Previously, the Phase III, open-label, randomised REFCLECT (NCT01761266) study showed non-inferiority for levantinib -a multikinase inhibitor of VEGFR 1–3, FGFR 1–4, PDGFRα, RET, and KIT- when compared to sorafenib in the first-line treatment of 954 unresectable HCC patients.

The study included a biomarker analysis, assaying VEGF, ANG2, FGF19, FGF21, and FGF23 by ELISA in serum samples (N=407) collected at baseline and during treatment and correlated these (Wilcoxon rank-sum test) with tumour response (mRECIST) by IRC.²Finn RS, et al. Ann Oncol 2018;29(8):mdy269.057

Both levantinib and sorafenib increased VEGF levels. However, FGF19 and FGF23 levels were only increased in patients treated with levantinib and was more profound in patients with a complete response or a partial response (PR) than in non-responders; FGF19: 55.2% vs 18.3%, P=0.0140; FGF23: 48.4% vs 16.4%; P=0.0022, respectively. Higher baseline-levels of VEGF, ANG2, and FGF21 correlated with a worse OS outcome for both sorafenib and levantinib, though patients with increased baseline FGF21 receiving levantinib (N=70) had longer OS compared to those patients receiving sorafenib (N=27); median OS 10.9 vs 6.8 months, respectively (HR=0.528; 95% CI, 0.328–0.849; P=0.0075).

Another presentation reported the patient-reported outcome (PRO) of second-line treatment with ramucirumab following sorafenib treatment in patients with advanced HCC from the Phase III REACH (NCT01140347) and REACH-2 studies (NCT02435433).³Zhu AX, et al. Ann Oncol 2018;29(8):mdy282.006 Patients in the REACH-2 were selected based on alpha-fetoprotein levels (AFP) ≥400 ng/mL, the study met its primary OS endpoint, consistent with what was observed in patients from the REACH-study who had AFP-levels ≥400 ng/mL .

Zhu and colleagues assessed the PRO of ramucirumab vs placebo in REACH-2 patients and in REACH patients with an AFP≥400 ng/mL by Functional Assessment of Cancer Therapy (FACT) and Hepatobiliary Symptom Index (FHSI)-8 and concluded that ramucirumab-treatment lead to a consistent trend in benefit for disease-related symptoms.

Biliary-Tract Cancer

Several Phase II studies reported the outcome of novel therapies for biliary-tract cancer (BTC). One open-label Phase II study investigated pan-fibroblast growth factor receptor (FGFR) kinase inhibitor infigratinib (BGJ398), a selective pan-FGFR kinase inhibitor, in 71 patients with previously treated advanced, intrahepatic cholangiocarcinoma (IHC) harbouring FGFR2 fusions or translocations (NCT02150967).⁴Javle M, et al. Ann Oncol 2018;29(8):mdy424.030 The investigators found that infigratinib had clinically meaningful activity after chemotherapy in patients with IHC with FGFR2 fusions and that the associated toxicity was manageable.

At a median duration of treatment of 5.5 months the investigator-assessed ORR by RECIST 1.1 was 31.0% (95% CI, 20.5–43.1%) and the confirmed ORR was 26.9% (95% CI, 16.8–39.1%). The confirmed ORR in patients (n=28) who had received ≤1 prior lines of treatment was 39.3%, compared to 17.9% in patients (n=39) who had received ≥2 lines of prior treatment. The disease control rate DCR was 83.6% (95% CI, 72.5–91.5%) and the median duration of response measured 5.4 months (95% CI, 3.7–7.4) with a median PFS of 6.8 months (95% CI, 5.3–7.6). The median OS was 12.5 months (95% CI, 9.9–16.6). Most common Grade 3/4 treatment emergent adverse events (TEAEs) in 47 pts (66.2%) were hypophosphatemia (14.1%), hyperphosphatemia (12.7%), and hyponatremia (11.3%).

Preliminary results of another open-label multicentre Phase IIa study (LUC2001, NCT02699606) in 11 out of 150 Asian patients with advanced cholangiocarcinoma (CCA) harbouring FGFR 2 or 3 alterations evaluated the efficacy and safety of erdafitinib after failing first-line treatment.⁵Chen YY, et al. Ann Oncol 2018;29(8):mdy282.008 In this study, erdafitinib showed encouraging clinical activity with 3 confirmed PR, 2 unconfirmed PR (uPR), 4 stable disease (SD), and 2 progression disease (PD, [both FGFR3 mutations]). The unconfirmed ORR measured 45.5% and the unconfirmed DCR 81.8%. Common adverse events (AEs) in >30% of patients included are hyperphosphatemia, dry mouth, stomatitis, diarrhea, nail disorder, and palmar-plantar erythrodysaesthesia syndrome. Grade ≥3 AEs occured in 7 patients, 3 subjects required dose reductions, but treatment-discontinuation or Grade 5 toxicities were not reported. Non-drug related serious AEs (SAEs) occurred in 3 patients.

The multicohort, Phase 2 KEYNOTE-158 (NCT02628067) study investigated pembrolizumab in 104 patients with advanced BTC and prior progression/intolerance on standard chemotherapy, measurable disease per RECIST 1.1, ECOG PS ≤1, and sufficient tumour tissues sample available for biomarker evaluation including programmed death ligand 1 (PD-L1) by the immunohistochemistry 22C3 antibody.⁶Ueno M, et al. Ann Oncol 2018;29(8):mdy282.009 In this study, pembrolizumab showed durable antitumour activity in a subset of advanced BTC patients regardless of the PD-L1 expression by combined positive score (CPS) and with a manageable toxicity-profile.

At median follow-up of 9.3 months (range, 0.6–23.6), 6 partial responses were observed matching an ORR of 5.8% (95% CI, 2.1–12.1). Patients with a CPS ≥1 (N=61) had an ORR of 6.6% (95% CI, 1.8–15.9) compared to 2.9% (95% CI, 0.1–15.3) in patients with a PD-L1 CPS <1 (N=34). Seventeen patients (16%) had SD. The median DOR was not yet reached (range, 6.2–15.7+ months). The median PFS was 2.0 months (95% CI, 1.9–2.1); 1.9 months (95% CI, 1.8–2.0) in CPS ≥1 patients vs 2.1 mo (95% CI, 1.9–2.6) in CPS <1 patients. Likewise, a CPS ≥1 did not correlate with an improved OS outcome. The median OS in all patients was 9.1 months (95% CI, 5.6–10.4), and measured 7.2 mo (95% CI, 5.3–11.0) and 9.6 mo (95% CI, 5.4–12.8) in patients with a CPS ≥1 and CPS <1, respectively. No patients were found to be microsattelite instability-high (MSI-H). Grade ≥3 AEs occurred in 13% of patients, 16% had immune-mediated AEs (irAEs). Six patients discontinued due to treatment-related AEs.

Finally, results were presented of a nationwide Japanese cancer genome screening project, the SCRUM Japan GISCREEN, identifying BTC cancer genome alterations by utilising Next Generation Sequencing in 167 patients scheduled for -or undergoing- chemotherapy for advanced BTC at 20 major cancer centres (UMIN000016344).⁷Morizane C, et al. Ann Oncol 2018;29(8):mdy282.007

In total 140 samples were analysed by the cut-off date of March 31, 2017. A successful sequence was performed in 92 (65.7%) of tumour samples (36 IHC, 35 extrahepatic bile duct (EHBD), 14 gallbladder (GB), and 7 ampulla of Vater (VP)). No gene fusions were detected. Frequently detected gene-alterations are listed in Table 3.