Search

6579 Background: Larotrectinib trials in Tropomyosin Receptor Kinase (TRK) fusion cancer population were single arm trials and therefore limited comparative effectiveness data with larotrectinib are available.MAIC is typically used to balance population characteristics to facilitate cross-study comparisons. The objective of this study was to use MAIC to compare efficacy of the highly specific TRK inhibitor larotrectinib vs. SoC. Methods: Individual patient data from larotrectinib trials (NCT02122913, NCT02637687, NCT02576431) were compared with published aggregate SoC data from patients with locally advanced/metastatic TRK fusion cancer identified in the Flatiron Health/Foundation Medicine clinico-genomic database (Demetri et al. Ann Oncol. 2021). Prior to matching, eligible patients were ≥18 years and had to have received ≤4 lines of prior therapy (LoPT). Patients were matched on available common baseline characteristics (Table). Overall survival was the only endpoint assessed and was defined as time from locally advanced/metastatic disease diagnosis to death. The analyses included: 1) a log-rank test of equality to test whether the two groups were similar before larotrectinib initiation; and 2) estimation of treatment effect of larotrectinib vs. SoC. Hazard ratios (HRs) were used to compare the 2 groups. MAIC assumes that all observed and unobserved prognostic factors are adjusted for in the analysis. Results: 85 larotrectinib patients and 28 SoC patients were included. After matching, log-rank testing suggested no difference between the 2 groups (P=0.26), and larotrectinib was associated with a 78% lower risk of death (adjusted HR: 0.22 [95% confidence interval: 0.09, 0.54]; P=0.001), compared to SoC. Conclusions: This analysis suggests longer overall survival with larotrectinib, compared to SoC, in adult patients with TRK fusion cancer. The current analysis was limited to the prognostic variables available in the RWD. Further data are warranted to confirm those results. [Table: see text]

2010 Background: Neurotrophic tyrosine receptor kinase ( NTRK) gene fusions are known oncogenic drivers in a variety of tumor types. Larotrectinib is a highly selective, CNS-active TRK inhibitor that demonstrated an objective response rate (ORR) of 30% and a 24-week disease control rate (DCR) of 73% across 33 evaluable adult and pediatric patients with TRK fusion primary CNS tumors, as of July 2020 (Doz et al, Neuro Oncol 2021). We report updated data on an expanded dataset of patients. Methods: Patients with TRK fusion primary CNS tumors in two clinical trials (NCT02637687, NCT02576431) were included. Larotrectinib was administered at 100 mg twice daily (BID) in adults and 100 mg/m2 (max 100 mg) BID in pediatric patients. Response was investigator-assessed. Results: As of July 2021, 38 adult and pediatric patients with TRK fusion primary CNS tumors were identified: high-grade glioma (HGG; n =23), low-grade glioma (LGG; n =9), and other (n =6; includes glioneuronal, neuroepithelial, diffuse leptomeningeal, neuroblastoma, recurrent small round blue cell, and not otherwise specified). Median age at enrollment was 10.8 years (range 1.3–79.0; 28 [74%] patients < 18 years old). The gene fusions involved NTRK2 (n = 28), NTRK1 (n = 6), and NTRK3 (n = 4). Sixteen (42%) patients received one prior line of systemic therapy and 16 (42%) received ≥2 prior lines. The ORR for 37 evaluable patients was 30% (95% confidence interval [CI] 16–47): three complete responses, eight partial responses, 21 stable disease (16 patients ≥24 weeks), and five progressive disease. The 24-week DCR was 73% (95% CI 56–86) for all patients, 68% (95% CI 45–86) for patients with HGG, and 89% (95% CI 52–100) for patients with LGG. Twenty-five of 31 patients (81%) with measurable disease at baseline had tumor shrinkage. Median time to response was 1.9 months. Median duration of response (DoR) was not reached; median follow-up was 25.6 months. The 12-month DoR rate was 64%. Median progression-free survival (PFS) was 16.5 months (95% CI 6.7–not estimable); median follow-up was 27.4 months. Median overall survival (OS) was not reached; median follow-up was 26.7 months. The 24-month OS rate was 65%. Treatment duration ranged from 0.1+ to 38.7+ months. Twenty-two patients (58%) progressed on treatment and three continued treatment post-progression for ≥4 weeks. Treatment-related adverse events (TRAEs) were reported in 21 patients (55%); the majority of these patients (18/21 [86%]) reported Grade 1 or 2 TRAEs. No Grade 3 or higher treatment-related neurological adverse events were reported. There were no treatment discontinuations due to TRAEs. Conclusions: Larotrectinib achieved a high DCR, rapid and durable responses, and a manageable safety profile in patients with TRK fusion primary CNS tumors. These results support testing for NTRK gene fusions in patients with CNS tumors. Clinical trial information: NCT02637687, NCT02576431.

2059 Background: Glioblastoma (GBM) is an aggressive brain tumor and despite efforts in developing new effective therapies, patient survival remains low. There is increasing interest in using immune checkpoint inhibitors (ICIs) for GBM, however the immunosuppressive (cold tumor) characteristics of GBM limit the efficacy of ICIs. Consequently, there is a need to identify patients with active tumor-infiltrating lymphocytes (hot tumor). Several studies have shown that brain extracellular matrix such as type IV collagen has a dynamic composition with protease-induced alterations. In this study, we evaluated the clinical utility of a non-invasive biomarker of granzyme B degraded type IV collagen (C4G) reflecting tumor-infiltrating lymphocyte activity and of matrix metalloproteinase (MMP) degraded type IV collagen (C4M) in GBM patients treated with nivolumab (anti-PD-1) and bevacizumab (anti-VEGF) in the recurrent setting. Methods: C4G and C4M were measured in serum from 22 controls and 39 GBM patients previously treated with surgery, radiotherapy and chemotherapy in the primary setting. After GBM recurrence, 18 patients underwent salvage resection (arm A) however 21 patients had no possibility for resection (arm B). All patients were treated with nivolumab and bevacizumab (NCT03890952 phase II study). Baseline GBM samples were taken before the second-line treatment. The association between C4G levels and outcome was evaluated by Cox regression analysis for overall survival (OS) and odds ratio (OR) calculations for complete response (CR) rate after dichotomizing patients into low vs high levels of C4G (median cutpoint). Results: C4G (p = 0.004), but not C4M (p = 0.166), was significantly elevated in serum from GBM patients compared to controls. Moreover, patients with high C4G levels had a significantly increased likelihood of experiencing CR (OR=6.68, p

9109 Background: Neurotrophic tyrosine receptor kinase ( NTRK) gene fusions have been identified as oncogenic drivers in a diverse array of tumor types including lung cancer. Larotrectinib is a first-in-class, highly selective, central nervous system (CNS)-active tropomyosin receptor kinase (TRK) inhibitor approved for the treatment of adult and pediatric patients (pts) with TRK fusion cancer, with an objective response rate (ORR) of 78% across multiple non-CNS cancers (McDermott et al, ESMO 2020). Here, we report the updated data on pts with lung cancer treated with larotrectinib. Methods: Pts with lung cancer harboring a NTRK gene fusion enrolled in two clinical trials (NCT02576431 and NCT02122913) were identified for this analysis. Larotrectinib 100 mg PO BID was administered on a continuous 28-day schedule until disease progression, withdrawal, or unacceptable toxicity. Response was assessed by the investigator per RECIST v1.1. Results: As of July 20, 2020, a total of 20 pts with TRK fusion-positive lung cancer (19 with non-small cell lung cancer and 1 with small cell lung cancer) were enrolled. Median age was 48.5 years (range 25.0–76.0). The gene fusions involved NTRK1 (n = 16; 80%) or NTRK3 (n = 4; 20%). Pts were heavily pre-treated with a median of 3 systemic therapies (range 0–6). Among 15 evaluable pts, the confirmed ORR was 73% (95% CI 45–92): 1 complete response, 10 partial responses (PR), 3 stable disease (SD) and 1 progressive disease (PD). The median time to response was 1.8 months. Among 8 evaluable pts with baseline measurable and non-measurable CNS metastases, the ORR was 63% (95% CI 25–91): 5 PR, 2 SD, and 1 PD. In all evaluable pts, the 12-month rates for duration of response and progression-free survival were 81% and 65%, respectively. Median overall survival was 40.7 months (95% CI 17.2 to not estimable) at a median follow-up of 16.2 months. Duration of treatment ranged from 0.03+ to 51.55+ months. Adverse events (AEs) were predominantly Grade 1–2. Treatment-related AEs were reported in 16 pts, of which 2 experienced Grade 3 events (myalgia, hypersensitivity, weight increase). There were no treatment discontinuations due to AEs. Conclusions: These data confirm that larotrectinib is highly active with rapid and durable responses, extended survival benefit, and a favorable long-term safety profile in pts with advanced lung cancer harboring NTRK gene fusions, including in pts with CNS metastases. These results underscore the importance of screening for NTRK gene fusions in pts with lung cancer. Clinical trial information: NCT02576431 and NCT02122913.

3114 Background: Larotrectinib is a highly selective, CNS-active tropomyosin receptor kinase (TRK) inhibitor that demonstrated rapid and durable responses in three phase I/II single-arm studies of patients (pts) with TRK fusion cancer. In single-arm studies the growth modulation index (GMI) can be used to provide a comparative analysis. GMI is an intra-patient comparison that uses pts as their own control by comparing progression-free survival (PFS) on current therapy to time to progression or treatment failure (TTP) on the most recent prior therapy; namely the ratio of PFS/TTP (EMA Guidelines. Guideline on the Evaluation of Anticancer Medicinal Products in Man, EMA/CHMP/205/95 Rev.5). A GMI ratio ≥1.33 has been used as a threshold of meaningful clinical activity. In a previous analysis of 122 pts with TRK fusion cancer treated with larotrectinib, 84 pts (69%) had a GMI ≥1.33. Conversely, 38 pts (31%) had a GMI < 1.33, but of these, 9 pts were ongoing treatment and censored for PFS as of July 2019 (Italiano et al, ESMO 2020). Here, we report the GMI of this initial group with a longer follow-up as well as an expanded dataset to more accurately assess the treatment effect of larotrectinib in pts with TRK fusion cancer previously treated with ≥1 line of therapy. Methods: Pts with TRK fusion cancer from three clinical trials on larotrectinib treatment with ≥1 prior line of systemic therapy were eligible for retrospective GMI analysis. TTP on the prior line of therapy was investigator-assessed. PFS on larotrectinib was determined by independent review committee per RECIST v1.1. Pts who had not progressed were censored as of date of last visit. Kaplan–Meier (KM) analyses were used to estimate median GMI, in addition to median PFS and TTP. The data cut-off was July 2020. Results: With an extended follow up of the original 122 pts, 90 (74%) pts had a GMI ≥1.33, including 6 of the 9 pts who were previously censored with a GMI < 1.33 and ongoing treatment; 6 pts (5%) had a GMI ≥1 to < 1.33 and 26 (21%) had a GMI < 1. The KM estimated median GMI increased from 7.6 (95% CI 5.7–88.0) to 9.5 (95% CI 5.7–17.4). In the expanded dataset of 140 pts, 103 pts (74%) had GMI ≥1.33, 7 (5%) had a GMI ≥1 to < 1.33 and 30 (21%) had a GMI < 1. Six of the 37 pts with a GMI < 1.33 were censored and still ongoing treatment. The KM estimated median GMI was 8.9 (95% CI 6.2–17.4). Among pts who had received 1, 2, or ≥3 prior lines of therapy, 74%, 65%, and 80%, respectively, had GMI of ≥1.33. Median TTP on the prior therapy was 3.0 months (95% CI 2.1–3.5) and median PFS on larotrectinib was 33.0 months (95% CI 16.6–34.9). Conclusions: With a longer follow-up, nearly three-quarters of pts with TRK fusion cancer treated with larotrectinib had a prolonged PFS compared to their most recent prior therapy. These results further validate the use of larotrectinib in treating patients with TRK fusion cancer. Clinical trial information: NCT02576431, NCT02122913, NCT02637687.

68 Background: HER2 amplification is an emerging biomarker in colorectal cancer (CRC) with increased copy number associated with improved clinical outcomes to HER2 targeting. RAS/RAF wildtype CRC also benefit from use of epidermal growth factor receptor inhibition (EGFRi). The sequencing of EGFRi versus HER2 inhibition in low copy number HER2 amplified CRC remains uncertain. Patient-derived cancer organoids (PDCOs) allow an ex vivo method to assess treatment sensitivity. We examined treatment sensitivity of a HER2 amplified PDCO at baseline and following resistance to panitumumab and trastuzumab. Methods: Following IRB-approval, fresh CRC tissue was cultured to maturation. After expansion, subcultures were treated with stepwise (20%) increase to physiologic Cmax of panitumumab (230ug/mL) and trastuzumab (180ug/mL). Threshold for escalation was median relative growth of +20% at 96h. Sensitivity was assessed on primary culture (RC1), panitumumab resistance (RC1-P) and trastuzumab resistance (RC1-T) using 96h of physiologic Cmax panitumumab, trastuzumab, and combination trastuzumab/pertuzumab. Individual sphere response was assessed for change in mean NADH autofluorescence intensity and ratio of NADH/FAD signal. Response was assessed at 96h in comparison to control using effect size of Glass’s Delta (GΔ). Results: Molecular profiling revealed HER2 copy number of 14 with no concurrent alterations in RAS, RAF, or PIK3CA. Time to resistance was similar between panitumumab (55 days) and trastuzumab (51 days). RC1 had baseline growth (+116%) which was reduced with single agent panitumumab (+17%, GΔ=1.40) with intermediate sensitivity to trastuzumab (+48%, GΔ=0.95) and trastuzumab/pertuzumab (46%, GΔ=0.99). Normalized NADH/FAD ratio revealed significant metabolic response to panitumumab (-20%, GΔ=0.66) and trastuzumab/pertuzumab (-35%, GΔ=1.16) with insignificant effect of single agent trastuzumab (-14%, GΔ=0.46). Following resistance to panitumumab, RC1-P had persistent growth with trastuzumab (+68%) which improved in combination trastuzumab/pertuzumab (+34%, GΔ=1.16). Following resistance to trastuzumab, RC1-T was insensitive to EGFRi with panitumumab including persistent growth (+58%, GΔ=0.70) and unchanged metabolism (+2%, GΔ=-0.10). Conclusions: Therapeutic dose escalation in a single PDCO of HER2 amplified CRC suggests improved sensitivity to EGFRi and dual HER2 targeting with trastuzumab/pertuzumab. Resistance to EGFRi resulted in persistent sensitivity to dual HER2 inhibition using trastuzumab/pertuzumab, however resistance to single agent trastuzumab. Resistance to trastuzumab resulted in future insensitivity to EGFRi. Molecular profiling at resistance revealed no pathologic alterations in EGFR or ERBB2 signaling, with ongoing analysis of transcriptional changes by RNAseq.

e15638 Background: Larotrectinib is a highly selective TRK inhibitor that demonstrated high response rates and durable disease control in three phase I/II single-arm studies of patients (pts) with TRK fusion cancer. While single-arm studies are often used for rare cancer populations, they do not provide comparative data. GMI utilizes pts as their own control and can be used in this setting. GMI is the ratio of progression-free survival (PFS) on the current therapy to time to progression (TTP) on the most recent prior line of therapy. A GMI ≥1.33 has been used as a threshold of meaningful clinical activity. We report GMI for TRK fusion cancer pts treated with larotrectinib. Methods: Data were pooled from 3 clinical trials of pts with TRK fusion cancer treated with larotrectinib. A retrospective, exploratory analysis of GMI was conducted in pts who had been on larotrectinib and followed up for ≥6 mos or pts who discontinued early, and had ≥1 prior line of systemic treatment for locally advanced or metastatic disease. TTP on the prior line of therapy was investigator assessed. PFS on larotrectinib was determined by independent review committee per RECIST 1.1. Pts who had not progressed were not censored from the analysis. Results: As of July 15, 2019, 122 pts were eligible for analysis. Fifteen different tumor types were represented, the most common being soft tissue sarcoma in 26 pts (21%), infantile fibrosarcoma in 22 (18%), and thyroid in 21 (17%). Median GMI was 3.35 (range 0.00–337.00; Table); In metastatic pts (n = 81), the proportion with a GMI ≥1.33 was higher in pts with a complete or partial response vs non-responders (88% vs 42%). In the whole analysis set (N = 122), median TTP on prior line of treatment was 2.7 mos (95% CI 2.0–3.1) and median PFS on larotrectinib was 33.4 mos (95% CI 13.8–NE; HR 0.20 [95% CI 0.14–0.29]). In metastatic pts (n = 81), median TTP on prior line of treatment was 2.3 mos (95% CI 1.9–3.0) and median PFS on larotrectinib was 23.4 mos (95% CI 10.9–NE; HR 0.24 [95% CI 0.16–0.36]). Conclusions: Greater than two-thirds of pts with TRK fusion cancer treated with larotrectinib had a GMI ≥1.33, demonstrating a clinically meaningful improvement in PFS compared to TTP on their prior treatment. Clinical trial information: NCT02122913, NCT02576431, NCT02637687.

3614 Background: NTRK gene fusions occur in diverse tumor types in adults and children. The selective TRK inhibitor, larotrectinib, has shown high response rates, durable disease control, and a favorable safety profile in patients (pts) with TRK fusion cancer. We report an expanded quality of life (QoL) analysis for pts treated with larotrectinib. Methods: QoL data were collected in two trials of larotrectinib in pts with TRK fusion cancer using EORTC QLQ-C30 (adults) and PedsQL (children) questionnaires, and were analyzed descriptively and longitudinally. EORTC QLQ-C30 global health scores (GHS) and PedsQL total scores range from 0 to 100, with higher scores indicating better QoL. We calculated the proportion of pts with normal/above and below normal QoL scores compared to values in the literature for the US general population. Results: By July 2019, 126 pts with TRK fusion cancer (74 adults, 24 children ≥2 yrs, and 28 infants

3610 Background: The highly selective TRK inhibitor larotrectinib is approved for the treatment of adult and pediatric cancers that harbor NTRK gene fusions; it achieves a 79% overall response rate (ORR) in this population (Hong et al., Lancet Oncol, 2020). The activity of larotrectinib in adults alone was further characterized in this update with a larger series of patients and more mature durability data. Methods: Adults (aged ≥18 y) with TRK fusion cancer treated in three larotrectinib clinical trials (NCT02122913, NCT02576431, and NCT02637687) were analyzed. Larotrectinib was administered 100 mg BID until disease progression, withdrawal, or unacceptable toxicity. ORR was investigator-assessed (RECIST v1.1). Compared to previously presented data on 74 patients, this expanded data set includes an additional 42 patients (data cutoff: July 15, 2019). Results: 116 adults (median age: 56 y, range 19–84 y; 53% female) with TRK fusion cancer were treated. Tumor types included thyroid cancer (22%), salivary gland cancer (19%), soft tissue sarcoma (16%), lung cancer (12%), colon cancer (7%), melanoma (5%), breast cancer (5%), GIST (3%), and 9 other types (≤2% each). NTRK fusions involved NTRK1 (43%), NTRK2 (3%), and NTRK3 (54%). 78% of patients had received prior systemic therapy (with 68% of those receiving ≥2 prior therapies). The ORR was 71% (95% CI 62–79): 10% complete response, 60% partial response (2% pending confirmation), 16% stable disease, 9% progressive disease, 3% not determined. In patients with brain metastases, the ORR was 71% (95% CI 42–92; 10 of 14 patients, all partial responses). Median duration of response for the overall data set (n = 116) was 35.2 mo (95% CI 21.6–not estimable [NE]). Median progression-free survival was 25.8 mo (95% CI 15.2–NE). Median overall survival was not reached (range 0.5+ to 51.6+ mo) at a median follow-up of 15.8 mo. Duration of treatment ranged from 0.10 to 51.6+ mo. 12% of patients had dose reductions. One patient (1%) discontinued due to a larotrectinib-related adverse event (AE). AEs were mostly grade 1–2; no new unexpected AEs were reported. Conclusions: In an expanded data set of adults with TRK fusion cancer, larotrectinib demonstrated robust and durable tumor-agnostic efficacy and favorable safety, supporting NTRK gene fusion testing in patients with solid tumors of any type. Clinical trial information: NTC02122913, NCT02576431, NCT02637687 .

177 Background: No current clinical tool can predict the efficacy of cancer therapeutics for patients with colorectal cancer (CRC). We recently demonstrated the feasibility of using patient-derived cancer organoids (PDCOs) to examine therapeutic response and tumor heterogeneity for individual patients with CRC via optical metabolic imaging (OMI). Here we expand these analyses in a cohort of patients with clinical outcomes. Methods: CRC tissue was collected from patients on IRB-approved protocols. PDCOs were matured and treated with chemotherapy regimens concurrent with patient treatment. Previously established effect size response thresholds were used for diameter ( > 1.5) and OMI ( > 0.5) following 48 hours of treatment. OMI measures the intrinsic autofluorescence of NAD(P)H and FAD using 2-photon microscopy without specific reagents or dyes. Clinical outcomes were prospectively collected by manual chart review. Results: 12 CRC PDCOs were established from patients with CRC. PDCOs were collected from initial diagnosis and advanced setting of both primary and metastatic sites by core needle biopsy and surgical resection. Differential growth rates were observed across lines. PDCOs with RAS/RAF alterations had more uniform growth, while PDCOs without these alterations demonstrated more heterogeneous growth and metabolism. Clinical correlation of PDCOs response with recurrence of disease in the adjuvant setting will be presented. Cases with prior 5-FU-based chemotherapy at the time of PDCO collection had intermediate sensitivity. For PDCOs collected pre-treatment, PDCO response predicted clinical response for 5 of 6 cases using predefined sensitivity thresholds. In the case that overall PDCO response did not predict clinical response, a heterogenous response was observed with distinct sensitive and resistant populations. Across PDCOs, greater post-treatment heterogeneity was observed in resistant lines compared to those with treatment sensitivity. Conclusions: Tumor heterogeneity in treatment response can be assessed using PDCOs growth and metabolism. The utility of PDCOs to predict clinical outcomes for patients with CRC deserves further prospective validation.

3122 Background: A broad range of pediatric and adult malignancies harbor TRK fusions involving the NTRK1, NTRK2, and NTRK3 genes. The highly-selective TRK inhibitor, larotrectinib, has previously shown a high overall response rate (ORR) and a favorable safety profile in patients (pts) with TRK fusion cancer. To better delineate efficacy in adults, as pediatric pts have a particularly high ORR, here we report updated efficacy and safety data from the adult subset of pts with TRK fusion cancer treated with larotrectinib. Methods: Adult pts (aged 18 or older) with TRK fusion cancer detected by local testing in 2 larotrectinib clinical trials (NCT02122913 and NCT02576431) were analyzed. Larotrectinib was administered 100 mg PO BID until disease progression, withdrawal, or unacceptable toxicity. Disease status was assessed by both investigator (INV) and independent assessment (IRC) using RECIST v1.1. Results: As of July 30, 2018, 83 adults (median age: 57 y, range 20–80 y) with TRK fusion cancer had been treated. Cancer types included salivary gland (23%) and thyroid cancer (19%), soft tissue sarcoma (14%), lung cancer (13%), colon cancer and melanoma (7% each), GIST (5%), and bone sarcoma, cholangiocarcinoma, and appendiceal, breast, and pancreas cancer (≤2% each). TRK fusions involved NTRK1 (40%), NTRK2 (2%), and NTRK3 (57%). 77% of pts had received prior systemic therapy (median lines: 2, range 0–10). In 74 pts evaluable per INV, the ORR was 76% with 9% CR, 57% confirmed PR, 9% PR pending confirmation, 12% SD, 11% PD, and 1% not determined; 9 pts were non-evaluable (NE) due to lack of post-baseline assessment. In 65 pts evaluable per IRC, the ORR was 68% with 17% CR, 51% PR, 15% SD, 12% PD, and 5% NE. With a median follow up of 17.2 and 17.5 mo per INV and IRC, respectively, the median duration of response had not been reached (ranges identical: 1.9+ to 38.7+ months). At data cutoff, 63% remained on treatment; 30% had discontinued due to disease progression. Adverse events were mostly grade 1–2. Conclusions: Larotrectinib demonstrated robust tumor-agnostic efficacy and a favorable safety profile in adult pts with TRK fusion cancer. These results support testing for TRK fusion cancer in pts with advanced solid tumors, regardless of site of primary diagnosis. Clinical trial information: NCT02122913 and NCT02576431.

2504 Background: Interleukin 1 receptor Accessory protein, IL1RAP, is expressed in several solid tumors, both on cancer cells and tumor-associated inflammatory cells. CAN04 is a first-in-class fully humanized monoclonal antibody targeting IL1RAP blocking IL-1 alpha and beta signaling and triggering antibody dependent cellular cytotoxicity. Methods: The primary objective was to assess safety and tolerability of weekly CAN04 in order to define the Recommended Phase 2 Dose (RP2D). Patients (pts) with relapsed or refractory non-small cell lung cancer (NSCLC), pancreatic ductal adenocarcinoma (PDAC), breast or colorectal cancer were included using a 3+3 dose escalation design. Key eligibility criteria were ECOG ≤1, normal organ function and no bleeding disorder/coagulopathy. Tumor responses were evaluated according to irRC every 8 weeks. PK and biomarkers were analyzed in serum. Results: 22 pts were enrolled across 5 cohorts (1-10 mg/kg). Demography: mean age 62 yrs (39-81); gender 14 M and 8 F; median number of prior lines of therapy 3 (range 1-11). AEs occurred mainly following the first dose and the most common AEs were: infusion related reaction (IRR) (41%), pyrexia (27%), fatigue (23%), chills (23%) and nausea (23%). AE grade 3 or 4: one IRR, one neutropenia/leukopenia and one hypokalemia, all of them grade 3. Serum CRP and IL-6 were reduced after two weeks of treatment. There were linear increases of AUC and Cmax (1-10 mg/kg) and CAN04 exposure at 10mg/kg was above levels associated with signs of efficacy in preclinical models. In pts receiving at least one dose of CAN04, 9/20 (45%) had SD by irRC (7/20 had SD by RECIST 1.1) at 8 weeks follow up. Two pts, one with NSCLC and one with PDAC, had SD for 6 and 4 months (latter still on therapy). Conclusions: CAN04 demonstrated a manageable safety profile and a RP2D of 10 mg/kg has been established. The dose expansion phase of the trial will evaluate CAN04 as monotherapy as well as in combination with relevant chemotherapy regimens in NSCLC and PDAC in separate arms. Clinical trial information: NCT03267316.

187 Background: BTCs are rare, aggressive malignancies with poor prognoses. Treatment options and outcomes after first-line therapy are not well defined. Median progression-free survival (PFS) in second-line BTC is < 5 mo. Combined BRAF + MEK inhibition is efficacious in BRAF V600–mutated anaplastic thyroid cancer, melanoma, and lung cancer, but less so in BRAF V600E-mutated colorectal cancer. Activating BRAF V600E mutations have been reported in 0% to 20% of BTCs; thus, D (BRAF inhibitor) + T (MEK inhibitor) was evaluated as a treatment for pts with BRAF V600E–mutated BTC in the ROAR basket trial. Methods: In this phase II, open-label trial, pts with BRAF V600E mutations in 9 rare tumor types, including BTC, received D (150 mg BID) + T (2 mg QD) until unacceptable toxicity, disease progression, or death. Eligible pts had advanced or metastatic cancer and had been treated with ≥ 1 prior systemic therapy. The primary endpoint was investigator-assessed overall response rate (ORR). Secondary endpoints included duration of response (DOR), PFS, overall survival (OS), biomarker correlates, and safety. Results: Thirty-three pts with BTC had enrolled at data cutoff (January 3, 2018). BRAF V600E mutations were centrally confirmed in 30 of 33 pts with BTC (91%). Median age was 58 y, and 78% had received ≥ 2 prior lines of systemic therapy. 32 of 33 pts with BTC were evaluable. Investigator-assessed ORR was 41% (13/32; 95% CI, 24%-59%), with 6 of 13 responses ongoing at data cutoff, 7 of 13 pts (54%) had a DOR ≥ 6 mo. Median PFS was 7.2 mo (95% CI, 4.6-10.1 mo), and median OS was 11.3 mo (95% CI, 7.3-17.6 mo). Grade 3/4 adverse events in ≥ 3 pts were increased γ-glutamyltransferase (n = 3 [9%]) and decreased white blood cell count (n = 3 pts [9%]). Biomarker analyses demonstrate a heterogeneous genetic landscape, and suggest a higher baseline expression of MAPK pathway genes in the pts who did not respond to D + T. Conclusions: D + T demonstrated promising efficacy in pts with BTC, with a favorable safety profile. These pts should be considered for BRAF mutation analysis, and D + T should be considered for pts with BRAF V600E-mutated BTC. Clinical trial information: NCT02034110.

2027Background: Bevacizumab in combination with chemotherapy has shown activity in recurrent glioblastoma patients. Patients who achieve response to bevacizumab have improved survival as well as qu...

e13077Background: In an ongoing, open-label Phase 2 study liposomal cisplatin (LiPlaCis) was evaluated in metastatic breast cancer (mBC) patients selected by a cisplatin-specific mRNA-based drug re...

e14086Background: Glioblastoma is the most malignant and lethal primary brain tumor. Standard therapy comprises maximal safe surgical resection, radiation therapy plus concomitant and adjuvant temo...

e15155Background: High-dose (HD) IL-2 is approved for patients with melanoma and renal cell carcinoma but is associated with significant toxicity, given only in an inpatient setting. RO6874281 is a...

PurposeTwo phase I studies were conducted of ABR-217620 alone or in combination with docetaxel. This is a recombinant fusion protein consisting of a mutated variant of the superantigen staphylococcal enterotoxin E (SEA/E-120) linked to fragment antigen binding moiety of a monoclonal antibody recognizing the tumor-associated antigen 5T4.Patients and MethodsPatients with non–small-cell lung cancer (NSCLC), pancreatic cancer (PC), and renal cell cancer (RCC) received 5 daily boluses of ABR-217620 (3-month cycles) in escalating doses to determine the maximum-tolerated dose (MTD; ABR-217620 dose escalation monotherapy [MONO] study). Doses were selected based on individual patient anti–SEA/E-120 titers pretreatment. Patients with NSCLC received 4 daily, escalating doses of ABR-217620 followed by docetaxel in 21-day cycles (ABR-217620 dose escalation combination with docetaxel [COMBO] study).ResultsThirty-nine patients were enrolled in the MONO study and 13 were enrolled in the COMBO study. The monotherapy MTD was 26 μg/kg (NSCLC and PC) and 15 μg/kg (RCC). Dose-limiting toxicities (DLTs) in the MONO study were fever, hypotension, acute liver toxicity, and vascular leak syndrome. In the COMBO study, the MTD was 22 μg/kg (neutropenic sepsis). Adverse events included grade 1 to 2 fever, hypotension, nausea, and chills. Treatment caused a systemic increase of inflammatory cytokines and selective expansion of SEA/E-120 reactive T-cells. Tumor biopsies demonstrated T-cell infiltration after therapy. Fourteen patients (36%) had stable disease (SD) on day 56 of the MONO study. Two patients (15%) in the COMBO study had partial responses, one in a patient with progressive disease on prior docetaxel, and five patients (38%) had SD on day 56.ConclusionABR-217620 was well tolerated with evidence of immunological activity and antitumor activity.

TPS2605 Background: HuMax-AXL-ADC is an antibody-drug conjugate (ADC) composed of an Axl-specific human monoclonal immunoglobulin G1 (IgG1κ) conjugated via a protease-cleavable valine-citrulline linker to the microtubule disrupting agent monomethyl auristatin E (MMAE). In vivo, HuMax-AXL-ADC demonstrated therapeutic anti-tumor efficacy in patient-derived xenograft models representing a variety of solid cancers, including pancreas, thyroid, lung, esophageal, cervical cancers and malignant melanoma. The non-clinical safety profile and pharmacokinetics (PK) of a once every 3 weeks (1Q3W) dosing schedule were established in cynomolgus monkeys. Methods: The primary objective of this trial is to determine the MTD and to establish the safety profile of HuMax-AXL-ADC in a mixed population of patients with specified solid tumors: ovarian, cervical, endometrial, thyroid cancer, NSCLC, and malignant melanoma. The trial consists of two parts, a phase I dose escalation part and a phase IIa expansion part. The dose escalation part explores two different dosing regimens: the first investigates doses from 0.3 up to 2.8 mg/kg to be administered 1Q3W. The second investigates doses in the range of 0.45 to 1.4 mg/kg to be administered weekly for 3 weeks followed by one treatment-free week (3Q4W dosing schedule). The second arm has a delayed start to inform a safe starting dose: when at least 8 patients have been evaluated for dose limiting toxicities, the 1.5 mg/kg cohort of the 1Q3W arm has been declared safe, and the predicted PK parameters of the starting dose in the 3Q4W arm are below pre-defined limits, the 3Q4W arm will be initiated. The 1Q3W arm follows a modified Bayesian Continuous Reassessment Method including escalation with overdose control in up to 41 patients on up to 7 main and 4 intermediate dose levels while the 3Q4W arm is run as a standard 3+3 trial design on up to 5-6 dose levels. In the phase IIa expansion part, further safety and biological activity data will be generated in selected indications using cohorts of 22 patients (11+11 patients in each cohort applying the Simon’s two-stage design). Clinical trial information: NCT 02988817.

6024 Background: ACCs have high levels of Notch-1 receptor expression and activation. LY3039478 (LY) is an orally bioavailable selective Notch inhibitor (Notch 1-4). Here we report on safety, pharmacokinetics (PK), pharmacodynamics (PD), and anti-tumor activity of LY in patients (pts) with ACC. Methods: Ongoing, multi-part, phase I trial enrolled pts with advanced or metastatic ACC, measurable disease, ECOG ≤1, and baseline tumor tissue. Eligible pts received LY 50 mg three times per week (TIW), on a 28-day cycle until disease progression. Safety assessments were based on CTCAE V4.0. Tumor responses were assessed using RECIST 1.1. Primary objectives are to confirm the recommended phase II dose of LY and document antitumor activity. Secondary objectives are safety and toxicity, PK and progression-free survival (PFS). Exploratory objectives include assessment for PET metabolic responses. Results: 22 pts have been enrolled and received LY 50mg TIW (13 men, 9 women; median age 60, range 41-82). All pts had metastatic disease; median treatment duration was 3 cycles (range 1-10) with 6 pts continuing on treatment. One pt had an unconfirmed partial response. Disease control rate (DCR) was 16/22 (73%), of which 4 pts had stable disease ≥ 6 months. In the overall group (n = 22) median PFS (mPFS) was 5.3 months (95% CI: 2.4, NE). mPFS was 7.7 (95% CI: 4.0, NE) for pts in second line (n = 7), while mPFS was 2.4 (95% CI: 1.1, NE) for pts in third line or more (n = 9). In pts without prior systemic therapy (n = 6) mPFS could not be estimated since 4 of those patients were censored. In preliminary analysis, 14 pts were assessed by PET, with 2 (14%) achieving partial metabolic response. Most frequent related adverse events (all grades) occurring in ≥20% of pts included diarrhea (55%), fatigue (45%), vomiting (36%), decreased appetite (27%), dry mouth (27%), and dry skin (23%). Grade 3/4 related treatment-emergent adverse events observed in more than one pt were diarrhea (n = 3) and squamous cell carcinoma of skin (n = 2). PK was assessed in 17 pts, with peak concentrations occurring approximately 2 hours post-dose. Biomarker and histologic analyses of pre and post treatment biopsies will be presented. Conclusions: LY showed activity (73% DCR) in ACC with a manageable safety profile. Clinical trial information: NCT01695005.

LBA2501 Background: Larotrectinib is the first selective small-molecule pan-TRK inhibitor. TRK fusions appear oncogenic independent of tumor lineage, are widely distributed across cancers, and affect all ages. We present an integrated dataset from 3 studies intended to support regulatory approval. Methods: All NTRK fusion pts with RECIST measurable disease enrolled to the adult (NCT02122913, n=8) and pediatric (NCT02637687, n=12) phase I trials and adult/adolescent phase 2 trial (NCT02576431, n=35) were analyzed. TRK fusion status was determined by local testing prior to enrollment. Pts were dosed predominantly at 100mg BID on a continuous 28-day schedule. Primary objective was investigator-assessed overall response rate (ORR) per RECIST v1.1. Secondary endpoints included duration of response (DOR) and safety. Data were cut on 31-JAN-2017. Results: 55 TRK fusion pts (12 peds, 43 adult, range: 4 mo.-76 yrs) were enrolled (median priors=2). Fusions involved NTRK1 (n=25), NTRK2 (n=1), and NTRK3 (n=29), and 14 unique partners. 13 discrete tumor types were treated: salivary (12), sarcoma (10), infantile fibrosarcoma (7), lung (5), thyroid (5), colon (4), melanoma (4), cholangio (2), GIST (2), and other (4). For the 46 pts evaluated to date, the ORR was 78% (95% CI: 64%–89%) with responses in 12 unique tumor types. Responses are ongoing in 29/33 (88%) pts, excluding 3 peds pts whose DOR was censored at attempted curative resection. A median DOR has not been reached as the majority of responders remain on treatment without progression. The longest responder remains on treatment at 23 mos., 8 pts remain in response at >12 mos., and 16 pts at >6 mos. NTRK solvent front mutations were detected in all 4 pts to develop acquired resistance. The most common TEAEs were fatigue (30%), dizziness (28%), and nausea (28%). 5 (11%) pts required dose reductions. Conclusions: Larotrectinib has demonstrated consistent and durable antitumor activity in TRK fusion cancers, across a wide range of ages and tumor types, and was well-tolerated. Larotrectinib could be the first targeted therapy developed in a tissue type-agnostic manner, and the first developed simultaneously in adults and pediatrics. Clinical trial information: NCT02576431, NCT02122913, NCT02637687.

85 Background: While phase I trials are essential for the development of new anticancer drugs, there is a limited chance of benefitting for cancer patients participating in such trials. The information dialogue is therefore of substantial importance for providing a foundation to make a decision, and the support from relatives of potential value for the patient. This systematic review investigated patients’ prerequisites for deciding to participate in a phase I trial by summarizing the existing knowledge regarding patients’ decision-making when entering a phase I trial, and patients’ and their relatives’ perception of the information prior to enrollment. Methods: The review is based on the principles of Preferred Reporting Items for Systematic Reviews. A comprehensive systematic search was performed using the PubMed, Embase and PsycInfo databases and supplemented by a search for unpublished literature. Results: We identified 36 studies for inclusion in this review. When patients are offered participation in a phase I trial, information procedures as well as the patients’ individual approach influence the decision-making and the perception of the information provided. Across the studies exploring patients’ perception of information, there was a limited understanding of trial purpose and unrealistic expectations of benefit. The relatives’ perception of information remains unexplored. Evaluation of the included studies demonstrated a comprehensive risk of bias in the majority of studies. Conclusions: The information dialogue between physician and the patient concerning participation in a phase I trial seems to benefit from exact information taking account of the perspectives for each individual patient as well as the need for further discussion of trial. While relatives intuitively function as resources for patients entering a phase I trial, this topic is still not investigated.