Clinical benefit and cost-effectiveness analysis of liquid biopsy application in patients with advanced non-small cell lung cancer (NSCLC): a modelling approach

Abstract

Purpose: Targeted therapies are effective therapeutic approaches in advanced stages of NSCLC and require precise molecular profiling to identify oncogenic drivers. Differential diagnosis on a molecular level contributes to clinical decision making. Liquid biopsy (LB) use has demonstrated its potential to serve as an alternative to tissue biopsy (TB) particularly in cases where tissue sampling is not feasible or insufficient. We aimed at evaluating the cost-effectiveness of ctDNA-based LB use (molecular multigene testing) according to German care guidelines for metastatic NSCLC.

Methods: A Markov model was developed to compare the costs and clinical benefits associated with the use of LB as an add-on to TB according to the guidelines for NSCLC patients. Usual care TB served as comparator. A microsimulation model was used to simulate a cohort of non-squamous NSCLC patients stage IV. The parameters used for modelling were obtained from the literature and from the prospective German CRISP registry ("Clinical Research platform Into molecular testing, treatment, and outcome of non-Small cell lung carcinoma Patients"). For each pathway, average direct medical costs, and QALYs gained per patient were used for calculating incremental cost-effectiveness ratios (ICER).

Results: The use of LB as an add-on was costlier (€144,981 vs. €144,587) but more effective measured in QALYs (1.20 vs. 1.19) for the care pathway of NSCLC patients (ICER €53,909/QALY). Cost-effectiveness was shown for EGFR-mutated patients (ICER €-13,247/QALY).

Conclusion: Including LB as an add-on into the care pathway of advanced NSCLC has positive clinical effects in terms of QALYs accompanied by a moderate cost-effectiveness.

Keywords: Cost-effectiveness; Liquid biopsy; Molecular profiling; NSCLC.

Fig. 1

Markov model and its health states used for economic modelling

Fig. 2

Model structure—Care pathway with liquid biopsy as an add-on. Schematic diagram shows the decision tree model structure. It illustrates the care pathway with LB as an add-on and the respective biopsy procedures. LB liquid biopsy, TB tissue biopsy. Alterations are divided into four gene-subgroups: ALK translocation, EGFR mutations, BRAF-V600 mutation, ROS1 translocation. No oncogenic driver includes wildtypes and other alterations that do not belong to the listed alterations of the genes ALK, BRAF, EGFR, and ROS1. Following agents were used for treatment: EGFR: afatinib (1st-line) – osimertinib (2nd-line); EGFR: osimertinib (1st-line) – atezolizumab + paclitaxel + carboplatin + bevacizumab (2nd-line) or pembrolizumab (2nd-line) or pembrolizumab + pemetrexed + carboplatin (2nd-line); ALK: alectinib (1st-line) – lorlatinib (2nd-line); BRAF-V600: dabrafenib + trametinib (1st-line) – pembrolizumab or pembrolizumab + pemetrexed + carboplatin (2nd-line); ROS1: crizotinib (1st-line) – pembrolizumab or pembrolizumab + pemetrexed + carboplatin (2nd-line); Wildtype and others: pembrolizumab (1st-line) – paclitaxel + carboplatin + bevacizumab (2nd-line); Wildtype: pembrolizumab + pemetrexed + carboplatin (1st-line) – docetaxel + nintedanib (2nd-line)

Fig. 3

Treatment lines of non-squamous NSCLC and corresponding progression-free survival. Illustration of the current personalized treatment options for non-squamous NSCLC. The therapies were selected according to evidence-based German guidelines. If more than two therapies were available for first- or second-line use the two most frequently prescribed substances were selected according to the German CRISP report 2020. ^aSince the combination of PEM + CAR + PBZ is currently under investigation in the second-line setting (KEYNOTE-789 (U. S. National Library of Medicine/ClinicalTrials.gov 2018)), the clinical parameters for the combination of PAC + CAR + BEV were used for modelling purposes. AFA afatinib, ALC alectinib, ATE atezolizumab, BEV bevacizumab, CAR carboplatin, CRI crizotinib, DAB dabrafenib, DOC docetaxel, LOR lorlatinib, NIN nintedanib, OSI osimertinib, PAC paclitaxel, PBZ pembrolizumab, PEM pemetrexed, TRA trametinib

Fig. 4

Progression-free survival (PFS) and overall survival (OS) derived from the respective therapy lines of the corresponding care pathways. A Subgroup (cohort I) includes all patients, regardless of whether a tissue biopsy can be taken, or the tissue is sufficient for molecular analysis. B Subgroup (cohort II) includes patients in whom tissue rebiopsy or molecular analysis on primary tissue sample is not possible. The figure indicates the survival data achieved in the respective care pathway. The survival data were calculated for different subgroups. Driver alterations are divided into four subgroups: ALK translocation, EGFR mutations, BRAF-V600 mutation, ROS1 translocation; The total cohort includes all NSCLC cases (wildtype, and driver alterations of the genes ALK, BRAF, EGFR, and ROS1)

Fig. 5

ICER of the competing care pathways. ICER indicates incremental cost-effectiveness ratio for the subgroups based on the underlying biomarker profile. The ICER of the total cohort (cohort I) is presented, and the ICER of the subgroup in which TB or molecular analysis on tissue sample is not possible (cohort II)

Fig. 6

Tornado diagram and the influence on the ICER. The tornado diagram is based on a cost-effectiveness simulation without microsimulation. The expected value (EV) deviates from the EV calculated by the 10,000 trials via microsimulation. Black: low value; grey: high value

Fig. 7

Cost-effectiveness acceptability curve. A curve illustrating the probability that a care pathway is cost-effective based on different WTP values for one gained QALY