• Benchmarked Cameron and CASST in intraoral myomucosal flap reconstructions. • Scores under-predicted tracheostomy versus real-world clinical decisions. • High specificity but low sensitivity; slight agreement (κ≈0.19/0.03) • Safe avoidance feasible: no airway events without tracheostomy; LOS 8 vs 20. Selective tracheostomy after resection of oral tongue cancer remains contentious, particularly when reconstruction uses intraoral myomucosal flaps whose bulk and edema are not well represented in existing airway risk scores. We retrospectively reviewed 62 consecutive patients who underwent partial glossectomy with myomucosal (buccinator/FAMM-type) flap reconstruction. Primary outcome was elective tracheostomy; secondary outcomes included airway events, ICU utilization, and length of stay (LOS). We compared clinical decisions with the Cameron and Gupta/CASST scores and fitted a multivariable logistic regression to evaluate associations with elective tracheostomy. Selective tracheostomy was performed in 43/62 patients (69%); no airway emergencies occurred among patients managed without tracheostomy. ICU admission occurred in 11% without tracheostomy versus 65% with tracheostomy, and median LOS was 8 versus 20 days, respectively. At prespecified thresholds, both scores showed high specificity, but low sensitivity compared with real-world decisions (Cameron ≥ 5: sensitivity 32.6%, specificity 94.7%; Gupta/CASST ≥ 7: sensitivity 4.7%, specificity 100%); agreement beyond chance was slight (κ≈0.19 and κ≈0.03). The multivariable model demonstrated acceptable discrimination (ROC-AUC 0.73); no individual covariate reached statistical significance, although advancing T stage showed a non-significant trend toward elective tracheostomy. All flaps survived, with only minor distal ischemia managed conservatively. These data suggest that, in the context of intraoral myomucosal reconstruction, institutional thresholds for elective tracheostomy exceed those implied by generic scores; conversely, carefully selected patients can safely avoid tracheostomy. Tailoring risk prediction to flap-specific mechanics may refine decision support and reduce unnecessary tracheostomies without compromising safety.
Manfuso et al. (Sun,) studied this question.