Noninferiority data move navigational bronchoscopy into frontline lung biopsy discussion

A fundamental facet of noninferiority trials is that your type 1 and type 2 errors are reversed. In a superiority trial, a type 1 error means that you falsely conclude that the "new" therapy is superior to control, and a type 2 error means that you falsely conclude that there is no difference. In a noninferiority trial (like the VERITAS trial) a type 1 error occurs when you falsely conclude that the difference between the "new" therapy and the standard/control therapy falls within the noninferiority margin when in reality the new therapy is inferior, and a type 2 error occurs when you conclude that the therapy is not within the noninferiority margin, when in reality it is. Because in the VERITAS trial, the noninferiority criterion was satisfied, we have to worry about a type 1 error: that navigational bronchoscopy is inferior or at least is not noninferior. How might that come to be?

Without delving into all the possible ways you may get a type 1 noninferiority error, I will point out just one: that navigational bronchoscopy had a leg up because it was performed only by the best operators at a select set of centers. These centers were chosen because of navigational bronchoscopy expertise, not transthoracic needle biopsy expertise (though the latter is a longstanding and commonly performed procedure, and it is expected that most centers will have high competence at it).