Abnormal chest computed tomography (CT) scans are a common finding in pulmonary practice. In a population of high-risk individuals screened for lung cancer, over 25% of patients had an abnormal chest CT scan.1–3 The probability of malignancy in an abnormal chest CT depends upon the characteristics of the abnormality and the patient’s clinical characteristics, such as age and smoking status, and range 0% to 80%.4 Many solitary pulmonary nodules (SPNs) contain non–small cell lung cancer (NSCLC), the most common type of lung cancer, and the larger the nodule, the more likely it is to contain cancer.3 As the early detection and resection of NSCLC has been shown to improve survival, prompt diagnosis and treatment of the SPN is vitally important.1
The probability of malignancy of a SPN may be estimated using existing models.2,5 If high, generally a probability of malignancy in excess of 60%, the American College of Chest Physicians clinical practice guidelines for the evaluation of patients with pulmonary nodules recommends going directly to surgical resection for both diagnostic and therapeutic reasons, either with video-assisted thoracoscopic surgery (VATS) or an open surgical procedure.3,6 This approach is expensive and commits the patient to the morbidity associated with VATS procedure but does have a very high likelihood of securing the diagnosis of NSCLC if it is present.7 Alternately, if the probability of malignancy is low, serial CT scans to monitor for increase in the size of the lesion over time is the preferred strategy.5–7 The management strategy for the SPN with an intermediate probability of malignancy or for patients with a high surgical risk frequently involves biopsy of the lesion, although the optimal biopsy approach has not been determined.
One diagnostic strategy is CT-guided biopsy which is 81% to 97% sensitive for the diagnosis of NSCLC.5,8 CT-guided biopsies, however, are associated with a pneumothorax rate of approximately 15%, over 40% of which require a chest tube to manage.9 This has led some to search for a better diagnostic technology.
Electromagnetic navigational bronchoscopy (ENB) with transbronchial biopsy is an increasingly common alternative technology for the diagnosis of the SPN. Navigational bronchoscopy has emerged as a technology that improves the sensitivity of conventional bronchoscopy.10–12 The technology has been reviewed elsewhere and increases the sensitivity of transbronchial biopsy to approximately 70% and it is associated with a low risk of complications, including approximately a 1.6% pneumothorax rate.12,13 It is, however, a new technology that is relatively expensive and its role in the multidisciplinary approach to the diagnosis of the SPN is not certain. Either CT-guided or ENB strategies may result in a nondiagnostic biopsy results, and if a patient is a surgical candidate VATS resection of the SPN is often performed to obtain a definitive diagnosis.14,15
Both the National Comprehensive Cancer Network and National Institute for Health and Clinical Excellence guidelines for the evaluation and treatment of lung cancer recommend a team-oriented, multidisciplinary approach.16,17 Multidisciplinary teams have shown to improve guideline adherence and process measures of quality.18–20 New and evolving diagnostic technologies might be best studied and implemented within a comprehensive, multidisciplinary thoracic oncology program where greater attention could be focused on how the technology is deployed in the nodule evaluation algorithm.21
We performed a cost-consequences analysis to understand the clinical consequences and cost differences of a CT-guided biopsy strategy versus an ENB biopsy strategy for the diagnosis of a SPN with an intermediate probability of NSCLC.
We created a decision model to evaluate the costs associated with both ENB and transbronchial biopsy versus CT-guided biopsy for the diagnosis of SPN (Fig. 1). A limited societal perspective was used, and indirect costs were thought to be negligible and similar between the 2 arms and were not included in the study. The inflation rate was estimated at 3%.22
The base case was 65-year old with a >40 pack-year smoking history with a 2 cm SPN. Such SPN would have approximately a 60% chance of malignancy.2,23 We assumed that there were no other features, radiographic or otherwise, to signal that there might be more advanced disease. In addition, we assumed that if a positron emission tomography (PET) scan were obtained before the biopsy, it would only affect the pretest probability of malignancy and not the operating characteristics of either diagnostic technology. The decision tree did not address the possibility of watchful waiting or follow-up imaging at a future time point, as it was assumed that a tissue diagnosis was desired. In addition, the location of the lesion within the chest or relative to an airway was not considered separately from the effects that those lesion features would have on the sensitivity of the 2 diagnostic technologies.
The first node on the decision tree represents the choice between ENB biopsy (upper path) and CT-guided biopsy (lower path).5,6,24 In the base case scenario, it was assumed that all nondiagnostic ENB and CT-guided biopsy attempts were sent directly to VATS for a diagnostic excisional biopsy.14,15 The decision tree shown in Figure 1 includes the possibility of performing the other technology if the first biopsy were nondiagnostic (ENB biopsy if CT-guided biopsy were performed first or CT-guided biopsy if ENB were performed first). However, in the base case the serial use of diagnostic testing was assumed to be 0 and all patients with nondiagnostic biopsies went on to VATS.
In our secondary analysis, we increased the rate of serial testing in both the arms to 100%, so that patients in the ENB arm who had a nondiagnostic ENB biopsy went on to CT-guided biopsy. Similarly, patients in the CT-guided biopsy arm with a nondiagnostic CT-guided biopsy all went on to ENB biopsy.
There are 3 main significant complications that can befall patients with either of these diagnostic technologies (Table 1).9,13 The first is pneumothorax, which occurs in approximately 15% of CT-guided biopsies and 1.6% of ENB biopsies. Approximately 56% of the time the pneumothorax resolves without further intervention. Chest tube placement and an observation stay in the hospital are required in approximate 44% of pneumothoraces, a number that is independent of the diagnostic technology that produced the pneumothorax.9,13,30
The second significant complication is clinically significant hemorrhage requiring hospitalization. This can either be in the airway (more common with navigational bronchoscopy with biopsy) or around the lung (more common with CT-guided biopsy) and occurs in approximately 1% of CT-guided biopsies and 0.1% of ENB biopsies.9,13 The final significant common complication captured by our decision tree is respiratory failure requiring hospitalization, and possibly mechanical ventilation. This occurs in approximately 1% of CT-guided biopsies and 0.1% of ENB biopsies.9,13
Baseline and periprocedural and complication utilities were all obtained from the literature.
The costs were obtained from the literature, the American Medical Association and private coding websites and are national Medicare reimbursement rates expressed in 2011 dollars (Table 2). For patients with ≥2 complications requiring hospitalization, the most expensive of the conditions subsumed the hospitalization costs of the other complications. A range of ±50% of values was used in the sensitivity analysis unless the numbers were clinically implausible.
The model did not include the cost of treating the ultimate diagnosis, be it either cancer or an unspecified benign condition, nor did it include any of the costs of the imaging carried out to evaluate the SPN before the diagnostic procedure. We assumed that there were no ultimate false-positive or false-negative results. An initial positive biopsy for either cancer or a noncancerous cause by either CT or navigational bronchoscopy was not confirmed with VATS in accordance with practice patterns and the literature.31
One-way and multivariable sensitivity analyses were performed, and multiple 1-way sensitivity analyses were plotted on a tornado diagram. Monte Carlo simulations with 104 iterations in the base case and in the secondary serial testing analysis were performed to allow multivariable sensitivity analysis and to estimate the 95% central cost range. All analyses were carried out in TreeAge Pro 2011 (TreeAge Software Inc., Williamstown, MA).
Results from the base case analysis and the alternate serial testing scenarios are both shown in Table 3. Both models had the same final probability of diagnosis in accordance with the assumption that there were no false-positive or false-negative diagnoses and that initial nondiagnostic biopsies went on for additional testing. In the base case scenario, the ENB with biopsy strategy was associated with a 20% increased rate of VATS surgery compared with the CT-guided biopsy strategy (30.0 per 100 patients vs. 10.0 per 100 patients). The ENB biopsy strategy, however, was associated with fewer complications. For every 100 ENB procedures, 13.4 fewer pneumothoraces were produced and 5.9 fewer chest tubes were placed compared with CT-guided biopsy. In addition, 0.9 fewer hemorrhages and 0.6 fewer cases of respiratory failure occurred.
The costs were greater in the ENB biopsy strategy. In the base case scenario, the ENB with biopsy strategy was on average $3719 per patient more expensive than the CT-guided biopsy strategy. Mean costs per biopsy were $6633 [95% confidence interval (CI), $1518-$18,511] and $2913 (95% CI, $1248-$18,241) in the ENB and CT-guided arms, respectively.
Figure 2 shows the tornado diagram of the univariate sensitivity analyses. Costs were most influenced by the sensitivity of the CT-guided biopsy. Given that a VATS procedure costs $16,993, a decrease in the sensitivity of the CT-guided biopsy led to an increase number of VATS procedures and greater costs. In fact, the cost of the VATS procedure itself is the second factor in the tornado diagram, highlighting its direct role in the overall costs. The third factor in the tornado diagram is the probability of performing an ENB biopsy after a nondiagnostic CT-guided biopsy, assumed to be 0 in the base case scenario.
The impact of the sequential use of diagnostic technologies is seen in the serial biopsy strategy shown in Table 3. If the use of CT-guided biopsy after nondiagnostic ENB biopsy (or conversely ENB biopsy after nondiagnostic CT-guided biopsy) were to increase from 0% in the base case scenario to 100%, use of VATS then falls to 3% in either arm. As expected, the rate of other complications increases in both the arms with a sequential approach. For example, the rate of pneumothorax in the ENB-first arm increases by 4.5 per 100 patients to 6.1 from 1.6 per 100 patients.
Costs are decreased in the both arms in the serial biopsy strategy. The average cost of the ENB biopsy strategy falls to $2406 (95% CI, $1518-$19,759) from $6633 (95% CI, $1518-$18,511), a savings of $4227 or 64%. Similarly, the average cost of the CT-guided biopsy strategy decreases by $978 or 34% to $1934 (95% CI, $1248-$19,759) from $2913 (95% CI, $1248-$18,241).
In 2-way sensitivity analysis, where 2 parameters are varied simultaneously, exploring the cost of the diagnostic strategies, we found an inverse linear relationship between the sensitivity of the CT-guided biopsy and the ENB biopsy (Fig. 3). For example, if ENB biopsy were to have a sensitivity of ≥95%, it would be the less expensive strategy if the sensitivity of CT-guided biopsy were <93.4%.
Additional 2-way sensitivity analysis revealed a curvilinear relationship between the sensitivity of either CT-guided or ENB biopsy and the probability of using the alternate diagnostic technology after an initial nondiagnostic biopsy and the overall cost. Figure 4 demonstrates the decreasing sensitivity of CT-guided biopsy needed to make it the less expensive technology as the probability of performing ENB after a nondiagnostic CT-guided biopsy increases from 0 (the base case scenario) to 1 (the serial biopsy scenario). For example, if the sensitivity of the CT-guided biopsy were 85% and VATS procedures were to follow all nondiagnostic CT-guided biopsies, ENB biopsy would be the less expensive strategy. However, if all nondiagnostic CT-guided biopsies were followed by ENB biopsy and not by VATS, not only do overall costs decrease, but the same CT-guided biopsy with 85% sensitivity would be less expensive than ENB. This shows that the likelihood of following 1 diagnostic technology with the other, in series, has an effect on overall costs as well as the relative sensitivities at which one of the diagnostic technologies is preferred.
Calculation of cost per quality-adjusted life year (QALY) allows comparison of the relative value of various diagnostic or therapeutic interventions. In an attempt to construct a QALY analysis, we evaluated the differences in utility (the numerator in QALY analyses, where, by definition, 1 y in perfect health has a utility of 1) based on published literature values. The baseline utility in our base case is approximately 0.84 to 0.88.29,39 The utilities associated with VATS and the various complications are similar: VATS (0.73 to 0.88), mechanical ventilation (0.76), pneumothorax (0.63), and relatively short-lived.25,29,40–43 Assuming a 2-week duration of disutility, an ENB-based strategy in the base case scenario would result in a loss of 0.002 QALY/person. Similarly, an ENB-based strategy in the serial biopsy scenario would experience a gain of 0.0009 QALY/person. Both of these values are significantly smaller than the precision of the QALY measurements themselves.
Electromagnetic navigation bronchoscopy is a relatively new technology that is associated with a decreased risk of pneumothorax and other complications versus the competing diagnostic technology, CT-guided biopsy, for the diagnosis of the SPN. The decrease in number of pneumothoraces, chest tubes, cases of hemorrhage, and respiratory failure seen with the ENB biopsy strategy, however, comes at a cost of an increased rate of VATS surgery and at an increased cost.
The average greater cost of ENB ($3719 additional cost/biopsy) versus CT-guided biopsy is largely driven by the high sensitivity of CT-guided biopsy and the need to use an alternate diagnostic technology to secure a diagnosis after a nondiagnostic ENB biopsy. Several factors have been reported to impact the sensitivities of ENB and CT-guided biopsies, like the “bronchus sign,” size of the lesion, and location of the lesion within the chest.13,44–47 These factors exert their influence on the relative sensitivities of ENB and CT-guided biopsies and were not considered individually.
Many of the studies performed when a new diagnostic technology emerges on the market look to define its operating characteristics, particularly its sensitivity and specificity relative to a current technology or a gold standard.13–15 In practice, diagnostic tests are often not used in isolation. Rather providers often seek to combine diagnostic tests to maximize certain parameters of their operating characteristics.
Combining ≥2 tests in a series serves to increase the sensitivity of the combined tests over each of the tests individually. Given that ENB and CT-guided biopsy have similar sensitivities for both benign and malignant disease, the combined serial sensitivity is: 1−(1−sensitivity [CT-guided])×(1−sensitivity[ENB]).13,30 The 3% serial nondiagnosis rate, and therefore VATS rate, is predicted by this formula. Such a combination assumes that there are no false-positives for either malignant or benign diagnoses.
PET scans are often obtained to evaluate the SPN, and a lesion that is PET avid has a higher posttest probability of cancer. We chose not to include PET scan status in our model explicitly because we felt that the PET result exerted its effects via the pretest probability of malignancy. As shown in Figure 2, the costs of both the CT-guided and the ENB biopsy strategy were relatively insensitive to the pretest probability of disease.
We did not include a watchful waiting arm in our decision tree. Because biopsies have similar sensitivities for both benign and malignant disease, the posttest probability of malignancy is not significantly decreased by a nondiagnostic biopsy.13,30 In addition, both the patient and the provider have a revealed preference to obtain a diagnosis by virtue of the fact that they chose to undergo a diagnostic procedure.5–7 We did not evaluate equivocal biopsy results (eg, “inflammation suggestive of a granuloma”), which would change the posttest probability of malignancy and could lead to a watchful waiting strategy.
Additional limitations include a lack of primary patient-level data and a reliance on literature values that might not be generalizable to all patient populations. Costs are given as Medicare costs and might not generalize to other countries, nor do they represent the cost to individual patients or health care systems.
We did not conduct a formal cost-utility analysis. The changes in QALY were small between the 2 arms. This was driven largely by the study design that considered only the peribiopsy period and did allow for different long-term QALY states. In addition, given the range of utility reported in the literature for the various health states considered, we did not feel that without patient-level utility data we could adequately comment on differences that were small in comparison with the precision of the measurements.
The context within which a diagnostic technology is used is an area that merits future study. We have shown that the costs and consequences of CT-guided biopsy and ENB biopsy are quite different depending upon how these technologies are used in the evaluation of the SPN in ways that go beyond their simple cost and sensitivity. Future work should focus on understanding the multidisciplinary environment in which the evaluation of malignancy occurs, including organizational factors like a “multidisciplinary nodule clinic” on resource utilization and patient-centered outcomes. In addition, provider and patient preference should be incorporated in decision analysis to understand why and in what circumstances providers and patients choose one diagnostic technology over another.
The use of navigational bronchoscopy with biopsy in preference to CT-guided biopsy of the SPN results in fewer episodes of pneumothorax, hemorrhage, and respiratory failure, but increases average costs by $3719 per case and is associated with a 20% absolute increase in VATS procedure rates. A diagnostic strategy that combines CT-guided biopsy and ENB biopsy in series decreases the rate of VATS procedures and decreases costs. An ENB-first sequential biopsy approach decreases average cost per case relative to CT-guided biopsy alone by $507. A CT-guided biopsy-first sequential biopsy approach decreases cost by $978 per case compared with CT-guided biopsy alone.
1. Aberle DR, Adams AM, Berg CD, et al. Reduced lung-cancer mortality with low-dose computed tomographic screening. N Engl J Med. 2011;365:395–409
2. Swensen SJ, Silverstein MD, Ilstrup DM, et al. The probability of malignancy in solitary pulmonary nodules: application to small radiologically indeterminate nodules. Arch Intern Med. 1997;157:849–855
3. Ost D, Fein AM, Feinsilver SH. The solitary pulmonary nodule
. N Engl J Med. 2003;348:2535–2542
4. Bach PB, Kattan MW, Thornquist MD, et al. Variations in lung cancer
risk among smokers. J Natl Cancer Inst. 2003;95:470–478
5. Ost DE, Gould MK. Decision making in the patient with pulmonary nodules. Am J Respir Crit Care Med. 2011;185:363–372
6. Gould MK, Fletcher J, Iannettoni MD, et al. Evaluation of patients with pulmonary nodules: when is it lung cancer
?: ACCP evidence-based clinical practice guidelines (2nd edition). Chest. 2007;132(suppl 3):108S–130S
7. Rivera MP, Mehta AC. Initial diagnosis of lung cancer
: ACCP evidence-based clinical practice guidelines (2nd edition). Chest. 2007;132(suppl 3):131S–148S
8. Yao X, Gomes MM, Tsao MS, et al. Fine-needle aspiration biopsy
versus core-needle biopsy
in diagnosing lung cancer
: a systematic review. Curr Oncol. 2012;19:e16–e27
9. Wiener RS, Schwartz LM, Woloshin S, et al. Population-based risk for complications after transthoracic needle lung biopsy
of a pulmonary nodule: an analysis of discharge records. Ann Intern Med. 2011;155:137–144
10. Shah PL, Singh S, Bower M, et al. The role of transbronchial fine needle aspiration in an integrated care pathway for the assessment of patients with suspected lung cancer
. J Thorac Oncol. 2006;1:324–327
11. Rand Du IA, Barber PV, Goldring J, et al. British Thoracic Society guideline for advanced diagnostic and therapeutic flexible bronchoscopy in adults. Thorax. 2011;66(suppl 3):1–21
12. Herth FJF, Eberhardt R. Flexible bronchoscopy and its role in the staging of non-small cell lung cancer
. Clin Chest Med. 2010;31:87–100
13. Wang Memoli JS, Nietert PJ, Silvestri GA. Meta-analysis of guided bronchoscopy for the evaluation of the pulmonary nodule. Chest. 2011;142:385–393
14. Makris D, Scherpereel A, Leroy S, et al. Electromagnetic navigation diagnostic bronchoscopy for small peripheral lung lesions. Eur Respir J. 2007;29:1187–1192
15. Gildea TR, Mazzone PJ, Karnak D, et al. Electromagnetic navigation diagnostic bronchoscopy: a prospective study. Am J Respir Crit Care Med. 2006;174:982–989
16. Ettinger DS, Akerley W, Bepler G, et al. Non-small cell lung cancer
. J Natl Compr Canc Netw. 2010;8:740–801
17. Baldwin DR, White B, Schmidt-Hansen M, et al. Diagnosis and treatment of lung cancer
: summary of updated NICE guidance. BMJ. 2011;342:d2110
18. Handy JR. Attributes contributing to superior outcomes in the surgical management of early-stage lung cancer
and examples of implementing improvement. Cancer J. 2011;17:57–62
19. Forrest LM, McMillan DC, McArdle CS, et al. An evaluation of the impact of a multidisciplinary team, in a single centre, on treatment and survival in patients with inoperable non-small-cell lung cancer
. Br J Cancer. 2005;93:977–978
20. Freeman RK, Van Woerkom JM, Vyverberg A, et al. The effect of a multidisciplinary thoracic malignancy conference on the treatment of patients with lung cancer
. Eur J Cardiothorac Surg. 2010;38:1–5
21. Davison AG, Eraut CD, Haque AS, et al. Telemedicine for multidisciplinary lung cancer
meetings. J Telemed Telecare. 2004;10:140–143
23. MacMahon H, Austin JHM, Gamsu G, et al. Guidelines for management of small pulmonary nodules detected on CT scans: a statement from the fleischner Society. Radiology. 2005;237:395–400
24. Ost D, Fein A. Management strategies for the solitary pulmonary nodule
. Curr Opin Pulm Med. 2004;10:272–278
25. Burfeind WR, Jaik NP, Villamizar N, et al. A cost-minimisation analysis of lobectomy: thoracoscopic versus posterolateral thoracotomy. Eur J Cardiothorac Surg. 2010;37:827–832
26. Rintoul RC, Slade MG. Another new tool for the diagnostic bronchoscopist. Thorax. 2011;66:1027–1028
27. Harewood GC, Wiersema MJ, Edell ES, et al. Cost-minimization analysis of alternative diagnostic approaches in a modeled patient with non-small cell lung cancer
and subcarinal lymphadenopathy. Mayo Clin Proc. 2002;77:155–164
29. Earle CCC, Chapman RHR, Baker CSC, et al. Systematic overview of cost-utility assessments in oncology. J Clin Oncol. 2000;18:3302–3317
30. Ost D, Shah R, Anasco E, et al. A randomized trial of CT fluoroscopic-guided bronchoscopy vs conventional bronchoscopy in patients with suspected lung cancer
. Chest. 2008;134:507–513
31. Zarbo R, Fenoglio-Preiser CM. Interinstitutional database for comparison of performance in lung fine-needle aspiration cytology. A College of American Pathologists Q-Probe Study of 5264 cases with histologic correlation. Arch Pathol Lab Med. 1992;116:463–470
33. Find-A-Code: Quick and Easy Medical Coding [Internet]. Find-A-Code. Available at: http://www.findacode.com
. Accessed December 3, 2011
34. Layfield LJ, Bentz JS, Gopez EV. Immediate on-site interpretation of fine-needle aspiration smears: a cost and compensation analysis. Cancer. 2001;93:319–322
37. Pearlstein DP, Quinn CC, Burtis CC, et al. Electromagnetic navigation bronchoscopy performed by thoracic surgeons: one center’s early success. Ann Thorac Surg. 2012;93:944–950
38. Edell E, Krier-Morrow D. Navigational bronchoscopy
: overview of technology and practical considerations—new current procedural terminology codes effective 2010. Chest. 2010;137:450–454
39. Hanmer J, Lawrence WF, Anderson JP, et al. Report of nationally representative values for the noninstitutionalized US adult population for 7 health-related quality-of-life scores. Med Decis Making. 2006;26:391–400
40. Timbie JWJ, Shahian DMD, Newhouse JPJ, et al. Composite measures for hospital quality using quality-adjusted life years. Stat Med. 2009;28:1238–1254
41. . Optimal strategy for the first episode of primary spontaneous pneumothorax in young men: a decision analysis. J Gen Intern Med. 2002;17:193–202
42. Handy JRJ, Asaph JWJ, Skokan LL, et al. What happens to patients undergoing lung cancer
surgery? Outcomes and quality of life before and after surgery. Chest. 2002;122:21–30
43. Ramsey SD, Shroyer AL, Sullivan SD, et al. Updated evaluation of the cost-effectiveness
of lung volume reduction surgery. Chest. 2007;131:823–832
44. Steinfort DP, Vincent J, Heinze S, et al. Comparative effectiveness of radial probe endobronchial ultrasound versus CT-guided needle biopsy
for evaluation of peripheral pulmonary lesions: a randomized pragmatic trial. Respir Med. 2011;105:1704–1711
45. Yamauchi Y, Izumi Y, Nakatsuka S, et al. Diagnostic performance of percutaneous core-needle lung biopsy
under CT scan fluoroscopic guidance for pulmonary lesions measuring ≤10 mm. Chest. 2011;140:1669–1670
46. Ernst A, Anantham D. Bronchus sign on CT scan rediscovered. Chest. 2010;138:1290–1292
47. Seijo LM, de Torres JP, Lozano MD, et al. Diagnostic yield of electromagnetic navigation bronchoscopy is highly dependent on the presence of a bronchus sign on CT imaging: results from a prospective study. Chest. 2010;138:1316–1321