Abstract Rationale Indeterminate pulmonary nodules (IPNs) are a frequent clinical challenge, and determining which nodules warrant biopsy versus surveillance remains difficult. Current guidelines recommend fluorodeoxyglucose positron emission tomography (FDG-PET) as an option, yet FDG uptake is nonspecific and may lead to unnecessary invasive procedures, particularly in granulomatous regions. Recently developed blood-based biomarker tests have demonstrated potential for refining malignancy risk assessment, but their optimal role in clinical pathways remains uncertain. We therefore undertook this study to compare diagnostic strategies that begin with either a blood-based biomarker approach or PET-CT in the evaluation and management of indeterminate pulmonary nodules. Methods A Monte Carlo simulation modeled 1,000 virtual patients per run across 1,000 independent runs, assuming a 20% cancer prevalence. Each simulated patient was assigned cancer status and test results according to published performance estimates: PET (sensitivity 0.89, specificity 0.75), and a combined biomarker approach with a rule-in autoantibody test (0.41, 0.93), and a rule out proteomic test (0.97, 0.44). Two diagnostic strategies were compared: (1) PET-first, followed by biopsy if positive, surveillance if negative. (2) Biomarker-first followed by biopsy if positive (plus PET for staging if cancer); surveillance if negative; PET if indeterminate. Primary outcomes were benign biopsies, cancers assigned to surveillance, and total orders. A secondary analysis simulated histoplasmosis-endemic regions by reducing PET specificity and sensitivity based upon published values (0.61 and 0.40). A sensitivity analysis varied cancer prevalence from 10%-50% to assess performance across clinical contexts. P values represent a paired t-test of difference within each run. Results At 20% cancer prevalence, the biomarker-first strategy increased the malignancy rate among biopsied nodules from 47.2% (95% CI, 44.2-50.4%) with the PET-first pathway to 53.5% (50.0-57.4%). This approach also reduced cancers assigned to surveillance (16 9-25 vs 22 14-31) and lowered benign biopsies from 199 (176-222) to 160 (137-182), while requiring 35% fewer PET scans (613 580-642 vs 1,000 1,000-1,000). All differences were statistically significant (p 0.001). Performance remained consistent across a 10-50% prevalence range. In histoplasmosis-endemic simulations, the biomarker-first approach maintained higher malignancy yield (45.6% 42.9-49.0 and 37.5% 35.3-40.0 vs 36.3% 34.2-38.6 and 27.0% 25.5-28.6 for PET-first). Conclusions A biomarker first approach may have the potential to increase the malignancy rate among biopsied nodules, while reducing unnecessary invasive procedures and PET utilization. Considering costs, radiation, and biopsy risks, this data supports use of biomarkers with confirmation through prospective pragmatic clinical studies. This abstract is funded by: Biodesix
Kammer et al. (Fri,) studied this question.