User login
FT. LAUDERDALE, FLA. – Appropriate staging of non–small cell lung cancer (NSCLC) is critical to patient treatment and prognosis. Endobronchial ultrasound–guided fine-needle aspiration (EBUS-FNA) has gradually gained acceptance as a diagnostic tool to pathologically stage the mediastinum in patients with NSCLC, according to Dr. Bryan A. Whitson at the annual meeting of the Society of Thoracic Surgeons.
Because EBUS-FNA is a test to detect cancer, it is critical that it have few false-negative results, i.e., have a high negative predictive value (NPV). However, by virtue of the size and volume of mediastinal lymph node FNA samples, nondiagnostic results occur with some frequency.
Nondiagnostic samples decrease a test’s diagnostic performance, since they cannot help guide a clinical decision, and a portion of these samples may actually be false negatives.
The standard calculation of NPV does not factor in nondiagnostic samples. According to a study by the researchers in the Division of Thoracic and Foregut Surgery at the University of Minnesota, which Dr. Whitson presented, one must take nondiagnostic samples into consideration to determine the true diagnostic performance of EBUS-FNA.
To conservatively calculate NPV, the researchers modified the standard definition (true negatives/[true negatives plus false negatives]), creating an alternative NPV definition (true negatives/[true negatives plus false negatives plus nondiagnostic]). This definition takes into account the possibility that these nondiagnostic samples may be false negatives, which assumes the worst-case scenario.
The thoracic team at the University of Minnesota then compared the results of these two definitions of NPV in a retrospective analysis of their prospective database of patients with NSCLC who had EBUS-FNA between January 2007 and July 2011.
Dr. Bryan Whitson, who is a surgeon at the University of Minnesota, presented the team’s results.
A total of 120 patients with NSCLC who underwent EBUS-FNA were included in the analsyis. EBUS-FNAs were assessed using rapid on-site cytologic evaluation (ROSE) and a false-negative definition consisting of negative FNAs coupled to NSCLC-positive surgical biopsy of the same site. Diagnostic performance was evaluated comparing results of the inclusion or exclusion of nondiagnostic samples in the calculation of NPV.
Seven (5.8%) of the 120 patients had nondiagnostic results. One patient had a false negative. When the seven nondiagnostic procedures were excluded, the NPV was 96.3%; when they were included, the NPV dropped to 76.5%.
Both sensitivity and accuracy also dropped with the inclusion of the nondiagnostic procedures. Sensitivity dropped from 98.8% to 91% and overall accuracy from 98.2% to 92.2% when nondiagnostic procedures were included.
"Our data indicate that a conservative calculation of NPV that includes nondiagnostic samples should be used to assess the real diagnostic accuracy of EBUS-FNA in patients with NSCLC. Otherwise, decisions based on these results could be flawed and patients inappropriately staged. In our experience, and using this assessment, EBUS-FNA shows a true NPV that is less than most clinicians assume," Dr. Whitson said.
Ultimately, the thoracic surgeons at the University of Minnesota believe that "EBUS-FNA should be the primary method to stage the mediastinum in patients with NSCLC, since it enables pathologic staging without the need for an incision, with preservation of tissue planes, and with minimal complications."
However, in view of the limitations of EBUS-FNA, "we perform an immediate surgical biopsy when in doubt of the reliability of an EBUS-FNA result. A thoracic surgeon facile in EBUS can offer most patients minimally invasive mediastinal staging, with selective surgical biopsy in the same anesthetic setting, which can ensure accuracy and streamline patient care," according to Dr. Whitson and his colleagues.
FT. LAUDERDALE, FLA. – Appropriate staging of non–small cell lung cancer (NSCLC) is critical to patient treatment and prognosis. Endobronchial ultrasound–guided fine-needle aspiration (EBUS-FNA) has gradually gained acceptance as a diagnostic tool to pathologically stage the mediastinum in patients with NSCLC, according to Dr. Bryan A. Whitson at the annual meeting of the Society of Thoracic Surgeons.
Because EBUS-FNA is a test to detect cancer, it is critical that it have few false-negative results, i.e., have a high negative predictive value (NPV). However, by virtue of the size and volume of mediastinal lymph node FNA samples, nondiagnostic results occur with some frequency.
Nondiagnostic samples decrease a test’s diagnostic performance, since they cannot help guide a clinical decision, and a portion of these samples may actually be false negatives.
The standard calculation of NPV does not factor in nondiagnostic samples. According to a study by the researchers in the Division of Thoracic and Foregut Surgery at the University of Minnesota, which Dr. Whitson presented, one must take nondiagnostic samples into consideration to determine the true diagnostic performance of EBUS-FNA.
To conservatively calculate NPV, the researchers modified the standard definition (true negatives/[true negatives plus false negatives]), creating an alternative NPV definition (true negatives/[true negatives plus false negatives plus nondiagnostic]). This definition takes into account the possibility that these nondiagnostic samples may be false negatives, which assumes the worst-case scenario.
The thoracic team at the University of Minnesota then compared the results of these two definitions of NPV in a retrospective analysis of their prospective database of patients with NSCLC who had EBUS-FNA between January 2007 and July 2011.
Dr. Bryan Whitson, who is a surgeon at the University of Minnesota, presented the team’s results.
A total of 120 patients with NSCLC who underwent EBUS-FNA were included in the analsyis. EBUS-FNAs were assessed using rapid on-site cytologic evaluation (ROSE) and a false-negative definition consisting of negative FNAs coupled to NSCLC-positive surgical biopsy of the same site. Diagnostic performance was evaluated comparing results of the inclusion or exclusion of nondiagnostic samples in the calculation of NPV.
Seven (5.8%) of the 120 patients had nondiagnostic results. One patient had a false negative. When the seven nondiagnostic procedures were excluded, the NPV was 96.3%; when they were included, the NPV dropped to 76.5%.
Both sensitivity and accuracy also dropped with the inclusion of the nondiagnostic procedures. Sensitivity dropped from 98.8% to 91% and overall accuracy from 98.2% to 92.2% when nondiagnostic procedures were included.
"Our data indicate that a conservative calculation of NPV that includes nondiagnostic samples should be used to assess the real diagnostic accuracy of EBUS-FNA in patients with NSCLC. Otherwise, decisions based on these results could be flawed and patients inappropriately staged. In our experience, and using this assessment, EBUS-FNA shows a true NPV that is less than most clinicians assume," Dr. Whitson said.
Ultimately, the thoracic surgeons at the University of Minnesota believe that "EBUS-FNA should be the primary method to stage the mediastinum in patients with NSCLC, since it enables pathologic staging without the need for an incision, with preservation of tissue planes, and with minimal complications."
However, in view of the limitations of EBUS-FNA, "we perform an immediate surgical biopsy when in doubt of the reliability of an EBUS-FNA result. A thoracic surgeon facile in EBUS can offer most patients minimally invasive mediastinal staging, with selective surgical biopsy in the same anesthetic setting, which can ensure accuracy and streamline patient care," according to Dr. Whitson and his colleagues.
FT. LAUDERDALE, FLA. – Appropriate staging of non–small cell lung cancer (NSCLC) is critical to patient treatment and prognosis. Endobronchial ultrasound–guided fine-needle aspiration (EBUS-FNA) has gradually gained acceptance as a diagnostic tool to pathologically stage the mediastinum in patients with NSCLC, according to Dr. Bryan A. Whitson at the annual meeting of the Society of Thoracic Surgeons.
Because EBUS-FNA is a test to detect cancer, it is critical that it have few false-negative results, i.e., have a high negative predictive value (NPV). However, by virtue of the size and volume of mediastinal lymph node FNA samples, nondiagnostic results occur with some frequency.
Nondiagnostic samples decrease a test’s diagnostic performance, since they cannot help guide a clinical decision, and a portion of these samples may actually be false negatives.
The standard calculation of NPV does not factor in nondiagnostic samples. According to a study by the researchers in the Division of Thoracic and Foregut Surgery at the University of Minnesota, which Dr. Whitson presented, one must take nondiagnostic samples into consideration to determine the true diagnostic performance of EBUS-FNA.
To conservatively calculate NPV, the researchers modified the standard definition (true negatives/[true negatives plus false negatives]), creating an alternative NPV definition (true negatives/[true negatives plus false negatives plus nondiagnostic]). This definition takes into account the possibility that these nondiagnostic samples may be false negatives, which assumes the worst-case scenario.
The thoracic team at the University of Minnesota then compared the results of these two definitions of NPV in a retrospective analysis of their prospective database of patients with NSCLC who had EBUS-FNA between January 2007 and July 2011.
Dr. Bryan Whitson, who is a surgeon at the University of Minnesota, presented the team’s results.
A total of 120 patients with NSCLC who underwent EBUS-FNA were included in the analsyis. EBUS-FNAs were assessed using rapid on-site cytologic evaluation (ROSE) and a false-negative definition consisting of negative FNAs coupled to NSCLC-positive surgical biopsy of the same site. Diagnostic performance was evaluated comparing results of the inclusion or exclusion of nondiagnostic samples in the calculation of NPV.
Seven (5.8%) of the 120 patients had nondiagnostic results. One patient had a false negative. When the seven nondiagnostic procedures were excluded, the NPV was 96.3%; when they were included, the NPV dropped to 76.5%.
Both sensitivity and accuracy also dropped with the inclusion of the nondiagnostic procedures. Sensitivity dropped from 98.8% to 91% and overall accuracy from 98.2% to 92.2% when nondiagnostic procedures were included.
"Our data indicate that a conservative calculation of NPV that includes nondiagnostic samples should be used to assess the real diagnostic accuracy of EBUS-FNA in patients with NSCLC. Otherwise, decisions based on these results could be flawed and patients inappropriately staged. In our experience, and using this assessment, EBUS-FNA shows a true NPV that is less than most clinicians assume," Dr. Whitson said.
Ultimately, the thoracic surgeons at the University of Minnesota believe that "EBUS-FNA should be the primary method to stage the mediastinum in patients with NSCLC, since it enables pathologic staging without the need for an incision, with preservation of tissue planes, and with minimal complications."
However, in view of the limitations of EBUS-FNA, "we perform an immediate surgical biopsy when in doubt of the reliability of an EBUS-FNA result. A thoracic surgeon facile in EBUS can offer most patients minimally invasive mediastinal staging, with selective surgical biopsy in the same anesthetic setting, which can ensure accuracy and streamline patient care," according to Dr. Whitson and his colleagues.