01/21/2018: Articles Summaries - GI & Hematological Cancers

View the most clinically relevant journal articles of the week. Updated each Sunday!

Please click on the tumour site to jump to summary


GI CANCERS Nivolumab in Patients With Advanced Gastric or Gastro-Oesophageal Junction Cancer Refractory to, or Intolerant of, at Least Two Previous Chemotherapy Regimens (ONO-4538-12, ATTRACTION-2): a Randomised, Double-Blind, Placebo-Controlled, Phase 3 Trial
Kang, YK et al. Lancet. 2017 Dec 2;390(10111):2461-2471.
PubMed ID: 28993052


Immune checkpoint inhibitors have proven anti-tumor activity in several solid malignancies such as breast and lung cancer, as revealed in several trials. In gastric cancer, PD-L1 expression is associated with depth of invasion, tumor size, lymph node metastasis and shorter median survival. Because advanced gastric and gastro-oesophageal cancers are characterized by a high mutational burden and overexpression of immune checkpoint proteins, the use of an immune checkpoint inhibitor like nivolumab is a potential therapy to target this disease.

Based on preliminary data from the phase I/II CHECKMATE 032 trial on anti-PD-1 therapy in advanced gastric and gastro-oesophageal junction cancer, investigators further examined the efficacy and safety of nivolumab in a heavily pretreated population unselected for PD-L1 tumor expression.



  • Patients: Histologically confirmed advanced unresectable or recurrent gastric or gastro-oesophageal junction adocarcinoma refractory to, or intolerant of, standard therapy, ECOG 0-1; patients with symptomatic brain metastases or requiring treatment were ineligible.
  • Study design: phase III, double-blind placebo-controlled RCT
  • Intervention
    • Experimental: Nivolumab 3mg/kg IV q2 wks. (6 wks.=1 cycle)
    • Control: Placebo
  • Endpoints
    • Primary: OS
    • Secondary: ORR, DCR, DOR, best overall response, maximum percentage change from baseline in the sum of diameters of target lesions, PFS, safety



  • 493 patients randomized 2:1 to nivolumab (n=330) and placebo (n=163)
  • Baseline patient characteristics: 71% male, median age 61 yrs., 46% Japanese, 45% Korean, 71% ECOG 1, 74% with ≥ 2 organs with metastases, 86% lymph node metastases, 22% peritoneal metastases, 41% ≥ 4 previous treatment regimens, 62% previous gastrectomy
  • Previous therapies: pyrimidine analogues (100%), platinum (95%), taxane (86%), irinotecan (75%), ramuirumab (12%)
  • Subsequent anticancer therapy: 47% vs 44.2% 

Table 1: Efficacy Outcomes

  • Safety
    • Serious treatment-related adverse events with nivolumab (≥ 2 patients): interstitial lung disease (n=3) and colitis, pyrexia, pneumonia, urinary tract infection, diabetic ketoacidosis (n=2 each)
    • Most common any grade adverse events: pruritus (9% vs 6%), diarrhea (7% vs 2%), rash (7% vs 3%), fatigue (5% vs 6%)
    • Adverse events (any grade) of special interest: interstitial lung disease (2% vs 0), maculopapular rash (1% vs 1%), colitis (1% vs 0), hyperthyroidism (1% vs 0)

Table 2: Adverse events and treatment summary



  • In heavily pretreated advanced gastric and gastro-oesophageal junction cancer, nivolumab provided marginal improvement in OS compared to placebo. Nivolumab improved OS by 1.1 mos. and the estimated 18-mo. OS was 3x longer than placebo. The risk of disease progression or death was reduced by 40% with nivolumab and at 6 mos., 20% were free from disease progression.
  • In a posthoc analysis, patients on nivolumab derived almost similar benefit regardless of PD-L1 status. However, only 26 patients were PD-L1 positive. Hence, these results should be interpreted with caution. Patients previously treated with ramucirumab (11%) likewise had better clinical outcomes compared to placebo.
  • Partial response was achieved in at least 11% on nivolumab and tumor responses were durable whereas none of the patients on placebo achieved PR.
  • The safety profile of nivolumab was similar to that previously reported in the literature. Adverse events were more frequently observed in the nivolumab arm, particularly pruritus, diarrhea and rash. Immune-related adverse events occurred in <3% of patients on nivolumab and were mostly low grade.
  • Patients were not preselected based on PD-L1 status and only 40% had archival tissue samples available for testing. Hence, an association between nivolumab and PD-L1 expression was not established in this trial, unlike studies in NSCLC which revealed greater magnitude of benefit with nivolumab in tumors with higher PD-L1 expression. QOL, which is an important outcome especially in heavily treated patients was not assessed in the study.
  • This is the first phase III trial to demonstrate improvement in survival for patients with advanced gastric and gastro-esophageal junction cancer previously treated with 2 or more chemotherapy regimens. While there was a 37% relative reduction in the risk of death the absolute improvement in survival was a modest 1.1 months. While nivolumab treatment can be considered in this patient population, further clinical trials of immune checkpoint inhibitors in non-Asian advanced gastric cancer patients should be considered.

More GI Cancer Updates







Bosutinib Versus Imatinib for Newly Diagnosed Chronic Myeloid Leukemia: Results From the Randomized BFORE Trial
Cortes JE et al., J Clin Oncol. 2018 Jan 20;36(3):231-237.
PubMed ID: 29091516


Bosutinib is a 2nd generation dual SRC/ABL kinase inhibitor with efficacy in Philadelphia (Ph) chromosome-positive chronic myeloid leukemia (CML) and is currently approved as 2nd or subsequent line therapy. In a study that compared bosutinib with imatinib in newly diagnosed Ph chromosome-positive chronic phase CML (CP CML), no difference was observed in complete cytogenetic response (CCyR) at 12 mos., despite improvement in major molecular response (MMR) at 12 mos. and a lower rate of transformation to accelerated phase (AP) or blast phase (BP) with bosutinib (J Clin Oncol 30:3486-3492, 2012). These results prompted Cortes and colleagues to further investigate the efficacy and safety of bosutinib at a lower initial dose (400 mg daily) compared with imatinib in a similar population.



  • Patients: CP CML diagnosed within 6 mos. prior to the study, Ph positive or Ph-negative/BCR-ABL1–positive, Eastern Cooperative Oncology Group (ECOG) performance status 0-1
    Ph-positive patients with typical BCR-ABL1 transcript types (e13a2 and/or e14a2); Ph-negative patients (ie, 0 out of ≥10 to 99 metaphases at baseline) and those with unknown Ph status and/or atypical transcript type were excluded from this population but included in the safety analyses and analyses of the ITT population of all patients randomly assigned.
  • Study design: Phase III, double-blind RCT
  • Intervention:
    • Experimental: Bosutinib 400 mg PO daily
    • Control: Imatinib 400 mg PO daily
  • Endpoints:
    • Primary: MMR at 12 mos.
    • Secondary: CCyR by 12 mos., MMR by 18 mos., duration of CCyR and MMR, event-free survival (EFS), overall survival (OS), safety



  • Treatment with bosutinib provided significantly better rate of MMR at 12 mos. compared to imatinib (47.2% v 36.9%; P=0.200 [mITT]). This advantage with bosutinib was consistently observed at 3, 6 and 9 mos., with a shorter time to response compared to imatinib.
  • Deeper molecular responses were achieved with bosutinib compared to imatinib at 3, 6, 9 and 12 mos. More patients on bosutinib achieved BCR-ABL1 transcripts ≤ 10% at 3 mos. Furthermore, CCyR rate at 12 mos. was significantly higher with bosutinib (77.2% v 66.4%; P = 0.0075).
  • Dose escalation due to suboptimal response was less common in the bosutinib arm (17.2% v 27.5%). CCyR was achieved by 7.11% on bosutinib v 15.8% on imatinib. The corresponding MMR rates were 3.4% and 10.6%, respectively.
  • Disease progression to AP or BP occurred in 1.6% (bosutinib) v 2.5% (imatinib). Five patients (bosutinib n=3, imatinib n=2) continued to receive treatment during AP and four patients achieved MMR. 12-mo. EFS (cumulative incidence) was 2x higher with bosutinib compared to imatinib (3.7% v 6.4%). 12-mo. OS was 99.6% v 97.9%, respectively.
  • Diarrhea and liver function abnormalities were the most common side effects with bosutinib, although diarrhea was transient and mostly low grade. Dose reductions (34.7% v 17.%), dose interruptions (56.3% v 35.8%) and treatment discontinuation (14.2% v 10.6%) due to adverse events were more frequent with bosutinib. The most common grade 3 or higher adverse events with bosutinib were elevated ALT (19%), thrombocytopenia (13.8%), elevated AST and lipase (9.7% each) and diarrhea (7.8%). Grade ≥ 3 liver function events were 6x higher with bosutinib (24.3% v 4.2%) and were the most common cause of treatment discontinuation.
  • Cardiac events (5.2% v 5.3%) were mostly low grade; grade 3 or higher events were reported in 0.7% and 0.4%, respectively. Peripheral vascular events were reported in <2% of patients (1.5% v 1.1%) while cardiovascular events occurred in 3% v 0.4%. (2.2% v 0.4% in the first year). One patient on imatinib died from a cerebrovascular event



  • Bosutinib demonstrated superiority as first line treatment over imatinib in Ph-positive CP CML. Deeper and earlier molecular response was achieved with bosutiinib, with higher rates of molecular and cytogenetic responses to a greater magnitude in those with higher Sokal risk score. The use of a lower bosutinib starting dose appeared to provide a more favorable safety profile. Cardiac, peripheral vascular and cardiovascular events were comparable in both treatment arms.
  • Study authors used modified ITT population for the primary endpoint (MMR at 12 mos.) that consisted of Ph-positive patients with typical BCR-ABL1 transcript
    types (e13a2 and/or e14a2). Hence, these findings are limited to this population.

More Hematological Cancer Updates