From Morgan Stanley....
Imminent lung readouts
BMY should report Opdivo combo Ph 3 first-line lung cancerIf trials succeed, we will likely have to await efficacy details at ESMO Sept. 26–
overall survival (OS) top-line results in mid-2019. Consensus
has low expectations for Part 1 (+ Yervoy) OS in PD-L1+ due
to disappointing PFS & OS in TMB-high. Consensus expects
Part 2 (+ chemo) to succeed but not match Merck’s KN- 189.
Oct. 1. For positive trials, Bristol usually only announces primary endpoint
achievement in the initial release. It subsequently discloses efficacy/safety details
at a medical conference. The next major oncology conference is ESMO in
Barcelona. If trials fail to achieve primary endpoint(s) but there is a positive
trend, Bristol may elect to highlight key efficacy metrics in its initial release.
Expectations are low for CM-227 Part 1 (Opdivo + Yervoy combo) OS in PD-L1+
patients; this trial is less important than Part 2. OS in PD-L1+ is a co-primary
endpoint of CM-227 Part 1. The other co-primary endpoint was PFS in TMB-high,
in which the combo showed positive but disappointing results (see detailed
discussion). Two other reasons most observers are cautious are that Opdivo
monotherapy failed to show an OS benefit in PD-L1+ in CM-026 and Yervoy has
not shown compelling efficacy as monotherapy in NSCLC. But if Part 1 shows an
OS benefit, we would expect BMY to file for approval of this chemo-free
CM-227 Part 2 is the more important readout; it is expected to succeed but not
match Merck's Keytruda KN-189 trial. Part 2 compares Opdivo + chemo vs.
chemo. It enrolled both non-squamous and squamous NSCLC patients, and
investors will compare the non-squamous cohort cross-trial to Keytruda KN-189
results. Bristol management has stated that it does not expect Part 2 to be able
to match KN-189's very impressive OS hazard ratio because Part 2's chemo
control group may perform better. Bristol has suggested that investors instead
focus on comparing landmark survival analysis at one year cross-trial.
Potential for CM-227 Part 2 positive/negative surprise. If Part 2 succeeds and
the non-squamous hazard ratio is surprisingly similar to KN-189, it could
potentially be a meaningful positive for BMY shares. Perception of Opdivo would
likely improve, and Opdivo lung cancer sales prospects would rise. Part 2 failing
would likely be a major negative for BMY shares. Although consensus does not
model significant Opdivo sales in 1L lung cancer, debate regarding whether
Opdivo may be inferior to Keytruda could impact Opdivo adoption in other
Opdivo + Yervoy combo showed relatively disappointing results in TMB-high patients.
CM-227 Part 1 showed PFS hazard ratio (HR) of 0.58 and OS HR of 0.77 (23 v. 17
months) for TMB-high patients, both of which are numerically worse than Keytruda HRs
in KN-024 and KN-189. Additionally, OS HR was 0.78 (16 v. 12 months) for TMB-low,
which raised questions about TMB as a potential biomarker for the combo regimen.
Investors expect CM-227 Part 2 to succeed but show non-squamous subgroup results
that are numerically inferior to KN-189 results cross-trial. Part 2 is comparing Opdivo +
chemo vs. chemo in squamous and non-squamous NSCLC while KN-189 was non-
squamous only. BMY management has downplayed HR (on the basis that chemo arm in
KN-189 underperformed with respect to historical trials) and suggested that investors
focus on landmark analysis for both Part 2 and CM-9LA when comparing to KN-189. For
comparison, KN-189 showed PFS of 34% for Keytruda vs. 17% for control and OS of 69%
vs. 49% at one year, but we don't know what landmark BMY will highlight.
Keytruda 1L lung survival results appear hard to beat. In non-
squamous NSCLC (KN-189), PFS HR was 0.52 (8.8 vs. 4.9 months) and OS HR was 0.49
(NR vs. 11.3 months). In
PD-L1+ ≥ 50% (KN-024), PFS HR was 0.50 (10.3 vs. 6.0 months) and OS HR was 0.60
(NR vs. NR). In PD-L1+ ≥ 1% (KN-042), PFS HR was 1.07 (5.4 vs. 6.5 months) and OS HR
was 0.81 (16.7 vs. 12.1 months).
Opdivo monotherapy one-year landmark survival analysis is numerically inferior to
that of Keytruda cross-trial (Exhibit 1), and we think CM-227 Part 2 will have to do
better in order for Opdivo to be competitive. We compared results from the 1L PD-L1+
NSCLC IO monotherapy trials Keynote-042 and CheckMate-026. The OS hazard ratios
(HRs) were 0.81 for Keytruda v. chemo and 1.07 for Opdivo v. chemo. The one-year OS
rates were 60% v. 50% for Keytruda v. chemo and 56% v. 54% for Opdivo v. chemo. So
cross-trial-wise, Keytruda was better than Opdivo on an absolute basis (by 4%) as well
as control-adjusted basis (by 8%), but Opdivo does appear more competitive based on