Valuing Innovation in the Bronchoscopy Suite Authors: Caroline M. Godfrey, Stephen A. Deppen Published: 2024 The landscape of minimally invasive lung cancer diagnostics has changed radically over the last decade. Traditionally, diagnostic bronchoscopy and endobronchial ultrasound bronchoscopy were performed in the operating room by surgeons, with a benign or malignant diagnosis equally likely. This invasive and inefficient approach became overwhelmed by the explosion of indeterminate pulmonary nodules (IPNs) discovered incidentally on imaging obtained for other purposes, accounting for 1.5 million annual “incidentalomas” (1), and through the recent expansion of lung cancer screening guidelines (2). This rapid rise in imaging-based IPN identification, coinciding with the technological advancement of the bronchoscopy suite, has necessitated a shift in the diagnostic approach to IPNs from surgeon-driven in the operating room to being led by interventional pulmonologists and radiologists using less invasive techniques. This shift has also resulted in a welcomed drop in the most invasive diagnostic modality, nontherapeutic thoracotomy (3). Currently published cost-effectiveness models comparing diagnostic approaches to IPNs focused on comparing specific diagnostic modalities and their associated cost-effectiveness to the current standard of care or to other specific diagnostic modalities. In this issue of AnnalsATS, Ost and colleagues (pp. 1759–1769) present a modality-agnostic model estimating the potential clinical benefits and quantifying the economic value of improving the sensitivity of pulmonary nodule diagnostics for lung cancer (4). Using a decision-tree model with a Markov lung cancer progression submodel, Ost and colleagues pose a baseline clinical scenario of a solid peripheral nodule whose population cancer prevalence is 67%. The lesion size, location, and patient histories were constant and consistent with the Clinical Evaluation of superDimension Navigation System for Electromagnetic Navigation Bronchoscopy trial (NAVIGATE) (5) and an older meta-analysis dominated by literature before 2010, larger (>20 mm median size) nodules, and studies reporting cancer prevalence exceeding 70%, which may not represent nodule clinic populations encountered today (6). Of note, the authors assumed immediate treatment after a cancer diagnosis at initial or subsequent evaluation (in the setting of initial false negative), with the specificity of the bronchoscopy modality fixed at 100% and the sensitivity varying from 70% at baseline compared with 80% sensitivity with the technological innovation. The assumption of immediate treatment does simplify the model; however, it is unlikely to be reflective of reality because of health care system limitations and loss to follow-up. Costs for bronchoscopy were assumed to be the same between the baseline and improved sensitivity methods, thus making only the incremental impact of improved sensitivity the metric of change. Costs for true negatives were also considered the same across technological modalities, and any healthcare activity specifically arising from noncancer results was not included in the model. The model was built from a payer perspective over a 30-year time horizon. The primary economic outcome was net monetary benefit, which is defined as the incremental health benefit, measured in quality-adjusted life years and subsequently monetized with the willingness to pay at $100,000 net of incremental costs. However, in this instance there were no cost differences between implementation of the low- and high-sensitivity bronchoscopy techniques outside of increased lifetime costs ($6,727) arising from more time under treatment in the high-sensitivity group; therefore, the net monetary benefit was driven by the reduction of costs incurred due to the resultant decreased time to treatment, in the number of subsequent diagnostic procedures and the stage shift toward earlier diagnosis. Quality-adjusted life years increased by 0.13 in the high-sensitivity population. The estimated net monetary benefit from increased sensitivity from 70% to 80% was $8,729, with most of the value ($3,668) accruing in the first 5 years from initial bronchoscopy. Most of this improvement in adjusted life years and indirect monetary benefit is driven by 100% immediate follow-up. In real life, the time to treatment is not immediate and appears to be increasing for some health systems as specialist and advanced imaging availability demand outstrips supply. As the time to treatment approaches 4 weeks, much of the observed technological benefit fades. Similarly, as the prevalence of lung cancer decreases to 50%, then the monetary benefit also decreases to $6,515. A modality-agnostic model does present limitations. Decision analysis models are useful primarily in their ability to model benefits as well as drawbacks in a comparison in which the cost-effective choice is not known. In this case, comparing a modality with improved sensitivity with no change in specificity, complication rate, or cost typically does not require a model to prove cost-effectiveness, as the improved sensitivity modality does not have any associated drawbacks, so the cost-effective choice is relatively easy to predict. Although it is standard practice with comparative effectiveness studies to report an incremental cost-effectiveness ratio, in this case, that value is of limited utility given no reported drawbacks to the high-sensitivity modality. In addition, although assuming no change in specificity or complication rate may be reasonable assumptions for the reasons outlined in the manuscript, the assumption that there is no change in cost of biopsy is more challenging to reconcile, as new technologies are typically associated with increased cost. Results from the efforts of Ost and colleagues define the improvement in time to diagnosis (4 wk) and quantify in dollars ($8,729) the potential impact of greater diagnostic sensitivity. Their work further emphasizes the importance of continued efforts to improve minimally invasive highly sensitive diagnostic methods for the rapid diagnosis of lung cancer. However, if time to treatment after diagnosis in your healthcare system approaches 4 weeks, efforts to decrease that time may well equal the benefit from future technological innovation.