-1 therapy and cetuximab. Saturation of the EGFR target was observed at BCA101 doses ≥750 mg. Dose proportional increase in Cmax and AUC were observed with doses of BCA101 750-1500 mg. Prolonged neutralization of plasma TGFβ1 was achieved at all doses ≥500 mg. Among paired tumor biopsies (n=23), pSMAD2 reduction up to 62% was observed at doses ≥500 mg. Conclusions: BCA101 is well tolerated and clinically active as a SA and in combination with PD-1 blockade with a predictable PK/PD profile. A recommended dose of 1500 mg both as SA and in combination has advanced to the part B expansion phase for pts with HNSCC, SCAC, and cutaneous SCC. Clinical trial information: NCT04429542.

1052 Background: Dysregulating mutations in PIK3CA, encoding the PI3K p110α subunit, occur in ̃40% of HR+/HER2– BCs. Inavolisib is a PI3Kα-specific inhibitor that also promotes degradation of mutant p110α. It has demonstrated encouraging preliminary antitumor activity in pts with PIK3CA-mutated HR+ BC as a monotherapy, and in combo with other anticancer agents. Methods: We included pts from NCT03006172 on treatment ≥1 year with inavolisib alone (Arm A), or in combo with palbo + letrozole (letro) (B), letro (C), fulvestrant (fulv) (D), or palbo + fulv (E; + metformin in Arm F for pts with body mass index ≥30 and/or HbA1c ≥5.7%). Inavolisib was administered orally daily (PO QD) at 3, 6, 9, or 12 mg (3+3 dose-escalation design); letro at 2.5 mg PO QD; palbo at 125 mg PO QD for 21/28 days; and fulv at 500 mg intramuscularly every 4 weeks, in 28-day cycles until intolerable toxicity/disease progression. Safety was assessed by NCI-CTCAE v4. Results: 57 female pts were included (cutoff 07/26/21; N = 1, 18, 6, 12, 15, 5 in Arms A–F); median age: 57 years (range 33–80); median lines of prior therapy: 2 (1–7). All but 2 pts, both in Arm B (3 mg), were assigned the 9 mg inavolisib recommended phase 3 dose. Overall median treatment duration: 19 months (range 12–45); median inavolisib cumulative dose intensity, 95%. The most frequent treatment-related adverse events (AEs; in ≥20 pts/35%) were hyperglycemia (68%), stomatitis (68%; grouped terms), neutropenia (58%), diarrhea (51%), nausea (39%), alopecia (35%), and rash (35%; grouped terms). The most frequent treatment-related Grade (G) 3–4 AEs (≥2 pts/4%) were neutropenia (47%), hyperglycemia (16%), leukopenia (9%), thrombocytopenia (9%), lymphopenia (7%), weight decreased, and hypokalemia (4% each). G3–4 neutropenia, leukopenia, thrombocytopenia, and lymphopenia were all reported in palbo arms. One G5 AE of pleural effusion was reported (disease progression-related). 39 pts (68%) had ≥1 AE resulting in study treatment modification (drug interruption/dose reduction/treatment withdrawal); 11 (19%) had an inavolisib dose reduction and 2 (4%) discontinued treatment due to an AE (1 related G2 diarrhea, 1 unrelated G3 cerebrovascular disorder). AEs typically occurred during the first 6 months and tended to be less frequent in later cycles. No new safety signals were observed with long-term inavolisib use. Conclusions: These data indicate acceptable long-term tolerability. The safety profile of pts on study treatment with inavolisib alone or in combo with endocrine-based anticancer therapies for ≥1 year was similar to that reported for the overall study population. Updated data will be presented. A phase 3 study of inavolisib + palbo + fulv is enrolling (NCT04191499; INAVO120). Clinical trial information: NCT03006172.

2546 Background: ctDNA kinetics with tumor-informed assays can predict treatment outcome in patients (pts) treated with anti-PD1 IO ( Bratman et al, Nature Cancer 2020). We evaluated whether early ctDNA kinetics with a tumor-naïve assay were associated with clinical outcomes in advanced solid tumor patients treated on early phase IO trials. Methods: Advanced solid tumor pts treated with investigational IO agents at the Princess Margaret Phase I program were enrolled. Baseline (B) and pre-cycle 2 (C2) (3-4 weeks after first dose) plasma samples were prospectively collected via an institutional liquid biopsy program (LIBERATE, NCT03702309). ctDNA was assessed using the tumor-naïve 425-gene Geneseeq Prime panel in a clinical laboratory. Mutations in each gene detected in ctDNA were measured as Variant Allele Fraction (VAF). Mean VAF from all mutations was calculated. Radiological response was measured per RECIST criteria and correlated using ROC curves. Hyperprogression (HPD) was defined using VHIO criteria ( Matos et al, CCR 2020). Survival outcomes were estimated using the Kaplan Meier method. Results: From 12/2017 to 3/2020, 162 plasma samples from 81 pts with 25 different tumor types were collected. Pts were treated within 25 different IO phase I/II trials, 72% of which involved a PD-1/PD-L1 inhibitor. Median age was 58y (range 21 – 79), 54% female, 76% ECOG1. Sarcoma and colorectal (11%, each) followed by breast (8%) and melanoma (7%) were the most frequent tumors. Median follow up was 10.3 months (m) (1.8-46.9). CR 4% (n = 3), PR 6% (n = 5), HPD 11% (n = 9). Clinical benefit (CB) rate (CR+PR+SD > 6 months) was 20% (n = 16). ctDNA was detected in 122/162 samples (75.3%) (60 at B, 62 at C2). The most frequent mutations were TP53 (32%), PI3KCA (12%), PKHD1 (11%), and KRAS (9%). Mean VAF at B below median was not associated with OS (HR = 0.68 95%CI 0.4-1.16; p = 0.16) or PFS (HR = 0.93 95%CI 0.56-1.54; p = 0.77). Mean VAF change (difference between mean VAF at B and at C2) was associated with response (AUC = 0.99) and CB (AUC = 0.86). A decrease in mean VAF from B to C2 was seen in 24 pts (37.5%) and was associated with longer PFS (median PFS 2.7 vs 1.8 m; HR: 0.43, 95%CI 0.24-0.77; p < 0.01) and OS (median OS 10.8 vs 9.1 m; HR: 0.54; 95%CI 0.3-0.96; p = 0.03) compared to an increase in mean VAF. These differences were more marked if there was > 50% decrease in mean VAF from B to C2 (n = 11, 17%) compared to decrease < 50% or increase: median PFS 3.6 vs 1.8 m (HR: 0.29, 95%CI 0.13-0.62; p < 0.01) and median OS not reached vs 9.6 m (HR: 0.23, 95%CI 0.09-0.6; p < 0.01). No differences in mean VAF change were seen between HPD and PD pts. Conclusions: In a pan-cancer solid tumor early phase trial IO cohort, a decrease in ctDNA within 4 weeks of treatment was associated with increased CB, OS and PFS. HPD pts did not show greater increases in ctDNA. Tumor-naïve ctDNA assays may be useful to identify early treatment benefit in phase I/II trials with IO.

e24084 Background: To direct limited specialized palliative care resources to patients in greatest need, we developed STEP (Symptom screening with Targeted Early Palliative care). STEP entails symptom screening (ESAS-r) at each oncology clinic visit and triggered alerts (for moderate-high physical and psychological symptoms) to a nurse who calls the patient to offer a palliative care clinic (PCC) visit. We conducted a phase III RCT to assess the impact of STEP versus usual care on quality of life and other patient-reported outcomes (PROs). Methods: Adults with advanced cancer were recruited from medical oncology clinics at the Princess Margaret Cancer Centre, Toronto, Canada. Consenting patients with oncologist-assessed ECOG 0-2 and estimated survival of 6-36 months were enrolled and block randomized (stratified by tumour site and symptom severity) to STEP or usual care. Participants completed measures of quality of life (FACT-G7), depression (PHQ-9), symptom control (ESASr-CS), and satisfaction with care (FAMCARE-P16) at baseline, 2, 4 and 6 months. The primary outcome was FACT-G7 at 6 months, with a planned sample size of 261/arm. Results: From 8/2019 to 3/2020, 69 patients were enrolled: 33 randomized to STEP and 36 to usual care. The trial was then halted permanently due to the COVID-19 pandemic, owing to substantial changes to elements of STEP (shift to virtual symptom screening and palliative care) and usual care (shift to virtual oncology care). Median age was 64 years (range 25-87) and 62% (43/69) were women; study arms were balanced at baseline except gender, with more women randomized to STEP. Within the STEP arm, 20 (61%) participants triggered a nurse’s call to offer a PCC visit, of whom 13 attended the clinic at least once. All outcomes tended to be better in the STEP arm compared to usual care, particularly depression and satisfaction with care at 6 months; however, results were not statistically significant (Table). Conclusions: STEP holds promise for improving quality of life and other PROs in patients with advanced cancer and effectively directing early palliative care towards those who need it most. In response to the pandemic, an online version of STEP has been developed and a further trial is in progress. Clinical trial information: NCT03987906. [Table: see text]

3119 Background: There is limited information about the clinical utility of targeted NGS panel testing to inform decision-making for patients with advanced solid tumors. The Ontario-wide Cancer Targeted Nucleic Acid Evaluation (OCTANE) is an ongoing prospective study that enrolled over 4,500 solid tumor patients for NGS panel testing. We performed a retrospective survey of 21 medical oncologists enrolling OCTANE patients at a single academic institution to evaluate the impact of NGS testing on treatment decisions. Methods: Patients and treating oncologists were identified at the Princess Margaret Cancer Centre between 2016-2021. Tumor-only sequencing was performed using a custom hybridization capture panel of 555 cancer genes (Hi5) or a commercial 161-gene amplicon DNA/RNA panel (Oncomine Comprehensive v3). Oncologists were asked to review testing results for individual patients and complete a survey indicating whether NGS testing impacted treatment decisions. Mutations were defined as actionable based on clinical judgment and compared to classifications provided by OncoKB, an FDA-recognized precision medicine knowledgebase. The primary outcome of this study was rate of treatment change based on mutation results. Patient, test, and physician factors were evaluated for association with treatment changes using univariate analyses and a mixed effects model. Results: Two cohorts were surveyed, the first between 2017-2019 and the second in 2021. Of the 582 surveys sent, 394 (67.7%) were completed. Each physician completed a median of 19 surveys (range, 9-48). We found that 188 (47.7%) patients had a mutation classified as actionable by the oncologist, of whom 134 (71.3%) had ≥1 OncoKB-defined actionable mutation(s). 62/394 (15.7%) patients were matched to a treatment, of whom 37 were enrolled in a clinical trial, 13 received an approved drug, 4 were prescribed off-label therapy and 8 avoided ineffective treatment. 127/188 (67.5%) patients with actionable mutations did not receive treatment due to lack of available therapy, stability on current regimen, clinical deterioration or patient decision. Rate of treatment change was highest for bowel (15/37, 40.5%), breast (14/52, 27.5%), biliary tract (6/22, 27.3%) and lung (4/17, 23.5%) cancers. Treatment decisions were not associated with patient age, gender, physician clinical experience, physician gender, testing experience, OncoKB mutation level or time from biopsy to sequencing. There was no difference in overall survival between patients with matched vs. no matched treatment ( p = 0.55, median survival not reached). Conclusions: OCTANE testing led to a change in drug treatment in 15.7% of patients, supporting the clinical utility of NGS panel testing for patients with advanced solid tumors. Patient, test, and physician characteristics were not significantly associated with treatment change.

2510 Background: VIGex is a 12- gene GES classifier initially developed on the Nanostring platform and validated for RNA-seq. VIGex classifies samples into Hot, intermediate-Cold (I-Cold) and Cold subgroups. The Hot subgroup as defined by VIGex has been associated with better (PFS) in patients (pts) treated on phase 1 ICT trials at Vall D’Hebron Hospital (VH) (ESMO2020). We investigated the performance of VIGex in pts treated with Pembrolizumab (P) in the INSPIRE clinical trial (NCT02644369) at Princess Margaret Cancer Centre (PM) and compared VIGex with other predictive ICT biomarkers. Methods: Pts with advanced solid tumors were treated with P 200 mg IV Q3wks. RNA-seq from baseline biopsies was performed using the Illumina NextSeq550 platform. Tumor RNA-seq data were transferred from PM to VH and classified by the VIGex algorithm blinded to clinical data. Bespoke circulating tumor DNA (ctDNA) was assayed at baseline (B) and start of cycle 3 (C3) using a pt-specific amplicon-based NGS assay (Signatera). Tumor mutational burden (TMB) was defined as the number of non-synonymous mutations per megabase and PD-L1 was assessed by immunohistochemistry (22C3). Hot subgroup (HOT) was compared to I-Cold + Cold (COLD). We defined 4 groups based on the combination of VIGex subgroups and the change in ctDNA at cycle 3 from baseline (ΔctDNA). Survival times were calculated with the Kaplan–Meier method and Cox proportional-hazard models were constructed. Results: Out of 76 pts, median age was 55y (range 21-81y), M:F 31:45, all ECOG 0-1, 16 High-grade serous ovarian, 12 triple negative breast, 12 head and neck, 10 melanoma and 26 other. Median no. of P cycles was 3 (range 1–35); follow up was 14m (range 1-67); Median PFS 10.9m and median overall survival (OS) 14m. Overall response rate (RECIST 1.1) was 24% in HOT and 10% in COLD (p = 0.22 two-sided Fisher's exact test). The HOT subgroup was significantly associated with higher OS and PFS when included in a multivariate model adjusted by tumor histology, TMB and PD-L1 (HR 0.43; 95%CI 0.23-0.81; p = 0.009) and (HR: 0.48; 95%CI 0.25-0.95; p = 0.036) respectively. A total of 57 pts had both VIGex and ΔctDNA data. The addition of ΔctDNA further improved the predictive performance of VIGex for OS (Table). Conclusions: VIGex maintained its predictive power for ICT outcomes when applied to an independent external dataset using RNA-seq. The predictive information provided by VIGex was independent of PD-L1 and TMB. Our data indicates that the addition of ΔctDNA to baseline VIGex may refine prediction for ICT outcomes. [Table: see text]

5514 Background: Mesothelin and its binder, antigen-CA125, are highly expressed in high grade serous and endometrioid ovarian cancers (HGOC) and, can inhibit paclitaxel-induced cell death. Anetumab ravtansine (AR) is a fully-human antibody directed at the mesothelin antigen, conjugated to a tubulin polymerization inhibitor. We assessed the safety and activity of the combination AR/bevacizumab (ARB) versus weekly paclitaxel/bevacizumab (PB) in patients (pts) with platinum resistant HGOC. Methods: An initial run-in phase I assessed the safety of ARB. After determination of the recommended phase 2 dose (RP2D), a multicenter 1:1 randomized phase 2 trial was designed to evaluate the progression free survival (PFS) in pts with platinum resistant/refractory HGOC. Pts were stratified by platinum resistant or refractory and prior bevacizumab (bev). Eligibility required measurable disease and mesothelin tested positive centrally on archival tissue by IHC. No limitation on the number of prior lines of therapy. A futility analysis was planned at 35 PFS events. The control arm was weekly intravenous paclitaxel 80mg/m2 with bev 10mg/kg every 2 weeks. A CT was performed every 8 weeks for RECIST1.1 assessment. The toxicities were reported according to CTCAE version 5.1. NCT03587311. Results: 7 pts were enrolled in the run-in phase 1 and the RP2D determined as bev (10mg/kg) biweekly with AR (2.2mg/kg) weekly on a 28 day-cycle. In the phase 2, 57 pts were enrolled, 28 pts in the ARB and 29 pts in the control group. The positivity rate for mesothelin screening was 88%. Pts were heavily pre-treated, median prior lines of 3 (range (1-9) with 24 pts received prior bev (42%) and 13 pts were platinum refractory (7 in ARB and 6 in PB). At the time of 35 PFS events, one CR and 4 PR were observed (ORR = 18%) in the ARB arm, versus no CR and 16 PR in the weekly PB (ORR = 55%). The estimated median PFS was 5.3 (95% CI: 3.7-7.4) months for ARB and 9.6 (95% CI: 7.4-17.4) months for PB (HR = 1.7(95% CI: 0.9-3.4)). The median number of cycles were 4 (1, 29) and 8 (1, 19) respectively. The most common treatment-related AEs in the ARB arm were mostly grade 1/2 increase AST (71%) and ALT (64%), thrombocytopenia (61%), fatigue (57%), and peripheral neuropathy (46%). In the PB arm, the most common treatment-related AE were anemia (66%), neutropenia (59%), epistaxis (48%), fatigue (45%) and peripheral neuropathy (45%). Conclusions: At the time of futility analysis, weekly PB had better outcome than weekly ARB leading to the study termination. Molecular and blood analyses are on-going to assess potential biomarkers of response. This study highlights the importance of randomization in assessment of novel therapies and potential for re-challenge with bevacizumab. These data show that weekly PB is an effective regimen and can be considered as the control arm in platinum resistant HGOC. Clinical trial information: NCT03587311.

The American Association for Cancer Research (AACR) Project Genomics Evidence Neoplasia Information Exchange (GENIE) is an international data-sharing consortium focused on enabling advances in precision oncology through the gathering and sharing of tumor genetic sequencing data linked with clinical data. The project’s history, operational structure, lessons learned, and institutional perspectives on participation in the data-sharing consortium are reviewed. Individuals involved with the inception and execution of AACR Project GENIE from each member institution described their experiences and lessons learned. The consortium was conceived in January 2014 and publicly released its first data set in January 2017, which consisted of 18,804 samples from 18,324 patients contributed by the eight founding institutions. Commitment and contributions from many individuals at AACR and the member institutions were crucial to the consortium’s success. These individuals filled leadership, project management, informatics, data curation, contracts, ethics, and security roles. Many lessons were learned during the first 3 years of the consortium, including on how to gather, harmonize, and share data; how to make decisions and foster collaboration; and how to set the stage for continued participation and expansion of the consortium. We hope that the lessons shared here will assist new GENIE members as well as others who embark on the journey of forming a genomic data–sharing consortium.

Purpose After chemotherapy, approximately 50% of patients with metastatic testicular germ cell tumors (GCTs) who undergo retroperitoneal lymph node dissections (RPNLDs) for residual masses have fibrosis. Radiomics uses image processing techniques to extract quantitative textures/features from regions of interest (ROIs) to train a classifier that predicts outcomes. We hypothesized that radiomics would identify patients with a high likelihood of fibrosis who may avoid RPLND. Patients and Methods Patients with GCT who had an RPLND for nodal masses > 1 cm after first-line platinum chemotherapy were included. Preoperative contrast-enhanced axial computed tomography images of retroperitoneal ROIs were manually contoured. Radiomics features (n = 153) were used to train a radial basis function support vector machine classifier to discriminate between viable GCT/mature teratoma versus fibrosis. A nested 10-fold cross-validation protocol was used to determine classifier accuracy. Clinical variables/restricted size criteria were used to optimize the classifier. Results Seventy-seven patients with 102 ROIs were analyzed (GCT, 21; teratoma, 41; fibrosis, 40). The discriminative accuracy of radiomics to identify GCT/teratoma versus fibrosis was 72 ± 2.2% (area under the curve [AUC], 0.74 ± 0.028); sensitivity was 56.2 ± 15.0%, and specificity was 81.9 ± 9.0% ( P = .001). No major predictive differences were identified when data were restricted by varying maximal axial diameters (AUC range, 0.58 ± 0.05 to 0.74 ± 0.03). The prediction algorithm using clinical variables alone identified an AUC of 0.76. When these variables were added to the radiomics signature, the best performing classifier was identified when axial masses were limited to diameter < 2 cm (accuracy, 88.2 ± 4.4; AUC, 0.80 ± 0.05; P = .02). Conclusion A predictive radiomics algorithm had a discriminative accuracy of 72% that improved to 88% when combined with clinical predictors. Additional independent validation is required to assess whether radiomics allows patients with a high predicted likelihood of fibrosis to avoid RPLND.

Purpose Fine-needle biopsy (FNB) and liquid biopsy are minimally invasive methods of tumor sampling that provide feasible means to assess tumor genotypes in real time. However, more data are needed to establish the strength of these methods by benchmarking against the current gold standard methods, core-needle biopsy (CNB) or surgical excision of the tumor. Patients and Methods Eligible patients with advanced solid tumors were prospectively recruited. We performed mutation profiling using matched tumor DNA obtained by CNB, FNB and liquid biopsy, and matrix-assisted laser desorption/ionization time-of-flight custom mass-spectrometry or targeted next-generation DNA sequencing. The actionability of detected mutations was determined using the OncoKB Web tool. Agreement between mutations detected in CNBs, FNBs, and circulating tumor DNA (ctDNA) was examined. Results Forty-one patients underwent tumor biopsy. Thirty CNBs (73%) and 34 FNBs (83%) had sufficient tumor and DNA for mutation profiling. Median DNA yield from CNB and FNB were 775 ng (interquartile range, 240 to 347 4ng) and 649 ng (interquartile range, 180 to1350 ng), respectively. Of 29 CNB/FNB pairs available for comparison, actionable mutation results were concordant in 28 (96%). Six of nine actionable mutations (67%) that were found by CNB, FNB, or both were detectable in ctDNA. Two additional actionable mutations were found exclusively in ctDNA. Conclusion Optimally processed FNB and liquid biopsy can be used routinely for tumor mutation profiling to identify actionable mutations.

303 Background: ICIs are a common therapeutic option for many solid tumors. While prior studies have shown that ATB exposure may negatively impact ICI outcomes through gut microbiome changes, many were small studies with heterogeneity in ATB classes and exposure windows. Here, we performed a population level retrospective cohort study to evaluate the impact of ATB exposure prior to ICI on OS. Methods: We used administrative data to identify a cohort of cancer patients > 65 years of age receiving ICIs from June 2012 to October 2018 in Ontario, Canada and deterministically linked with databases to obtain socio-demographic and clinical co-variates and ATB prescription claims. Multivariable cox-proportional hazard models evaluated the impact of ATB exposure both within 1 year and 60 days prior to starting ICI on OS, adjusted for age, gender, body mass index, comorbidities, autoimmune history, hospitalization in the past year and treatment facility level at start of ICI therapy. Results: Among 2737 patients, median age 73; 43% received Nivolumab, 41% Pembrolizumab and 13% Ipilimumab; 53% were lung cancer, 34% melanoma. Median ATB treatment duration for patients receiving ATB within 1 year (59%) and 60 days (19%) prior to ICI were 14 days (SD = 32) and 9 days (SD = 13) respectively. Median OS estimate was 306 days. Any ATB exposure within 1 year prior to ICI was associated with worse OS (aHR = 1.12 95% CI [1.12-1.23] p = 0.03). A non significant dose effect was seen based on weeks of ATB exposure 1 year prior to ICI (aHR = 1.01 per week [1.00-1.02] p = 0.10). ATB class analysis identified fluoroquinolone exposure within 1 year (aHR = 1.26 [1.13-1.40] p < 0.001) and 60 days before ICI (aHR = 1.20 [0.99-1.45] p = 0.06) were associated with worse OS; with a dose effect based on total weeks of exposure over 1 year (aHR = 1.07 per week [1.03-1.11] p < 0.001) and 60 days (aHR = 1.12 per week [1.03-1.23] p = 0.01). Subgroup analysis showed similar results for patients receiving anti-PD1 ICIs, where patients exposed to fluoroquinolone both 1 year (aHR = 1.28 [1.15-1.44] p < 0.001) and 60 days (aHR = 1.19 [0.98-1.44] p = 0.08) before ICIs had poorer OS with dose effects observed based on weeks of fluoroquinolone exposure. Similarly, subgroup analyses based on disease site identified that lung cancer patients exposed to fluoroquinolones 1 year before starting ICIs (aHR = 1.22 [1.06-1.39] p = 0.005) and melanoma patients exposed to fluoroquinolones 60 days before starting ICIs (aHR = 1.66 [1.12-2.47] p = 0.01) had poorer OS. Conclusions: Exposure to ATBs and specifically fluoroquinolones prior to ICI therapy is associated with worse OS. Interventions aimed at altering the gut microbiome may be required to help improve outcomes for patients on ICIs with prior ATB exposure.

304 Background: ICIs are becoming a common therapeutic option for many solid tumors. Prior studies have shown that ATB exposure can negatively impact ICI outcomes through gut microbiome changes leading to poorer overall survival; however, less is known about the potential impact of ATB exposure on toxicities from ICI. We undertook a population-based retrospective cohort study in patients receiving ICIs to evaluate the impact of ATB exposure on early acute care use, defined as emergency department visit or hospitalization, within 30 days of initiation of ICI therapy. Methods: Administrative data was utilized to identify a cohort of cancer patients > 65 years of age receiving ICIs from June 2012 to October 2018 in Ontario, Canada. We linked databases deterministically to obtain socio-demographic and clinical co-variates, ATB prescription claims and acute care utilization. Patients were censored if they died within 30 days of initiating ICI therapy. The impact of ATB exposure within 60 days prior to starting ICI on early acute care use was evaluated using multi-variable logistic regression models, adjusted for age, gender, rurality, recent hospitalization within 60 days prior to starting ICI and comorbidity score. Results: Among 2737 patients (median age 73 years), 43% received Nivolumab, 41% Pembrolizumab and 13% Ipilimumab, most commonly for lung cancer (53%) or melanoma (34%). Of these patients, 19% had ATB within 60 days prior to ICI with a median ATB treatment duration of 9 days (SD = 13). 647 (25%) patients had an acute care episode within 30 days of starting ICIs; 182 (7%) patients passed away within 30 days without acute care use and were censored from further analyses. Any ATB exposure within 60 days prior to ICI was associated with greater likelihood of acute care use (aOR = 1.34 95% CI [1.07-1.67] p = 0.01). A dose effect was seen based on weeks of ATB exposure within 60 days prior to ICI (aOR = 1.12 per week [1.04-1.21] p = 0.004) and early acute care use. ATB class analysis identified that exposure to penicillins (aOR = 1.54 [1.11-2.15] p = 0.01) and fluoroquinolones (aOR = 1.55 [1.11-2.17] p = 0.01) within 60 days of starting ICIs were associated with a greater likelihood of acute care use, while there was no significant association between cephalosporin exposure and early acute care use (p > 0.05). Conclusions: Exposure to ATBs, specifically fluoroquinolones and penicillins, prior to ICI therapy is associated with greater likelihood of hospitalization or emergency room visits within 30 days after initiation of ICIs, even after adjustment for relevant co-variates including age, comorbidity score and recent hospitalization prior to ICI initiation. Further studies are required to better understand the mechanisms of recent ATB exposure on early acute care use among patients receiving ICIs.

Purpose AKT1 E17K mutations are oncogenic and occur in many cancers at a low prevalence. We performed a multihistology basket study of AZD5363, an ATP-competitive pan-AKT kinase inhibitor, to determine the preliminary activity of AKT inhibition in AKT-mutant cancers. Patients and Methods Fifty-eight patients with advanced solid tumors were treated. The primary end point was safety; secondary end points were progression-free survival (PFS) and response according to Response Evaluation Criteria in Solid Tumors (RECIST). Tumor biopsies and plasma cell-free DNA (cfDNA) were collected in the majority of patients to identify predictive biomarkers of response. Results In patients with AKT1 E17K–mutant tumors (n = 52) and a median of five lines of prior therapy, the median PFS was 5.5 months (95% CI, 2.9 to 6.9 months), 6.6 months (95% CI, 1.5 to 8.3 months), and 4.2 months (95% CI, 2.1 to 12.8 months) in patients with estrogen receptor–positive breast, gynecologic, and other solid tumors, respectively. In an exploratory biomarker analysis, imbalance of the AKT1 E17K–mutant allele, most frequently caused by copy-neutral loss-of-heterozygosity targeting the wild-type allele, was associated with longer PFS (hazard ratio [HR], 0.41; P = .04), as was the presence of coincident PI3K pathway hotspot mutations (HR, 0.21; P = .045). Persistent declines in AKT1 E17K in cfDNA were associated with improved PFS (HR, 0.18; P = .004) and response ( P = .025). Responses were not restricted to patients with detectable AKT1 E17K in pretreatment cfDNA. The most common grade ≥ 3 adverse events were hyperglycemia (24%), diarrhea (17%), and rash (15.5%). Conclusion This study provides the first clinical data that AKT1 E17K is a therapeutic target in human cancer. The genomic context of the AKT1 E17K mutation further conditioned response to AZD5363.

11522 Background: The use of immune checkpoint blockade (ICB) in non-inflamed (cold) tumors is associated with limited clinical efficacy. Combination of ICB with certain molecularly targeted agents (MTA) is hypothesized to increase tumor immunogenicity by recruiting tumor infiltrating lymphocytes in cold tumors, such as LMS. Here, we present the results of LMS cohort treated on the DAPPER study (NCT03851614). Methods: LMS pts with ECOG 0-1 were randomized to either D+O (arm A), or D+C (arm B). In a 28-day cycle, D 1500mg i.v. q4w with either O 300mg bid po qd or C 20mg po qd 5d/week were administered. Overall response rates (ORR) were determined using RECISTv1.1. Evaluation of tumor kinetics (TK) was performed by calculating tumor growth rate (TGR) of target lesions on CT images at baseline and on-treatment, adjusted to account for the time difference between scans. TGR is expressed as % tumor growth/week (Ferte C et al. CCR, 2014). Additionally, paired FFPE samples (from baseline and on-treatment biopsies) were assessed using multispectral fluorescent immunohistochemistry (IHC) panel: CD3, CD8, CD20, CD68, FOXP3 and cytokeratin. Tumor areas were identified by a pathologist and immune cells were quantified using InForm image analysis software. Results: 25 metastatic LMS pts were randomized to arm A (n = 11) or B (n = 14) over 21 months. Median age was 53 years, 96% were females and 60% of pts had ≥3 lines of therapy. In 23 evaluable pts, no responses were seen, 7 pts had stable disease (SD) while 16 has progressive disease (PD). TK analysis was evaluable for 18 pts (arm A = 8, B = 10). 5/8 pts (62.5%) in arm A and 6/10 pts (60%) in arm B showed decreased TK (defined as TGRbaseline > TGRon-treatment). In 4/5 (80%) pts who had deceleration of TK in arm A, SD was maintained for ≥6 months. The reduction in TGR on treatment, compared to baseline was significant in arm A but not in arm B (measured as median % tumor growth/week of 0.5 vs 5.1, 95% CI 0.2-4.3, p = 0.035 in arm A; and 1.3 vs 2.9, 95% CI 0.2-2.7, p = 0.088 in arm B). The median PFS of arm A and B were 9 (95% CI 3-12.8) and 4 (95% CI 2.2-4.6) months respectively. There were no statistically significant differences in tumor-infiltrating immune cells when comparing baseline and on-treatment biopsies from arm A or B. In arm A, one pt with SD > 6 months had a 2.5-fold increase in CD8 (CD3+CD8+) T cells and a 7.6-fold increase in macrophages (CD68+). Conclusions: D+O or D+C resulted in stable disease in 30% of pts, mostly on arm A (D+O). TK analysis may identify pts with prolonged SD on treatment. Although a cold-to-hot immunophenotype change was not generally seen, changes in tumor infiltrating immune cell subsets were observed in one patient with prolonged stable disease. These findings support further molecular and immunophenotype characterization in LMS patients treated with D+O or D+C. Clinical trial information: NCT03851614.

12051 Background: There are limited data describing the patients’ experience of symptomatic adverse events (syAEs) on early phase clinical trials. This information is critical to understand the tolerability of experimental agents. The patient reported outcome version of the CTCAE (PRO-CTCAE) evaluates syAE components such as severity and interference in daily life. The aim of this study was to correlate clinician reported early, high grade (grade 3-4) AEs with patients’ reported experience of these toxicities. Methods: Advanced solid tumor patients (pts) enrolled on early phase clinical trials at Princess Margaret Cancer Centre were surveyed electronically using the full library of 78 items for PRO-CTCAE v1.0, which was administered at baseline (prior to therapy), mid-cycle 1, and mid-cycle 2. AEs on study were recorded by physicians using the CTCAE v4.0. Worst responses for severity items are ‘severe’ and ‘very severe’ and for interference items are ‘quite a bit’ and ‘very much’. A logistic regression model was used to assess the association between CTCAE grade and PRO-CTCAE severity and interference. Results: A total of 292 pts were approached in phase 1 clinics from May 2017 to January 2019, and 219 pts were included in the analysis: median age 60 years (range 18-82), 111 (51%) were male; all were ECOG ≤1. A total of 140 pts (64%) received combination therapy (immunotherapy and targeted therapy), and 73 pts (33%) had received ≥3 previous lines of treatment. In terms of patient reported syAEs, a total of 114 pts (52%) reported a symptomatic AE as either severe or very severe at any timepoint and 79 pts (36%) reported a syAE with an interference that was either ‘quite a bit’ or ‘very much’. With regards to clinician reported AEs, a total of 82 pts (37%) had a clinician reported grade 3 or 4 syAE, and of these 34 pts (41%) reported these as either severe or very severe; and 26 pts (32%) found these AEs interfered with daily life either ‘quite a bit’ or ‘very much’. Additionally 137 pts (66%) had a clinician reported grade 1 or 2 syAE, and of these 39 pts (28%) reported these as either severe or very severe; and 19 (14%) found these AEs interfered with daily life either ‘quite a bit’ or ‘very much’. Higher grade clinician reported syAEs (CTCAE grade 3-4 vs 1-2) was associated with higher patient reported severity levels (odds ratio, OR = 1.78, 95% CI 1-3.16, p = 0.049), and was associated with higher patient reported interference levels (OR = 2.88, 95% CI 1.47-5.64, p = 0.002). Conclusions: A majority of patients had a very negative experience of syAEs on a phase I trial. Higher grades of clinician reported AEs correlated with greater severity and interference with daily living. Future phase I studies could incorporate the PRO-CTCAE and other PRO tools which could inform the tolerability of experimental regimens and enhance the description of symptomatic AEs.

TPS2679 Background: Immune checkpoint inhibitors (ICI) have demonstrated efficacy in a wide variety of cancers. Nevertheless, only a small proportion of patients derive a durable benefit. Mechanisms underlying primary and acquired resistance are still incompletely understood. They comprise tumor-intrinsic factors such as genomic and transcriptomic changes; upregulation of immunosuppressive subsets; T cell exhaustion; and promotion of an immune-tolerant tumor microenvironment. The collection of tumor biopsy at disease progression (PD) is challenging both in clinical and research settings as this often occurs at the time of treatment discontinuation. However, the analysis of these samples can lead to novel strategies to prevent or reverse immune resistance. Thus, the current approach to begin a profiling study with patients at the time of PD on ICI enables access and interrogation of such samples. Methods: IRIS is a prospective, investigator-initiated trial at the Princess Margaret Cancer Centre that aims to extensively characterize the genomic, transcriptomic, epigenetic and immunophenotypic profiles of tumors with primary versus acquired resistance to ICI-based therapy. Primary resistance is defined as PD at the first on-treatment imaging, whereas acquired resistance is defined as PD occurring after an initial partial or complete response or following disease stability lasting ≥6 months. Additional objectives include the evaluation of radiomic parameters on standard radiological imaging, investigation of fecal microbiome, generation of patient-derived organoids and facilitation of data and sample sharing with the research community. The planned samples size is 100 patients. A one-time fresh tumor biopsy, blood and stool samples and archival tissue (when available) are collected at the time of PD on ICI (baseline) from all the participants. Longitudinal blood samples are obtained every 2-3 months (around the time of tumor imaging) until PD in patients receiving a subsequent treatment. Subjects who are not amenable for therapy undergo blood collections at the time of further PD. Molecular characterization of tumor samples includes: DNA/RNA sequencing, Assay of Transposase Accessible Chromatin (ATAC)-sequencing, Cellular Indexing of Transcriptomes and Epitopes (CITE)-sequencing, multiplexed immunohistochemistry and flow cytometry. Results of NGS performed on the first biopsy core are returned to patient and physician. Key eligibility criteria include diagnosis of solid tumor, progression to ICI as the most recent line of treatment and disease amenable to core needle biopsy. The IRIS trial, activated in October 2020, is currently open to enrollment. As of January 2021, 21 patients have been enrolled and a total of 92 tissue cores, 42 blood and 20 stool samples have been collected. Clinical trial information: NCT04243720.

3074 Background: Therapeutic targeting of EGFR has demonstrated efficacy in common advanced malignancies such as colorectal cancer (CRC), squamous cell of the head and neck (SCCHN), and non-small cell lung cancer (NSCLC). TGFβ has pro-tumorigenic effects on migration, invasion and tumor-specific immunosuppression, which may enhance the oncogenic effects of activated EGFR. Inhibition of EGFR signaling can lead to TGFβ upregulation as a resistance mechanism. BCA101 is a bifunctional recombinant fusion protein consisting of a chimeric anti-EGFR antibody and an extracellular domain (ECD) of human TGFβRII which demonstrated anti-tumor activity in several preclinical models. Methods: Patients with EGFR-driven advanced solid cancers refractory to standard therapies received BCA101 at escalating doses from 64 mg to 1000 mg intravenously (IV) weekly across 6 dose levels using a 3+3 design to determine dose limiting toxicities (DLT, established within 21 days of initial dosing), maximum tolerated dose (MTD) and/or recommended dose (RD). Secondary endpoints include detailed pharmacokinetic (PK), pharmacodynamic (PD) studies in serial tumor and/or blood samples and assessment of anti-tumor activity. Results: As of 2/11/2021, 21 patients received single agent BCA101 at 64 (n = 3), 240 (n = 7), 500 (n = 2), 750 (n = 3), 800 (n = 3) or 1000 (n = 3) mg IV weekly, including patients with CRC (n = 6), SCCHN (n = 5) uveal melanoma (n = 2), ovarian cancer (n = 2), glioblastoma multiforme (n = 2), conjunctival melanoma, chordoma, pancreatic cancer and anal squamous cell carcinoma (all n = 1). These patients had 1-7 prior lines of antineoplastic therapy (median 4), including 3 patients with prior EGFR inhibitor exposure. Adverse events (AE) related to BCA101 observed in > 1 patient included grade (G) 1-2 rash (n = 9), G 1-2 lipase elevation (n = 2). G3 vitreous hemorrhage at the 240 mg dose level has been the only DLT. The MTD has not been reached and the RD will be based on safety, exposure and pending PD data. Saturation of the clearance was observed at doses above 500mg. Dose proportional increase in Cmax and AUC were observed with doses of 750-1000mg. Best RECISTv1.1 response was stable disease (SD) in 3/10 evaluable patients, with 1 patient on drug ≥4 months. Conclusions: BCA101 is well tolerated at biologically active doses. BCA101 is now being tested in combination with the PD-1 antibody pembrolizumab in patients with SCCHN and anal carcinoma. Clinical trial information: NCT04429542.

2032 Background: 30-50% of patients with non-early NSCLC will eventually develop BM, with a median survival of less than one year from BM diagnosis. There are no widely accepted clinical risk models for development of BM in patients without them at baseline. We predicted the binary risk of BM using clinical and genetic factors from a large multi-institutional cohort. Methods: Stage II-IV NSCLC patients from the AACR Project GENIE Biopharma Consortium dataset were eligible. This consisted of 4 academic institutions who curated clinical data of patients who had somatic next-generation tumor sequencing (NGS) between 2015-2017. We excluded patients who had BM at baseline, died within 30 days of NSCLC diagnosis, or did not undergo brain imaging. Covariates included demographics, anticancer therapies (received up to 90 days prior to BM development and within 5 years from NSCLC diagnosis), and NGS data; radiotherapy (RT) data were not available. NGS features included mutations and copy number alterations. These features were restricted to those classified as oncogenic by OncoKB. Univariate feature selection with Fisher’s test (p

e15038 Background: Recurring oncogenic non-V600E BRAF mts have been identified in many cancer types. Preclinical data indicate that some BRAF non-V600E mts can be targeted with BRAF + MEK inhibitors. BEAVER is an investigator-initiated study designed to test the safety and efficacy of binimetinib and encorafenib (B+E) in patients (pts) with non-V600E BRAF mts. Methods: Key eligibility criteria are: pts with advanced solid tumors with BRAF non-V600E activating (class 1 and 2) or inhibitory (class 3) mts, and no prior BRAF/MEK inhibitors. Pts receive binimetinib (45mg PO BID) and encorafenib (450mg PO daily) on a 28-day cycle until intolerable toxicity or progression. The primary objective is OR rate (ORR) as per RECIST 1.1. In this Simon 2-stage trial, ≥1 of 7 pts need to have an objective response (OR) before commencing second stage of study (26 pts total). Secondary objectives include: safety, DCR, PFS, and OS. Exploratory objectives include: genomic profiling of tumors, evaluating circulating tumor DNA dynamics and development of patient derived xenograft (PDX) models. Results: From June 2019 to Feb 2021, 12 pts were screened and 9 pts enrolled; 9 are evaluable for safety and 8 for efficacy. Tumor types were melanoma and colon (n=2 each), as well as gallbladder, lung, breast and uterine cancers (n=1 each). Median age was 62 yrs (range 40-72). Median number of prior treatments was 2 (range 0-6). 1 pt had a class 1, 3 pts had class 2, and 5 pts had class 3, non-V600E BRAF mts. The median number of cycles was 2 (range: 1-7). Common treatment-related adverse events were mostly grade ≤ 2, and included: Blurred vision (78%), fatigue (67%), nausea (44%), vomiting (33%), and rash (33%). Dose reductions were required in 4/9 pts (44%) due to: blurred vision (22%), central serous retinopathy (11%), malaise (11%) and increased lipase (11%). Drug-related grade 3 AEs occurred in 2/9 pts and included: malaise (11%), confusion (11%), fatigue (11%) and increased lipase (11%). All eye toxicities were reversible with dose interruption. ORR was 12.5% (1/8) with one unconfirmed PR in a melanoma pt (BRAF G469S), treated for 6.5 months. One gallbladder cancer pt (BRAF D594N) had SD, and 6 pts had PD as best response. Genomic profiling was performed on archival tumors for 8 pts. Two PDX models were established. Responses to B+E in PDX models mirrored responses in 2 corresponding pts who had PD. Genomic and molecular profiling of pt tumors and corresponding PDXs identified multiple potential mechanisms of B+E resistance including: activation of EGFR and Akt pathways and inactivation of NF1 and Rb1. Conclusions: Preliminary data confirmed the safety of B+E and showed preliminary evidence of anti-tumor activity in advanced cancer pts with non-V600E BRAF mts. This study met the criterion for advancing to stage 2. Enrolment in the BEAVER trial and correlative biomarker analyses are ongoing. Clinical trial information: NCT03839342.

10535 Background: Homologous recombination is a major mechanism of defective DNA repair, but it remains uncertain whether homologous repair deficient (HRD) tumors have favorable prognosis or are more/less likely to respond to treatment than tumors lacking such mutations. Objective: To determine whether lung (NSCLC) and colorectal (CRC) HRD+ tumors have better survival or response to chemotherapy than HRD- tumors. Methods: Patients with de novo stage IV NSCLC or CRC who had next generation sequencing (NGS) between 2015-2018 from one of four cancer centers were identified. Records were curated using the PRISSMM framework to ascertain treatment, overall survival (OS) and progression free survival based on imaging (PFS-I) and oncologists’ notes (PFS-M). Each NSCLC or CRC tumor was categorized as HRD+ if NGS revealed an oncogenic/likely oncogenic mutation in: ATM, BAP1, BARD1, BLM, BRCA1, BRCA2, BRIP1, CHEK2, FAM175A, FANCA, FANCC, NBN, PALB2, RAD50, RAD51, RAD51C, RTEL1, or MRE11A based on the OncoKB database. The tumor was categorized as HRD- if no oncogenic mutation in any of these genes was evident and HRD indeterminate (HRD?) if no mutation was identified but the panel did not include all genes. OS, PFS-I and PFS-M from start of first line therapy were reported by HRD status. The percentage with a good response to first line therapy (≥2x the median) and exceptional response (≥3x the median) was estimated for each endpoint. Results: For NSCLC 4% were HRD+, 59% HRD- and 37% HRD?. For CRC there were 5% HRD+, 60% HRD- and 35% HRD?. There were no significant differences for any survival endpoint between patients who were HRD+ vs HRD- in univariable analyses. The proportion of good and exceptional responders to first line systemic chemotherapy also did not vary by HRD status, though patients with HRD+ CRC were potentially more likely to be exceptional responders. Similarly, no differences between HRD+ and HRD- tumors were apparent for the subgroup receiving platinum containing therapy. Conclusions: NSCLC and CRC patients with somatic mutations in HRD oncogenic genes did not differ from patients lacking such a mutation with respect to OS or PFS. CRC patients with HRD+ tumors may be more likely to be exceptional responders, but sample sizes are limited. By May, the analysis will include breast and pancreatic cancer cases.[Table: see text]

223 Background: Platinum-based chemotherapy agents cause significant hearing loss in 40-80% of treated cancer pts. Lack of follow-up serial testing has created gaps in knowledge regarding hearing loss onset, progression, and possible recovery between treatment cycles. This study aims to determine barriers and facilitators to implementation of a tablet-based point-of-care hearing test, as a serial screening tool to address these knowledge gaps. Methods: From Jul 2019 to Mar 2020, 53 pts receiving high dose platinum agents were recruited from three clinics (Thoracic, head and neck, and testicular cancer) at a comprehensive cancer centre, to undergo serial audiometry testing. Baseline hearing tests, mid cycles (3,6, and 9 weeks), and post treatment tests (3,6,9,12, 19 and 24 months) were completed during the pts’ clinic appointments. Clinical research coordinators (CRCs) collected feedback from physicians, nurses, and pts to identify barriers and facilitators of implementing serial point-of-care hearing tests in these clinics. An inductive and iterative approach was used to identify themes. Implementation was tailored and mapped to the CIHR Knowledge to Action Framework (KTA). Results: Barriers: Logistical barriers included: locating quiet and accessible rooms to administer the test; pts being distracted or interrupted while completing the test; presence of family members adding to noise levels; concerns over the serial testing during treatment; length of each test; and clinic staff burden. Facilitators: User-friendly self-administered tests; increasing healthcare staff education and pt management. Adapting to the local context: Logistical barriers were resolved by CRCs designating quiet spaces for the study to occur, and meeting pts upon arrival to utilize their wait time. A ‘hearing test in progress’ sign put on exam room doors prevented interruptions. CRCs utilized the test’s ‘assisted mode’ feature to keep pts attentive and/or accelerate the process. Low noise level was emphasized to obtain accurate test results. Pt engagement in their test results facilitated retention in the study. Test length may be shortened in the future by omitting low frequency testing. Conclusions: Participants and stakeholders expressed support for in-clinic hearing tests and identified personal and systemic barriers to implementation. These findings suggest that implementation should focus on addressing concerns related to accessible rooms, pt time investment and overall clinic flow.

3102 Background: Recent work from our laboratory demonstrated that T cell-derived acetylcholine induces vasodilation and increases T cell migration to infected tissues in response to viral infection (Cox et al. Science 2019). Choline acetyltransferase catalyzes the production of acetylcholine from choline and acetyl-CoA, however acetylcholine is challenging to quantify due to its extremely short half-life while choline is stable. This study is the first reported attempt to correlate serum choline levels in patients (pts) with advanced solid tumors receiving pembrolizumab with treatment outcomes. Methods: Blood samples were collected pre-treatment in 106 pts treated with pembrolizumab 200 mg IV Q3W in the investigator-initiated INSPIRE study (NCT02644369). Of these, 81 pts had on-treatment blood samples collected at week 7 (pre-cycle 3). Serum choline was analyzed with an HPLC-tandem mass spectrometry assay. PD-L1 staining was performed in baseline tumor tissues using 22C3 antibody and scored using modified proportion score. Tumor mutational burden (TMB) was calculated based on number of nonsynonymous mutations detected using whole exome sequencing. Multivariable Cox models were used to assess the impact of choline on PFS and OS, while adjusting for cohort, PD-L1 expression and TMB. Results: This pan-cancer group of 106 pts (median age 55, 62% females) comprised of 5 cohorts: squamous cell carcinoma of the head and neck = 19 pts, triple negative breast cancer = 22, high grade serous ovarian cancer = 21, melanoma = 19, mixed solid tumors = 32. With a median follow-up of 11 months, the median PFS = 1.9 months and median OS = 13.9 months for the entire cohort. In univariable analysis adjusted by cohort, baseline serum choline levels in 106 pts did not correlate with PFS or OS. However, an increase in serum choline level at week 7 compared to pre-treatment (D choline) in 81 pts was significantly associated with a better PFS (aHR 0.49, 95% CI 0.28-0.85, p = 0.01), and a trend towards a better OS (aHR 0.57, 95% CI 0.32-1.03, p = 0.064). In multivariable analysis, D choline remains significantly associated with an improved PFS (p = 0.0087) after adjustment for cohort, PD-L1 and TMB. Conclusions: This is the first exploratory report of serum choline levels in pan-cancer pts receiving pembrolizumab. The association between improved PFS and D choline suggests a possible role for the cholingeric system in the regulation of antitumor immunity. Further nonclinical and clinical studies are required to validate this finding.

1088 Background: Select patients (pts) with metastatic breast cancer (MBC) treated with capecitabine (cape) experience long-term disease control. We aimed to evaluate the clinical and biological predictors of long-term response. Methods: Pts receiving cape monotherapy for breast cancer from 01/2006 to 01/2016 at Princess Margaret Cancer Centre in Toronto, Canada (N = 352) or Alberta, Canada (N = 798) were identified through central pharmacy records. A median time-to-progression (MTTP) 3-fold higher (19.3 months) than those seen in published studies for patients treated with cape monotherapy at 1000mg/m2 was applied to select for pts with long-term response. MBC pts meeting these criteria were identified through chart review, with collection of clinical, pathological, and survival outcomes in addition to oncologist assessed (o) best response (eg: oCR, oPR, oSD) by radiology report review. Descriptive statistics, Kaplan-Meier and Cox-regressions were applied. Results: Overall, 41 (4%) pts met long-term response criteria. Median age of the study cohort was 62 (range 40-80), with 39% metastatic at diagnosis and 76% post-menopausal. At initiation of cape, a majority of pts were HR positive (85%), with an LDH < ULN (56%), and had bone (83%) or visceral (63%) metastases. Only 1 (2%) patient was HER2+. In the metastatic setting, most patients were chemotherapy-naive (83%) and had received 0-1 lines of hormonal therapy (61%). Median treatment duration with cape was 2.2 years (range 1.7-5.7) with 37% of pts having ≥40 cycles (range 22-94). Visceral (49%), bone (34%), and lymph nodes (24%) were the most common sites of progression, with 41% requiring 1 dose reduction and 27% requiring 2 dose reductions of cape. Overall response rate was 56% (oCR = 2%, oPR = 54%) with 44% having oSD as best response. Median follow-up was 6.8 years (95%CI: 5.5-8.2). Median PFS was 2.3 years (2.0-2.6), with a median OS of 3.8 years (2.9-4.6). 5 patients (12%) remained alive at data-cutoff, and 1 (2%) remained on treatment after 94 cycles of cape. Conclusions: Clinical features associated with long-term response on cape include HR positive, HER2 negative postmenopausal patients with bone predominant metastasis who have received 0-2 lines of prior hormone therapy and 0-1 lines of chemotherapy in the metastatic setting. This is the largest reported analysis of MBC patients with prolonged responses to cape. In the absence of randomized controlled trials; real-world evidence could aid clinicians in the optimal patient selection for treatment with cape.

2014 Background: Typically symptomatic adverse events (sy-AEs) on clinical trials are reported by clinicians using the CTCAE. To complement clinician collected sy-AEs and understand tolerability better, the patient report outcome version of the CTCAE (PRO-CTCAE) has been developed to provide the patient (pt) perspective on severity of AEs (graded scale 0-4) and their interference in daily life (scale 0-4). The aim of this study was to correlate PRO responses with the grade (G) of AEs, dose interruptions/reductions and dose limiting toxicities (DLTs). Methods: Pts enrolled on phase 1 clinical trials at Princess Margaret were surveyed electronically on tablet using the full library of items for PRO-CTCAE. The PRO-CTCAE was administered at baseline (prior to therapy), mid-cycle 1, and mid-cycle 2. AEs on study were recorded by physicians using the CTCAE. The electronic medical records were analyzed for an association between reported sy-AEs and PRO score. Summary statistics were used to describe patient and disease characteristics, as well as the outcomes. Spearman’s method was used to correlate PRO severity and interference responses. Logistic regression was used to assess which factors were associated with CTCAE G 3-4 vs G 2 AEs. Results: We analyzed 158 pts: median age 60yrs, 77 (49%) were male; all were ECOG ≤1 and 22, 55 and 81 pts completed 1, 2 and 3 surveys, respectively. Clinician reported G2, 3 and 4 sy-AEs occurred in 81, 47 and 3 pts, respectively and all of these were related to a PRO item except 5% (4/81), 9% (4/47) and 33% (1/3), respectively because either the AE occurred after 3rd time point or patient not able to complete the PRO (encephalitis). Sy-AEs causing dose interruptions, reductions, DLTs and discontinuations occurred in 45 (28%), 12 (7.5%), 5 (3%) and 12 (7.5%) pts, respectively; with a corresponding PRO item in 40 (89%), 12 (100%), 4 (80%) and 11(92%) pts, respectively. For patients who had CTCAE G2, G3/4 AEs, interruptions and discontinuations, their severity and inference levels were positively correlated (coefficient 0.49, p < 0.001; 0.45, p < 0.001 0.59, p < 0.001, 0.86, p < 0.001). Dose interruptions (p = 0.0027) and reductions (p = 0.0061) were significantly associated with G3-4 compared to G2 AEs. Conclusions: This is the first time an association between PRO-CTCAE severity and interference; and CTCAE G2, 3, 4 AEs, dose interruptions and discontinuations has been demonstrated. Additional modelling and more patient data are being analyzed to explore the relationship.

e19303 Background: There is limited real-world evidence of impact of large clinical panel sequencing on treatment-matching for patients with advanced solid tumors. The province of Ontario has a single payer, publicly funded health care system. We linked genomic testing results from a prospective province-wide trial, OCTANE (Ontario-Wide Cancer TArgeted Nucleic Acid Evaluation), to administrative data to determine the feasibility of this approach for evaluating survival and the impact of sequencing on treatment matching. Methods: We linked all Ontario patients from Princess Margaret (PM) with panel testing results (tumor-only 555-gene panel) to province-wide administrative data on treatments and outcomes. Patients were recruited from August 2016 to August 2018. Only clinically actionable variants based upon OncoKB annotation (Level 1 and 2) were assessed for genotype-informed treatment matching. Results: All 888 eligible patients were successfully linked to administrative data. Mean age was 58 (±13) years, 635 (71.5%) were female. Most common disease sites were ovary (26.4%), uterus (14.0%), colorectal (11.8%) and breast (9.5%). Administrative data vital status was more complete than trial collected data with 262 of 476 deaths only recorded in administrative data. Median survival was 1.70 years (95% confidence interval 1.50-1.91). 247 (27.8%) had actionable mutations, most commonly PIK3CA (54.7%), BRCA1 (15.8%), BRCA2 (15.0%) and BRAF (8.9%). 37 (15.0%) and 42 (17.0%) patients with actionable mutations received targeted therapy within 6 and 12 months of test report date, respectively. Conclusions: This is the first known feasibility study of linked administrative data to measure outcomes of large clinical panel sequencing for patients with advanced solid tumors. Vital status was more complete with administrative data compared to trial-collected data, and treatment data was successfully linked. About one in twenty-one enrolled patients received genome-informed treatments within 12 months, or about one in six of all patients with actionable mutations. This may be due to short interval follow up, trial and drug access, successful standard of care treatments, early patient deterioration or limited alterations covered by the panel, among other causes.

3562 Background: Next generation sequencing targeted panels increasingly inform clinical decisions, but may miss actionable findings detectable by whole exome sequencing (WES) and RNA-seq. There has been no direct comparison of WES plus RNA-seq against targeted panel sequencing to determine its added utility. To address this, we performed WES and RNA-seq analysis in a cohort of 100 patients with no actionable findings on prior panel sequencing. Methods: Ontario-wide Cancer Targeted Nucleic Acid Evaluation (OCTANE; NCT02906943) has sequenced 2,106 patients using a 161-gene Oncomine or 555-gene Hi5 panel. 100 patients (98 Hi5, 2 Oncomine) were chosen for further sequencing. Tumor (100x coverage) and normal (50x) exomes and tumor transcriptomes were sequenced on Illumina HiSeq2500 or NextSeq550. Interpretation included knowledgebase annotation (e.g. OncoKB, CIViC), mutation signatures, homologous recombination deficiency (HRD) scores, gene expression, and pathway analysis. Findings were deemed “actionable” if they could directly inform management or trial eligibility. Results: WES and RNA-seq identified one or more novel actionable findings in 38 patients. Of these, the main actionable finding was tumor mutation burden (TMB), mutation signature, or HRD score in 19 (50%), a copy number variant in 16 (42%), a fusion in 2 (5%), and a point mutation in 1 (2.6%). WES identified a MALAT1-GLI1 fusion in a cancer of unknown primary (CUP) whose transcriptome was consistent with gastric cancer, together suggesting the diagnosis of a rare gastroblastoma. To date, two cases have received exome-supported targeted therapy: (1) a metastatic high grade serous ovarian cancer, HRD-high, treated with olaparib then cisplatin for a combined 15 months, and (2) a metastatic neuroendocrine rectal tumor with RICTOR amplification treated with everolimus starting in Dec 2016 until last follow-up in Sep 2019. Of 62 patients with no actionable finding, expanded sequencing identified one or more known cancer drivers in 25 (40%): 17 CNVs, 3 fusions, and 5 point mutations or indels. In 16 patients, an oncogenic variant found on panel was not captured by WES, and may represent artifacts, germline mutations, or subclonal/localized variants. Conclusions: WES and RNA-seq expanded detection of actionable biomarkers and oncogenic mutations, especially CNV, TMB/signatures, and HRD. Two cases have undergone exome-supported targeted treatment. We performed the first WES of a rare gastroblastoma, originally a CUP but reclassified by expanded fusion detection and RNA-seq.

396 Background: Most guidelines recommend active surveillance (AS) as initial management for stage I testis cancer (TC). AS entails blood work and imaging at regular intervals requiring multiple clinic visits spanning 5 years. This can be time-consuming, difficult to adhere to, dissatisfying and costly for patients and health care systems. We innovated a secure online platform, WATChmAN (Web-based virtuAl Testicular CANcer clinic) allowing asynchronous communication between patients, results, and physician team. Methods: We are conducting an RCT (NCT03360994) where patients with stage I TC on AS are randomized to virtual care (WATChmAN), or standard in-person care. Primary endpoint is safety: examining loss-to-follow-up and compliance with AS schedules, incidence of relapse, delays in detection of relapse, and burden of relapse. Non-compliance represents a) patient-derived delay in visit; or b) follow-up visit with incomplete testing. Secondary endpoints include: patient/physician satisfaction and cost savings. Results: At present, 102 of a planned 144 patients are enrolled: 51 to virtual care, and 51 to standard in-person care. More patients in the virtual arm have been compliant with AS schedules (89% vs 73%) with shorter median compliance delays (12 vs. 14 days). To date, 10 patients have relapsed: 6 virtual (11.8%) and 4 standard (7.8%). Median time to relapse was shorter for the virtual arm (8 vs. 9.5 months), with no difference in burden of disease at relapse. Response rates to 6-month surveys were 90% and 59% for virtual and standard arms respectively. When asked if satisfied with their care, on the virtual arm 67% reported “extremely satisfied”, and 33% “satisfied” compared to 50% and 45% for the standard arm. When WATChmAN patients were asked if the application is able to provide the same excellence of care as in-person appointments, 82% reported “strongly agree” or “agree”. Conclusions: Interim results suggest virtual care in stage I TC is feasible and safe with improvements in patient satisfaction. Through semi-structured interviews and cost-effectiveness analyses, we anticipate more insight into virtual care. This may serve as a potential model for virtual care for other cancers. Clinical trial information: NCT03360994.

402 Background: Recent trials randomising patients (pts) receiving systemic cancer therapy showed that prophylactic anticoagulation (PAK) halves the risk of venous thromboembolic events (VTE) and doubles the risk of bleeding (Khorana et al. & Carrier et al., both NEJM 2019). In pts with metastatic germ cell tumors (mGCT) VTE is a frequent complication but it remains unclear whether PAK should be recommended because the number of mGCT pts in those trials was small. We aimed to determine the risk of VTE before, during and after chemotherapy and in mGCT pts without and with risk factors for VTE (retroperitoneal lymph nodes, Khorana score, venous access device) and to calculate the number needed to treat (NNT) and number needed to harm (NNH) of PAK. Methods: This retrospective analysis included mGCT pts treated with first-line platinum-based chemotherapy. We excluded patients who received PAK, with a known history of coagulopathy or VTE and extracted data about VTE and bleeding events. Cumulative VTE incidence was calculate for patients without and with increasing number of known risk factors for VTEs. NNT and NNH were calculated by assuming similar hazard ratios (HR) to reduce VTEs and increase bleeding as previously published (HR 0.66 and 1.96, Khorana et al., NEJM 2019). Results: Out of 1039 pts, 132 (13%) presented with VTE, 6 (1%) with bleeding. One patients died of VTE and 1 because of bleeding. Patients without any VTE risk factors experience VTE in 20/347 (5%) which translated into a NNT of 55 compared to the NNH of 84 respectively. Before start of chemotherapy 52 (5%) pts (NNT=60) presented with VTE of which 22 were reported symptomatic 21 asymptomatic/incidentally detected VTE on staging scans (9 unknowns). During chemotherapy 79 (8%) pts (NNT=40) were diagnosed with VTE whereas 19 (2%) pts (NNT=162) were diagnosed with VTE after chemotherapy. Conclusions: Our analysis revealed that even mGCT patients without risk factors for VTE show a relevant cumulative VTE incidence of 7%. Especially before and during but not after chemotherapy the benefits of PAK to prevent VTE outweighs the small increased risk of bleeding.

198 Background: Docetaxel and androgen receptor (AR) targeted agents have become standard options for mCSPC. CGP data is limited in mCSPC, however has the potential to guide treatment selection. We aimed to establish the feasibility of using CGP to characterize genomic alterations (GAs) in mCSPC. Methods: Patients with mCSPC were prospectively recruited at the Princess Margaret Cancer Centre to the OCTANE trial (NCT02906943), which aims to establish a genomically-characterized and clinically-annotated patient base that can be enrolled into specific research initiatives. Archival tumour specimens (after clinical testing completed) were profiled using next-generation sequencing (NGS) with a custom hybridization capture DNA-based panel (555 genes) or a targeted DNA/RNA amplicon panel (Oncomine Comprehensive Assay v3, 161 driver gene panel). GAs were classified according to the 2017 AMP/ASCO/CAP consensus recommendation. Clinical data were extracted from electronic health records. Results: Since 2016, 17/31 (55%) mCSPC patients who enrolled had sufficient tissue (15 from primary, 1 lymph node, 1 bone) for CGP. All had adenocarcinoma histology, most had high volume (76%) or de novo (71%) mCSPC. Median presenting PSA was 56 µg/L (range 8.4 – 1394). Patients received docetaxel (88%) or abiraterone (12%). Tier I/II GAs consistent with possible benefit from approved or investigational targeted therapies were identified in 14/17 patients (82%). AR pathway GAs (in 29%) were present regardless of prior androgen deprivation therapy or disease volume. DNA damage repair GAs (in 65%) included BRCA2, ATM, PALB2, MLH1/3, CDK12, FANCA, ERCC4, and SPOP. Other GAs involved the cell cycle, mTOR, EGFR, wnt, hedgehog, notch, and epigenetic pathways. Conclusions: In this small cohort, CGP was feasible in the setting of mCSPC for half of the patients. Potentially actionable GAs were identified in over 80% of patients tested. Pending confirmation from larger cohorts, these data support CGP in mCSPC prior to systemic therapy initiation, and highlight its potential role in future biomarker development and trial design in mCSPC.

Genotyping of patient tumors has been rapidly incorporated into both clinical trials and clinical practice over the last several years, with the goal of advancing personalized or precision cancer medicine. The crux of these efforts is to identify actionable alterations— genetic changes that can be used to guide individualized treatment. Dramatic advances in sequencing technologies and DNA isolation methods now enable multiplex hotspot mutation testing or targeted exome se

Purpose Cisplatin-based chemotherapy, a mainstay of treatment for disseminated germ cell tumors (GCTs), is associated with venous thromboembolism (VTE). Many patients with disseminated GCTs have large retroperitoneal lymph node (RPLN) metastases that may cause venous stasis and increase the risk of VTE development. We hypothesized that there was an association between large RPLN and chemotherapy-associated VTE risk. Patients and Methods The training cohort was composed of patients with disseminated GCT receiving first-line chemotherapy at Princess Margaret Cancer Centre between January 2000 and December 2010. Large RPLN was defined as more than 5 cm in maximal axial diameter. The predictive and discriminatory accuracies of a model using large RPLN in predicting VTE were compared with high-risk Khorana score (≥ 3) using logistic regression and area under receiver operator characteristic curves (AUROCs). The model was externally validated in a cohort of patients treated at the London Health Sciences Centre. Results The training cohort comprised 216 patients, 21 (10%) of whom developed VTE during chemotherapy. VTE was associated with large RPLN (odds ratio [OR], 5.26; P = .001), high-risk Khorana score (OR, 11.8; P < .001), intermediate-/poor-risk disease (OR, 3.76; P = .005), and hospitalization during chemotherapy (OR, 4.24; P = .002). Large RPLN showed higher discriminatory accuracy than high-risk Khorana score (AUROC, 0.71 v 0.67, respectively). Superior discriminatory accuracy of large RPLN over high-risk Khorana score was validated in the London cohort (AUROC, 0.61 v 0.57, respectively). Conclusion Large RPLN is associated with VTE in patients with disseminated GCT and provides higher discriminatory accuracy than high-risk Khorana score. Results should be validated in larger, prospective studies. Prophylactic anticoagulation may be considered in high-risk patients.

1034 Background: Anti-HER2 therapy has improved survival in HER2+ MBC. Yet, large patient cohorts with no evidence of disease (NED) with long-term follow up are incompletely described in the literature. We evaluated the clinical characteristics of patients with HER2+ MBC and prolonged response to anti-HER2 therapy, exhibiting NED vs Residual disease (RES). Methods: Patients treated with chemotherapy plus trastuzumab (CT) from 2005-2013, or taxane plus trastuzumab-pertuzumab (TTP), or trastuzumab-emtansine (TDM1) from 2012-2016 for HER2+ MBC at Princess Margaret Cancer Centre in Toronto, Ontario or in Alberta, Canada were included. Duration on anti-HER2 therapy (without switch) was collected. Patients with median duration of response (MDR; months) 2x higher than phase II/III trials for each regimen (CT = 18.2; TTP = 40.4; TDM1 = 25.2) were included to select for prolonged response. Clinical features (ie: stage at diagnosis, survival, etc) and oncologist/radiologist reported best response were collected. Responses were grouped as NED (including sclerotic bone metastases) or RES. Clinical variables were evaluated by Chi-square and survival by Kaplan-Meier (log-rank). Results: 2403 patients (CT = 1830; TTP = 394; TDM1 = 179) were evaluated. After cut-off, 119 patients (5%) were included, with NED in 41% (49/119) and RES in 59% (70/119). More women aged < 50 vs ≥50 (p = 0.015) had NED (61%) than RES (39%). No breast surgery (curative or metastatic) showed higher (p = 0.015) rates of NED (49%) relative to RES (27%). More women having NED (92%) than RES (37%) were still alive (p < 0.0001) at data cut-off, with improved mPFS (years; p < 0.0001) for NED (11.2, 95%CI: 11.2-NR) vs RES (3.0, 2.7-3.6), and mOS (years; p < 0.0001) for NED (NR) vs RES (3.4, 95%CI: 2.8-4.2). Significance persisted with TDM1 patients excluded. Treatment type (p = 0.053) and number of organs with metastases (p = 0.067) were of borderline significance. Conclusions: Patients with NED have improved survival compared to RES. Younger age and avoiding breast surgery correlates with NED. Evaluation of genomic factors influencing NED are planned.

TPS2664 Background: Differences in microbiome diversity and composition in immune checkpoint inhibitor (ICI)-responders vs non-responders have been demonstrated. Transplantation of responder feces in mouse models recapitulated the ICI-responsive phenotype. MET-4 is an oral alternative to fecal transplant consisting a well-defined mixture of intestinal bacteria isolated from healthy donor stool sample. We hypothesize that co-administration of MET-4 with ICI is safe and results in alterations of the gut microbiota. Methods: Three cohorts (n = 65) of subjects with any advanced solid tumor type treated with monotherapy anti- PD1/PD-L1 antibody outside of a therapeutic clinical trial will be enrolled. Group A: safety cohort of 5 subjects already on ICI will receive MET-4 in addition to standard of care (SOC) ICI. If < 2 subjects report adverse events of CTCAE grade ≥3 within 4 weeks at least possibly related to MET-4, this dose will be declared safe and groups B/C may start enrolling. Group B (n = 40): subjects with advanced solid tumors starting on ICI, randomized 3:1 to MET-4 plus SOC vs. SOC. Group C (n = 20): subjects with advanced solid tumors already on ICI with first unconfirmed disease progression randomized 1:1 to MET-4 plus ICI continuation vs. continuing ICI. Serial stool samples will be collected for taxonomic composition, diversity, metagenomics content and MET-4 species abundance. We anticipate the following analyses: 16S rRNA sequencing, shotgun metagenomics sequencing, qPCR, Nanostring nucleic acid detection and metabolomics profiling. Serial blood sampling for flow cytometry/CyTOF. Immune microenvironment of tumor specimen will be examined using immunohistochemistry. Other major inclusion criteria: willingness to provide correlative samples, RECIST v1.1 measurable disease and ECOG 0-2. Subjects unable to swallow oral medications are excluded. For the primary objective of cumulative relative abundance and changes of ICI-responsiveness associated species between baseline and day 12 MET-4, assuming a change of 0.5 standard deviation (SD) of microbial alpha diversity, our study will have ≥84% power to identify a significant difference given a significance level at 0.05 in group B. Assuming a change of 0.9 SD of microbial alpha diversity, we will have ≥83% power to identify a significant difference in group C. Response rates and progression free survival will be assessed per RECIST v1.1 and compared with historical data. Clinical trial information: NCT03686202.

1086 Background: Hyperprogression of disease (HPD), a rapid acceleration of tumor growth rate (TGR) has been reported with IO in other tumor types. Here, we explore HPD in aTNBC. Methods: A retrospective chart review identified aTNBC patients who consented for IO clinical trials at Princess Margaret Cancer Centre between June 2013 and June 2018. Demographic data, medical history, details of trial enrolment and RECIST 1.1 response to study treatment were recorded. Patients with RECIST 1.1 measurable disease on CT scans or physical examination before trial entry, at trial baseline and at protocol-defined interval following IO start were evaluable for TGR as defined by Champiat et al. Clin Cancer Res 2017. HPD defined as a ≥2-fold increase in TGR between baseline and on-trial restaging assessment. Univariable logistic regression used to identify variables [age, co-morbidity index, prognostic index, performance status, distant disease free interval (dDFI), lactate dehydrogenase, no. of metastatic sites, visceral disease and no. of prior treatment lines] associated with HPD. Overall survival (OS) curves were estimated with the Kaplan-Meier method and compared by the log-rank test. Results: 99 patients with aTNBC consented for 15 IO clinical trials, 60% IO monotherapy, 22% chemotherapy+/-IO and 18% IO combinations. Median age 52 (range 25-78), median no. of lines of prior systemic therapy for advanced disease 1 (range 0-8). 15% had de-novo metastatic disease, 58% recurred after a dDFI of < 3 years and 25% after a dDFI of > 3 years. 61% had < 3 metastatic disease sites, and 71% had metastases involving the viscera. 66 received IO treatment, 40 patients (20 monotherapy, 7 IO combination, 13 chemotherapy+/-IO) were evaluable for TGR. Median TGR pre-IO was 74.3 (range -17 – 1680) and post-IO was 2.5 (-71.4 – 223). 4 patients (10%) met criteria for HPD. All 4 treated with monotherapy PD1 inhibitor and received at least 2 further lines of therapy post-trial; 1 patient treated with IO as first-line therapy, 3 in the second or later lines. There was no significant difference in the overall OS of patients with HPD and patients who did not meet definition for HPD HR 0.89, (95% CI: 0.26-3.01; p = 0.41). Univariable analysis did not identify factors associated with HPD. Conclusions: HPD was observed in 10% of aTNBC treated on IO clinical trials. HPD was not associated with worse survival outcomes or known prognostic factors in our analysis.

e16062 Background: Testicular cancer survivors may experience long-term psychological distress related to their diagnosis and treatment. Measurement of mental health service use (MHS) is an objective way to quantify the incurred burden of psychological distress. Here we describe patterns of MHS use among survivors of testicular cancer in a single-payer health system. Methods: The Ontario Cancer Registry was linked to electronic treatment records to identify all incident cases of testicular cancer treated with orchiectomy in the Canadian province of Ontario during 2004-2010. Mental health clinical visits were identified from records of hospital admission, emergency room visits and out-patient physician billing records. Results: The study cohort included 1877 patients; mean age was 35 and 61% had pure seminoma. Two thirds of patients (66%, 1230/1877) were initially treated with active surveillance. In the 2-year period prior to orchiectomy, 24% (443/1877) of patients had MHS use. Post-orchiectomy, the prevalence of MHS use was 18% (331/1877), 23% (431/1877) and 30% (572/1877) at 6, 12 and 24 months, respectively. The use of MHS was greatest in the peri-diagnostic/surgical period. Rates of MHS visits, 3 months before, 0-3 months after, 4-6 months after and 2 years after orchiectomy were 11%, 14%, 10%, and 6%, respectively. Post-orchiectomy MHS use was greatest among those with baseline MHS use. The prevalence of MHS use for patients with and without baseline MHS use was 40% (175/443) vs 11% (156/1434, p < 0.001) at 6 months; 51% (225/443) vs 14% (206/1434, p < 0.001) at 12 months; and 61% (271/443) vs 21% (301/1434, p < 0.001) at 24 months post-orchiectomy. Post-orchiectomy MHS use did not vary by initial treatment strategy (active surveillance vs upfront treatment, p = 0.435). Older patients were more likely to have MHS visits at all time periods than younger patients; MHS rates at 6 months were 15% (93/631) vs 18% (112/625) vs 20% (126/621) for ages 16-29, 30-39, and 40+ years respectively (p = 0.035). Conclusions: MHS use among survivors of testicular cancer is common, particularly in the peri-diagnostic and surgical period. The rate of MHS use is highest among those patients with a history of prior MHS use. Clinicians should screen for the presence of psychological distress among survivors of testicular cancer.

2542 Background: Limited data exist in the clonal dynamics of serial ctDNA as a predictive biomarker in advanced solid tumor pts receiving immune checkpoint blockade. Methods: Pts with mixed solid tumors received single agent P (anti-PD-1) 200 mg IV Q3wks in the investigator-initiated phase II INSPIRE trial (NCT02644369). ctDNA was assayed at baseline (B) and start of cycle 3 (C3) using a pt-specific amplicon-based NGS assay (Signatera™). Samples were considered ctDNA positive if ≥2 of 16 pt-specific targets met the qualifying confidence score threshold. Results: Results of 70 pts are presented. Demographics: male 46%; median age=60 yrs (range 21–82); head and neck (20%), triple negative breast (14%) and ovarian (14%) cancers comprised the major malignancies. Median no. of P cycles=4 (range 2–35); follow up was 14m (range 2–29); RECIST responses: CR 2.9% (n=2), PR 17% (n=12), CBR (CR+PR+SD≥6 cycles) 31% (n=22), RECIST/clinical PD (n=43/10; 65%/15%). Median PFS=3.3m and median OS=17.8m. 68/70 pts had ctDNA detected at baseline (median=16/16 variants) demonstrating 97% sensitivity. Table shows correlation of ΔctDNA (ctDNAB compared to ctDNAC3) with clinical efficacy parameters, whereas ctDNAB values did not reach statistical significance. Conclusions: A strong correlation exists between ΔctDNA with OS, PFS, CBR and ORR with P, suggesting it is a potential predictive biomarker in pts with mixed solid tumors. Clinical trial information: NCT02644369. [Table: see text]

e16056 Background: Bleomycin is commonly used to treat testicular cancer and can be associated with severe pulmonary toxicity. There is limited information about how clinicians monitor patients during treatment and the incidence of pulmonary toxicity in routine practice. Methods: The Ontario Cancer Registry was linked to electronic records of treatment to identify all incident cases of testicular cancer treated with orchiectomy and bleomycin, etoposide, and cisplatin (BEP) chemotherapy in the province of Ontario during 2005-2010. Health-administrative databases were used to describe use of pulmonary function tests (PFTs), chest imaging and physician visits. Results: 475 patients were treated with orchiectomy and chemotherapy. Complete chemotherapy records were available for 93% (368/394) of men treated with BEP. Bleomycin was omitted among 32% (116/368) of patients. PFTs were performed in 17% (63/368), 17% (61/368) and 29% (106/368) of patients before BEP, during BEP, and within 2 years of finishing BEP, respectively. During chemotherapy, 62% of patients (227/368) had chest imaging. In the two years following BEP, 23% (85/368) had a physician visit for respiratory symptoms; this rate was substantially higher among men with greater exposure to bleomycin; 40% (24/60) for 10-12 doses bleomycin vs 21% (53/250) for 7-9 doses vs 14% (8/58) for 1-6 doses (p = 0.002). Two percent of men (8/368) had visit codes for pulmonary fibrosis. Conclusions: A substantial proportion of men treated with BEP will seek medical attention after chemotherapy for respiratory symptoms and this is associated with cumulative dose of bleomycin. Use of PFTs and chest imaging during treatment is common. Whether PFT test results or clinical symptoms are leading to bleomycin dose omission is uncertain.

523 Background: Bilateral testicular germ cell tumours (BTC) form a small minority of testicular cancer and detailed management data are sparse. Methods: Bilateral testicular cancer (BTC) patients managed at a single cancer centre were retrospectively analyzed. Synchronous BTC was defined as uni+contralateral presentation within 3 months. Patient characteristics, treatment and outcomes were collected. Kaplan-Meier method was used to calculate the overall survival (OS) and relapse-free survival (RFS). Results: Between Jan 1971 to Jun 2018, 118 pts were included. Nine patients (7.6%) had cryptorchidism. Twenty-two patients (18.6%) had synchronous BTC at median age of 30(21-54) years, 11 presented with concordant histology (10-seminoma). Median follow-up time was 96(1-220) months. Two of 14 patients (14%) with stage I disease on surveillance had retroperitoneal nodal recurrence, other 3 (21%) had testicular recurrence after partial orchiectomy alone. No recurrence occurred for 8 stage II/III patients (36%) who received stage-appropriate treatment. All patients were alive without disease at last follow-up. For metachronous BTC, the median age was 27(16-68) and 37(19-78) years for first and second diagnosis, respectively. The median time interval was 88 (8-352) months, with shorter interval when second primary was non-seminoma, median 69 vs. 92 months. Concordant histology was present in 58 (38-seminoma) patients and discordant in 38 patients. There were 66, 23, 7 and 84, 9, 3 patients with stage I, II, III disease for first and second testicular cancer (TC), respectively. For all stage I disease, 69% of non-seminoma (n = 33) and 79% of seminoma (n = 81) were on surveillance, of whom the crude relapse rate was 15%. The median follow-up time after second diagnosis was 87 months. In all, 35 patients (30%) with recurrence except 1 were successfully salvaged. The 10-year OS and RFS for whole cohort was 99% and 69.8%, respectively. Conclusions: In our series, seminoma was the more common pathology, and management based on pathology and stage yielded excellent outcomes regardless of prior therapy. Metachronous BTC may occur at extremely long time intervals such that longer follow-up is needed to capture the majority of contralateral primary TC.

519 Background: Even in patients (pts) with widely metastatic germ cell tumors cure rates are high with cisplatin based chemotherapy. Survivors of testicular cancer often have impaired gonadal function possibly related to chemotherapy. Most pts develop temporary azoospermia with recovery in about 50% after 2 years and 80% after 5 years. Platinum persists in blood in these pts, however, it is not known whether platinum also persists in semen. Methods: Pts who completed cisplatin > 3 months previously were enrolled. Age, total cisplatin dose and date of completion of treatment were collected. A semen sample was collected from each patient for analysis to assess semen quality. A blood sample was collected to assess FSH/LH/testosterone/creatinine. Serum and semen platinum levels were measured using HPLC-tandem mass spectrometry. Results: Between 11/2017 and 09/2018, 9 pts (median age 32 years; range: 18-52) were enrolled to the study, all were treated with standard Bleomycin, Etoposide, Cisplatin regimen. Blood samples were collected from 7 pts, and semen samples from 4 pts. Median total cisplatin dose given was 658 mg (range: 570-780). Median serum platinum concentration was 0.395 ng/mL (range: 0.10-0.94) at a median of 10 months (range 4.5-13) post completion of treatment. Median semen platinum concentration was 1.06 ng/mL (range: 0.25-1.47) at a median of 10 months (range: 6-14) post completion of treatment. Semen platinum levels were associated with total cisplatin dose administered (Pearson correlation, r = 0.645) and time from completion of treatment (Pearson correlation, r = 0.386). In all 4 pts with matched samples, the semen platinum level was higher than the serum platinum level. Semen analysis showed 2 pts were azoospermic post treatment, 1 of these was also azoospermic prior to treatment; 1 pt had normal semen analysis post treatment and 1 had a low sperm count post treatment. Conclusions: This is the first study demonstrating that platinum persists in semen > 6 months post completion of cisplatin-containing chemotherapy. Platinum levels are higher in semen than in blood following treatment. Our preliminary findings may have important implications for reproductive health of survivors of advanced testicular cancer.

1029Background: Immune checkpoint inhibitors are active in a broad range of advanced cancers. Antibody dependent cell mediated cytotoxicity is a mechanism of action of trastuzumab. The combination ...

TPS12127Background: Genomic profiling of cancers is increasingly used to refine prognostication and aid in treatment decisions by allowing matching of targeted agents to specific genetic variants. ...

3063Background: A subset of patients (pts) treated with immune checkpoint inhibitors experience an accelerated tumor growth rate (TGR) compared with pre-treatment kinetics, known as hyperprogressio...

2515Background: Notch signaling can be deregulated in human cancer. AL101 (formerly BMS-906024), a gamma secretase inhibitor that potently inhibits all Notch receptors, was evaluated. The primary o...

e18551Background: Cancer immunotherapy drugs are now approved to treat many types of cancer. Little is known about patient knowledge, attitudes, and expectations of cancer immunotherapies. Methods:...