A REAL-WORLD EVIDENCE STUDY OF EPIGENOMIC SUBTYPING IN LIQUID BIOPSY-BASED LUNG CANCER COHORTS
Author(s)
Jayati (Joy) Saha, MPH, PhD1, Sheila R. Solomon2, Shaun Forbes, PhD3, Nicole Zhang, MPH4;
1Guardant Health, Wilmette, IL, USA, 2Redwood City, CA, USA, 3Guardant Health, Redwood City, CA, USA, 4Guardant Health, Palo Alto, CA, USA
1Guardant Health, Wilmette, IL, USA, 2Redwood City, CA, USA, 3Guardant Health, Redwood City, CA, USA, 4Guardant Health, Palo Alto, CA, USA
OBJECTIVES: Histological subtyping guides lung cancer therapy, but small biopsy samples, tumor heterogeneity, and ambiguous morphology often complicate accurate classification and delay treatment. The plasma-based Molecular Lung Subtype Predictor (MLSP; Guardant360 Liquid/Infinity) classifies adenocarcinoma (LUAD), squamous cell (LUSC), small cell lung cancer (SCLC), and mixed histology using circulating hypermethylated DNA. We characterized real-world (RW) demographics, genomic profiles, treatment patterns, and outcomes across MLSP-defined subtypes.
METHODS: Using the GuardantINFORM™ clinico-genomic database, we identified baseline cohorts with MLSP results of ≥90% probability predicting “pure” LUAD, LUSC, SCLC, or mixed lung cancer (LC). Patients used in classifier training were excluded. We summarized demographics, subtype-specific mutation frequencies, and first-line (1L) treatment categories—chemotherapy, immunotherapy (IO), chemo-IO, SCLC-specific regimens, and targeted therapy. RW time to treatment discontinuation (rwTTD) and time to next treatment (rwTTNT) were estimated via Kaplan-Meier methods.
RESULTS: Among 8,559 MLSP-tested patients, 69.4% were predicted as LUAD, 12.4% LUSC, 3.7% SCLC, and 14.6% mixed LC. Median age was 69-72 years; males predominated in LUSC and mixed LC. Ever-smoking was common across all subtypes. Genomic profiles were subtype-specific: LUAD enriched for KRAS, EGFR, BRAF; LUSC for TP53, PIK3CA, NFE2L2; SCLC for TP53 and RB1 loss; mixed LC showed overlapping NSCLC-SCLC features. 1L treatment varied: LUAD received chemo-IO (28.5%), chemotherapy (28.2%), IO (14.5%), targeted therapy (27.1%); LUSC and mixed LC were dominated by chemo-IO and chemotherapy; SCLC primarily chemo-IO. RW outcomes reflected biology: LUAD with targeted or IO-based regimens had longest rwTTD/rwTTNT; SCLC shortest durability; mixed LC intermediate. IO improved outcomes in LUAD/LUSC; SCLC benefitted most from chemo-IO.
CONCLUSIONS: Plasma-based MLSP show distinct demographic, genomic, and therapeutic patterns consistent with tissue-based histology. Mixed LC exhibits hybrid features and intermediate outcomes. RW effectiveness varied by subtype and regimen, supporting MLSP utility for histology-informed decisions when tissue is unavailable or discordant.
METHODS: Using the GuardantINFORM™ clinico-genomic database, we identified baseline cohorts with MLSP results of ≥90% probability predicting “pure” LUAD, LUSC, SCLC, or mixed lung cancer (LC). Patients used in classifier training were excluded. We summarized demographics, subtype-specific mutation frequencies, and first-line (1L) treatment categories—chemotherapy, immunotherapy (IO), chemo-IO, SCLC-specific regimens, and targeted therapy. RW time to treatment discontinuation (rwTTD) and time to next treatment (rwTTNT) were estimated via Kaplan-Meier methods.
RESULTS: Among 8,559 MLSP-tested patients, 69.4% were predicted as LUAD, 12.4% LUSC, 3.7% SCLC, and 14.6% mixed LC. Median age was 69-72 years; males predominated in LUSC and mixed LC. Ever-smoking was common across all subtypes. Genomic profiles were subtype-specific: LUAD enriched for KRAS, EGFR, BRAF; LUSC for TP53, PIK3CA, NFE2L2; SCLC for TP53 and RB1 loss; mixed LC showed overlapping NSCLC-SCLC features. 1L treatment varied: LUAD received chemo-IO (28.5%), chemotherapy (28.2%), IO (14.5%), targeted therapy (27.1%); LUSC and mixed LC were dominated by chemo-IO and chemotherapy; SCLC primarily chemo-IO. RW outcomes reflected biology: LUAD with targeted or IO-based regimens had longest rwTTD/rwTTNT; SCLC shortest durability; mixed LC intermediate. IO improved outcomes in LUAD/LUSC; SCLC benefitted most from chemo-IO.
CONCLUSIONS: Plasma-based MLSP show distinct demographic, genomic, and therapeutic patterns consistent with tissue-based histology. Mixed LC exhibits hybrid features and intermediate outcomes. RW effectiveness varied by subtype and regimen, supporting MLSP utility for histology-informed decisions when tissue is unavailable or discordant.
Conference/Value in Health Info
2026-05, ISPOR 2026, Philadelphia, PA, USA
Value in Health, Volume 29, Issue S6
Code
CO102
Topic
Clinical Outcomes
Topic Subcategory
Clinical Outcomes Assessment
Disease
SDC: Oncology, SDC: Respiratory-Related Disorders (Allergy, Asthma, Smoking, Other Respiratory), STA: Genetic, Regenerative & Curative Therapies, STA: Personalized & Precision Medicine