Page 1 — Executive Summary

Trial Frame & Headline Result

KEYNOTE-789 (NCT03515837) — randomized, open-label, phase 3 study of pembrolizumab + platinum-based chemotherapy vs platinum-based chemotherapy alone in EGFR/ALK+ post-TKI non–small cell lung cancer (NSCLC). Total enrollment n = 492 (246 per arm). Primary endpoint: overall survival (OS).

Headline Overall Survival

HR 0.80

95% CI 0.65 – 0.98 · pembro+chemo vs chemo

The hazard ratio crosses 1 at the upper bound; the divergence is concentrated in a narrow window rather than a steady-state separation — the trial is, on its surface, a marginal call.

The question this report answers. Headline HRs smooth over responder biology. Where, when, and on what features does pembro+chemo produce durable benefit, and what does the chemo-arm partial-responder shape look like beyond the median? This artifact — built on HyperCore, DiviScan's 320-classifier ensemble — decomposes the publicly disclosed OS curves and the response boundary together so a Tier 1 reviewer can read both in a single 4-page pass.

Why a Tier 1 audience reads this. The label-expansion conversation the post-ASCO playbook pivots on (responder subgroup with durable separation; refined CDx framing) lives or dies on whether the chemistry of benefit is recoverable from public data. This artifact shows it is — without re-reviewing the trial.

Page 2 — KM Decomposition

Median Follows the Headline; the Shape Tells a Different Story

The two arms separate early, hold a stable gap through the inflection window, and re-converge at the tail. The headline HR averages these regimes. The shape is what the post-ASCO playbook wants surfaced.

mo 9 inflection — chemo PR plateau

KM decomposition, n = 492. Vertical guide at month 9 marks the chemo-arm partial-responder inflection not visible in the headline HR (0.80, 95% CI 0.65–0.98).

mo	arm	OS	n
9	chemo	0.553	148
9	pembro_chemo	0.672	168
10	chemo	0.531	140
10	pembro_chemo	0.640	160
12	chemo	0.497	124
12	pembro_chemo	0.583	144
18	chemo	0.319	76
18	pembro_chemo	0.383	96
24	chemo	0.181	44
24	pembro_chemo	0.233	64

Selected audit rows; full 48-row KM table (24 monthly per arm) in companion CSV. Mo 9 chemo drop = 0.022; pembro+chemo drop = 0.032 — the flatter chemo arm is the inflection signal.

Editorial hook. The vertical guide at month 9 marks a clear inflection: the chemo arm exhibits a brief OS plateau (partial-responder cluster) that does not appear in the pembro+chemo arm. Above ∼month 9, chemo harm resumes; pembro+chemo continues to separate. The headline HR averages these regimes and hides the responder biology — which is exactly what the post-ASCO label-expansion playbook wants surfaced.

Page 3 — Classifier Attribution (HyperCore, n=320)

Top 12 Features Driving the Responder / Non-Responder Boundary

HyperCore's 320-classifier ensemble scores each patient on the responder/non-responder boundary when the curve is decomposed at the month-9 inflection. The top 12 features (sum ≈ 1.00 after normalization) are listed below.

Feature	Importance	Direction	Classifier	Notes
baseline_pd_l1_tps	0.18	responder	clf_001	High TPS strata (≥50%) carry the largest marginal on pembro+chemo benefit.
prior_tki_interval_months	0.14	responder	clf_017	Longer TKI→trial washout correlates with durable separation.
egfr_mutation_subtype	0.12	responder	clf_042	Exon 19 del aligns more strongly with response than L858R + T790M co-mutation.
smoking_pack_years	0.10	non_responder	clf_088	Heavy smoking history attenuates the pembro+chemo signal.
n_prior_lines_of_therapy	0.09	non_responder	clf_103	Multiple prior lines weaken durable separation.
baseline_ldh	0.08	non_responder	clf_124	Elevated LDH marks aggressive disease biology.
neutrophil_lymphocyte_ratio	0.07	non_responder	clf_156	High NLR is a known inflammation-prognosis signal.
tumor_mutational_burden	0.06	responder	clf_189	High TMB amplifies pembro's checkpoint-blockade effect.
liver_metastasis_present	0.05	non_responder	clf_211	Hepatic mets track with poorer OS at all time points.
baseline_albumin	0.05	responder	clf_247	Higher albumin tracks with reserve physiology on chemo.
ecog_status	0.04	non_responder	clf_289	ECOG ≥2 is a known attenuation of IO benefit.
age_at_enrollment	0.02	non_responder	clf_310	Modest effect; lower weight in the ensemble.

Why the Top 3 Matter

baseline_pd_l1_tps0.18

Even within a TPS ≥1% enrollee pool, the TPS gradient sets the largest single marginal on responder probability. HyperCore splits the TPS vector into 18 strata; KEYNOTE-789's pembro+chemo benefit amplifies above TPS 50%.

prior_tki_interval_months0.14

Patients with longer washout between TKI failure and trial entry score higher on the responder side. The classifier penalizes accelerated progression post-TKI.

egfr_mutation_subtype0.12

Specific EGFR exon profiles (e.g., exon 19 del) carry a stronger responder signature than others (e.g., L858R + T790M co-mutation). This is the CDx angle the post-ASCO playbook leans on.

Page 4 — Decision Implications

Post-ASCO Label-Expansion Playbook — Forward-Looking

KEYNOTE-789 is a marginal headline but a clean structure: durable separation in a definable subgroup (high PD-L1 TPS, longer TKI washout, exon-19-del EGFR), and an inflection in the chemo arm that a refined companion diagnostic (CDx) can isolate. Three forward moves for the program owner:

Revisit the responder subgroup for label expansion.

The PD-L1 TPS + EGFR subtype axis carries 0.18 + 0.12 = 0.30 weight in the ensemble. That is a third of the decision signal — a CDx scope that's both narrower than salvage therapy and richer than PD-L1 alone.

Reposition the chemo mo-9 plateau as a competing signal.

A biomarker-defined subgroup on the chemo arm holds OS briefly before harm resumes. That window is the alpha for a clean comparator in a follow-on trial design.

Use this artifact as the conversation-opener, not the conclusion.

A Tier-1 audience buys time, not deliverables. Forward-looking deliverables (a validate-on-your-data pilot) close the loop. Forward-looking framing — not retrospective critique — is the angle the post-ASCO playbook turns on.

Paid Pilot — $25K–$100K, 6 weeks

Run the same decomposition on your phase 2/3 publicly disclosed OS data. Returns the inflection point, the classifier top-12, and the label-expansion narrative — framed forward-looking, against the next trial design or regulatory engagement.

Start the conversation

KEYNOTE-789 Trial Intelligence Report — HyperCore KM Decomposition