Airway obstruction after oral cancer surgery occurs as a result of a combination of multiple factors, such as large area excision of the mandible, tongue, and floor of the oral cavity; bilateral neck dissection; use of bulky reconstruction flaps; postoperative hematoma and pharyngolaryngeal edema; and relaxation of the tongue muscle [14, 15]. To avoid critical situations occurring as a result of postoperative airway obstructions, pathophysiological observations, such as those based on tracheal tug, neck ultrasound, SpO2, and monitoring by a capnometer, are necessary. However, reliability pertaining to diagnosis is low with any of these approaches. Therefore, in situations where difficulties in postoperative airway management can be predicted, the breathing status of patients should be carefully monitored to detect any abnormalities at an early time, enabling performance of emergency measures. Furthermore, patients at risk of airway obstruction should be identified such that they can undergo continued intratracheal intubation or selective tracheotomy [16, 17]. Typically, intratracheal intubation under sedation is used to maintain the airway between 24 and 48 hours after surgery. Selective tracheotomy is recommended in cases where intratracheal intubation must be maintained for more than 2 days. In a national survey in the UK, tracheotomy was selected for 69% (39/57) of patients who underwent free-flap head and neck reconstructive surgery [18]. However, complications, such as bleeding, occlusion, local infection, and pneumonia occur at a rate of 4–8% in tracheotomy [17–20]. These complications result in prolonged recovery of the patient and longer hospital stays. Appropriate strategies for airway management, therefore, remain controversial, and selective tracheotomy is determined based on the surgeon’s experience, which often varies between individuals. Therefore, research to establish criteria for performing tracheotomy is required. There have been several reports regarding evaluation methods to predict the necessity for selective tracheotomy in patients with oral cancer [9–12]. However, these methods may not be applicable for all patients due to disparities among institutions or differences in patient backgrounds. In this study, we examined the level of agreement between the need for tracheotomy based on these evaluation methods and the surgeon’s decision, as well as cases with a discrepancy between them.
Kruse et al. were the first to report a scoring system that can be used to determine the need for tracheotomy [9]. They scored and evaluated 5 components (tumor localization, tumor size, pathological chest X-ray findings, multimorbidity, and alcohol consumption) to predict the risk of postoperative respiratory failure in 928 patients. Kruse reported sensitivity as 97.6%. Notably, the scoring system by Kruse et al. does not evaluate various surgical factors. The expected reason for this may be that surgery is often avoided for patients with suspected disease based on chest X-rays; moreover, alcohol consumption is subject to recall bias. Surgical procedures, such as neck dissection and flap reconstruction, are highly invasive procedures; thus, these often are factors contributing to avoidance of surgery in patients with systemic disease. Considering such patients, this scoring method is not suitable for predicting postoperative airway management after surgical procedures or in the presence of localized tumors in patients with systemic disease.
Cameron et al. also reported a method that can be used to identify patients who require tracheotomy. Patients (n = 148) were scored based on tumor location, resected area of the mandible, neck dissection procedures, and reconstruction procedures [10]. This scoring system was created by combining elements used in evaluation based on the same items.
Based on the analysis of 486 patients, Gupta et al. reported scoring based on small and large categories, including resection, reconstruction methods, previous experience of radiotherapy, and degree of mouth opening; these are factors that may obstruct airway management even before surgery [12]. Gupta et al. reported that the sensitivity of the clinical assessment scoring system for tracheotomy (CASST) was 95.5%; selectivity was 99.5%; positive predictive value (PPV) was 96.9%; and negative predictive value (NPV) was 99.3%. Sensitivity was low in our patients, although selectivity and PPV were greater than 90%. In this system, scores are given for surgery (especially areas of resection) and reconstruction procedures. In terms of prediction of airway management associated with surgery, details of surgery are incorporated within this system in comparison with the other systems; therefore, this system is expected to be useful for surgeons.
Whether high sensitivity or high selectivity is required for these tests depends on the clinical state and study population. Since a number of analyses showed false-negative results depending on the criteria, scores became relatively low in partial glossectomy, as well as in cases where forearm flap reconstruction was performed with pull-through or supraomohyoid neck dissection. It is controversial as to whether to perform tracheotomy or to simply maintain intratracheal intubation under sedation in such cases. However, it is understood that tracheotomy is not performed on a large number of cases, despite high risks of postoperative airway obstruction. Careful postoperative monitoring and structures for managing emergencies are required for such cases to avoid malpractice. In addition to a system that can be used for distinguishing at-risk patients from those with false-negative results, it is important to combine several systems for evaluation. Similarly, false-positive diagnosis for tracheotomy must be avoided for patients who do not require this procedure. In this study, the κ coefficients for the Cameron and Gupta Scores were 0.61 and 0.6, respectively, indicating substantial agreement. The reason why the patients who needed tracheotomy based on the scores actually did not undergo tracheostomy was considered as follows: the Cameron and Gupta Scores were high in patients who underwent resection of the mandibular area and surrounding tissue (e.g. buccal mucosa or floor of mouth) and hard tissue reconstruction. However, in these cases, the surgeons decided that intubation under sedation was possible when postoperative aspiration was not a concern. We confirmed that the Cameron and Gupta Scores were consistent with the surgeon's decision to some extent and could be applied generally to clinical decisions in hospitals in Japan.
These scores are weighted for bilateral neck dissection. However, when unilateral total neck dissection was performed along with resection of the primary lesion, the score was not always high, which was considered to be the cause of discrepancy. These values are affected by prevalence, although scores were effective for screening postoperative airway management. However, Schmutz et al. have reported that patient population differs by institution; therefore, they failed to predict the need for tracheotomy based on these clinical scoring systems [21]. Similarly, Lee et al. reported that they could not identify correlations between the need for tracheotomy and the clinical findings for patients with oral cancer based on the Cameron score [22]. Moreover, Benetar et al. conducted analysis based on the Cameron score for performing elective tracheotomy in oral cancer surgery. Analysis revealed high selectivity and PPV (90% for both), low sensitivity (70%), and NPV (67%); these findings made it difficult to determine whether tracheotomy was necessary [23]. Upon reviewing reports that re-evaluate these scores, patients who actually require tracheotomy may not be accurately identified, and tracheotomy is suggested in a large proportion of cases. This is likely due to large differences in choices of surgical methods, decisions based on surgeons’ experience, and patient population. Moreover, postoperative hematoma and pharyngolaryngeal edema cannot be predicted directly from the scores investigated in this study. Clinically, airway obstruction rapidly advances in such cases, which then becomes irreversible. Intratracheal intubation or tracheotomy must be selected under emergency situations; these urgent decisions are the largest problem. These scoring systems were reported years ago, and since then, there have been advancement in equipment (e.g., energy devices), improvement in surgical techniques and perioperative care, and extensive changes in the applicability criteria of surgical procedures. Establishment of a scoring system that accommodates such advancements and changes is required in the future.
There were a few limitations in this study. Notably, we excluded patients who underwent cerclage because the risk of postoperative airway obstruction was expected to be low in such cases. However, there have been reported cases of severe outcomes resulting from occlusions, even if resection was solely performed on the frontal part of the mandible, as well as in cases where only the primary lesion was excised or a single neck dissection was performed [22, 23]. Even in cases that appear to be low risk, resection or abrasion of the genioglossus muscle, geniohyoid muscle, or mylohyoid muscle may cause deterioration in airway obstruction due to the loss of support of the hyoid bone. Such procedures with moderate surgical invasion are managed outside the intensive care unit and, therefore, pose a risk of delayed treatment of the airway obstruction. Evaluations of such cases are required in future.