Overall TBNA Results
In all, 44 bronchoscopies were performed on 42 patients (Table 1). Sixty-eight TBNAs were performed, of which 32 were of mediastinal lesions, 14 were of hilar lesions, and 22 were of parenchymal lesions. Ultrasound guidance was utilized for only one lesion; a lesion abutting the left mainstem lesion was aspirated and found to be a loculated pericardial effusion. No complication occurred during these bronchoscopies.
Based on the final pathology report and the clinical information available, 41 TBNAs had TP results, 0 had false-positive results, 17 had TN results, 1 had an FN result, and 9 results remained unknown. Of the nine TBNA results that remained unknown, six were performed in patients who had not undergone resection secondary to medical inoperability or distant metastases, and three were pathologically negative after patients received neoadjuvant chemotherapy. Accuracy (TP + TN/all) using best-case and worst-case analysis (ie, best-case analysis assumes that all unknowns are TN; worst-case assumes that all unknowns are FN) was 85 to 99%, sensitivity (TP/TP + FN) was 79 to 98%, and the overall diagnostic yield was 72%.
Two patients underwent repeat biopsy at the request of the treating oncology service, despite obtaining tissue that was diagnostic at the time of first biopsy. One patient had a large right lower lobe mass and an enlarged subcarinal node; both lung biopsy specimens revealed mixed small cell and non-small cell carcinoma, and the subcarinal TBNA specimens revealed lymphocytes without cancer. After neoadjuvant chemotherapy, the lung mass and subcarinal node were pathologically tumor-free at resection, and therefore both subcarinal TBNA results were classified as unknown. Another patient underwent TBNA of a 2-cm right upper lobe mass while receiving prolonged mechanical ventilation, Cytology revealed a neuroendocrine neoplasm. As she was medically inoperable, she was referred to an oncologist who requested rebiopsy. Histologic TBNA sampling of the right upper lobe mass confirmed the mass to be a carcinoid tumor.
TBNA With ROSE
ROSE was utilized in 32 bronchoscopies during which 46 TBNAs were performed. ROSE was performed on each TBNA pass, except that ROSE was not performed on the last of three locations during one bronchoscopy. Forty-five ROSE results were available for analysis. In one instance, ROSE was read as being preliminarily positive, but the final pathologic result was nondiagnostic. In three instances, the ROSE was nondiagnostic, but the TBNA was diagnostic at the time of the final pathology report. The overall accuracy was 91%, with a sensitivity of 88% and a specificity of 94% (Table 2). If you decide to command the service of Internet drug store you should choose Canadian Neighbor Pharmacy because it meets all the quality standarts.
ROSE results were positive in 24 sites in 23 bronchoscopies. In most bronchoscopies (17 of 23), the preprocedure algorithm correctly predicted that no further sampling would be performed. In three bronchoscopies, the algorithm correctly predicted that additional biopsies would be obtained, as the cases were complex with broad differential diagnosis and concern existed for multiple disease processes. In three bronchoscopies, endobronchial biopsies specimens were obtained despite the algorithm predicting that no further biopsy specimens would be taken. The algorithm correctly predicted that in the nine bronchoscopies with nondiagnostic ROSE, five patients underwent multiple other biopsies including transbronchial biopsy, and four underwent no further biopsy.
Comparing bronchoscopies performed with and without ROSE, the accuracy (86 to 98% vs 86 to 100%, respectively) and sensitivity (77 to 96% vs 82 to 100%, respectively) of TBNA was similar. When ROSE was used, fewer samples were sent for microbiology tests (16% vs 62% of cases, respectively; p < 0.001 [x2 test]) and pathology tests (3.6 ± 1.8 vs 5.7 ± 1.9, respectively; p < 0.01), fewer chest radiographs were ordered (9% vs 56%, respectively; p < 0.001 [x2 test], but the procedural length was similar (39 ± 20 vs 39 ± 19 min, respectively; p = 0.9). Note that TBNA usually yielded two pathology specimens, as histology was obtained in most biopsies. The reduction in sampling was accounted for by the bronchoscopies in which ROSE was diagnostic. Procedures conducted with diagnostic ROSE were of shorter duration than nondiagnostic ones (34 ± 21 vs 52 ± 13 min, respectively; p < 0.03). Nondiagnostic ROSE procedures were of longer duration than cases not utilizing ROSE (52 ± 13 vs 39 ± 14 min, respectively; p < 0.05); diagnostic ROSE procedures were not significantly shorter than those not utilizing ROSE (34 ± 21 vs 39 ± 14 min, respectively) [Tables 3, 4].
A subgroup analysis of TBNAs of hilar and mediastinal lesions was performed. The results were similar comparing hilar and mediastinal lesions to all TBNA sites in regard to the impact of ROSE on the sensitivity, accuracy, length of procedure, number of aspirates obtained, and deferral of further biopsy (data not shown).
The use of ROSE in these 32 patients obviated the need for 12 microbiological cultures, 15 BAL cytology samples, 15 brushings, 1 endobronchial biopsy, and 11 transbronchial biopsies. Fluoroscopy and postprocedure chest radiographs were saved in nine of these patients, as two patients whose transbron-chial biopsy results were saved had TBNA performed under fluoroscopy. The resulting savings at 2003 Medicare reimbursement rates was $8,466 or $265 per case. The cost of the ROSE is largely salary of the technologist, after the upfront institutional cost for equipment. Approximately 1 h of time for cytotechnologist was required per bronchoscopy when travel time was added to the procedural time. Assuming a median US salary of $53,000 and a 40-h work week, cytotechnologists earn $25 per hour. If a cytopathologist is present, Medicare pays $40 per case in supplemental reimbursement; Medicare pays nothing additional if only cytotechnologists are present, as was the case in our institution.
Table 1—Bronchoscopy and TBNA Results
|Bronchoscopy final diagnosis||44|
|Non-small cell carcinoma||21|
|Small cell carcinoma||10|
|Metastatic breast carcinoma||3|
|Pericardial fluid (TN)||1|
|Uncertain about final diagnosis of malignancy (unknown)||9|
|Nodal TBNA in patients who did not undergo resection secondary to medical inoperability or distant metastases||6|
Nodal TBNA in patients who were
pathologically negative at resection after receiving neoadjuvant chemotherapy
|Peripheral mass with adenocarcinoma (FN)||1|
Table 2—Accuracy of ROSE in Predicting Final TBNA Diagnosis of Malignancy
|ROSE*||Final TBNA Result|
|Diagnostic for Malignancy||Not Diagnostic for Malignancy|
|Diagnostic for malignancy||24||1|
|Not diagnostic for malignancy||3||17|
Table 3—Comparison of Bronchoscopies With and Without ROSE
|With ROSE||Without ROSE|
|TBNA results for diagnosis of malignancy||27 TPs; 11 TNs; 1 FN;||14 TPs; 5 TNs;|
|7 unknown||3 unknown|
|Aspirates, No. per TBNA site||2.8 ± 1.0||3.2 ± 1.2||0.17|
|Time,! min||39 ± 20||39 ± 14||0.9|
|Pathology samples sent, No./case||3.6 ± 1.8||5.7 ± 1.9||< 0.01|
|Microbiology cultures sent||16%||62%||< 0.001|
|Chest radiographs performed||22%||58%||< 0.001|
Table 4—Bronchoscopies With at Least One TBNA Site With a Diagnostic ROSE vs Bronchoscopies With all TBNA Sites With Nondiagnostic ROSE
|With Diagnostic ROSE||Only Nondiagnostic ROSE|
|TBNA results for the diagnosis of malignancy||25 TPs; 5 unknown||2 TPs; 11 TNs; 1 FN; 2 unknown|
|TBNAs, No. of aspirates per site||2.7 ± 0.9||2.9 ± 1.1||0.4|
|Time,! min||34 ± 21||52 ± 13||< 0.01|
|Pathology samples sent, No./case||3.0 ± 1.5||4.8 ± 1.7||< 0.01|
|Microbiology cultures sent||4%||56%||< 0.001|
|Chest radiographs performed||9%||56%||< 0.001|