Introduction: Due to the large numbers of negative results of sputum examinations in the face of malignant disease, we used a decision analytic model to determine whether, and under what conditions, sputum cytology (Spt) might have a cost-effective role to play in the approach to lung lesions.
Methods: We constructed a decision analytic model to compare the utility of Spt, image-directed fine-needle aspiration (FNA), bronchoscopic examination (Bronch), and open biopsy (OBx) in the evaluation of lung lesions. Prevalence and cost data were derived from local databases and diagnosis-related groups. Diagnostic sensitivity (sens) and specificity (spec) of the tests were derived from the literature and local data. Output of the model was lowest cost per correct malignant diagnosis and included surgical treatment costs. We did not attempt to model survival data or morbidity. Sensitivity analyses were performed using cost, test sensitivity, and lesion size variables.
Results: In the baseline case, a patient who is a surgical candidate with lesion size 2.8 cm, prevalence of malignancy=0.67, FNA sens=0.95, Bronch sens=0.80, Spt sens=0.51, OBx is the best initial procedure with a cost per correct diagnosis of $12,888. Sputum examination has the highest cost per correct diagnosis of $63,424. FNA and bronchoscopy have cost per correct diagnosis of $21,543 and $16,615, respectively. Sens analysis shows that OBx is the preferred strategy in the workup of lung lesions in patients who are surgical candidates older than 30 years of age. Spt is the preferred strategy only when the patient is not a surgical candidate, the lesion size is large (>4.7 cm), and only if sputum sens is >0.45.
Conclusions: Spt is not cost-effective in clinical practice, except when the patient has a large clinically unresectable lesion and the laboratory can demonstrate sufficient test sens.