Application of machine learning methods to guide patient management by predicting the risk of malignancy of Bethesda III-V thyroid nodules.

OBJECTIVE: Indeterminate thyroid nodules (ITN) are common and often lead to (sometimes unnecessary) diagnostic surgery. We aimed to evaluate the performance of two machine learning methods (ML), based on routinely available features to predict the risk of malignancy (RM) of ITN. DESIGN: Multi-centric diagnostic retrospective cohort study conducted between 2010 and 2020. METHODS: Adult patients who underwent surgery for at least one Bethesda III-V thyroid nodule (TN) with fully available medical records were included. Of the 7917 records reviewed, eligibility criteria were met in 1288 patients with 1335 TN. Patients were divided into training (940 TN) and validation cohort (395 TN). The diagnostic performance of a multivariate logistic regression model (LR) and its nomogram, and a random forest model (RF) in predicting the nature and RM of a TN were evaluated. All available clinical, biological, ultrasound, and cytological data of the patients were collected and used to construct the two algorithms. RESULTS: There were 253 (19%), 693 (52%), and 389 (29%) TN classified as Bethesda III, IV, and V, respectively, with an overall RM of 35%. Both cohorts were well-balanced for baseline characteristics. Both models were validated on the validation cohort, with performances in terms of specificity, sensitivity, positive predictive value, negative predictive value, and area under the receiver operating characteristic curve of 90%, 57.3%, 73.4%, 81.4%, 84% (CI95%: 78.5%-89.5%) for the LR model, and 87.6%, 54.7%, 68.1%, 80%, 82.6% (CI95%: 77.4%-87.9%) for the RF model, respectively. CONCLUSIONS: Our ML models performed well in predicting the nature of Bethesda III-V TN. In addition, our freely available online nomogram helped to refine the RM, identifying low-risk TN that may benefit from surveillance in up to a third of ITN, and thus may reduce the number of unnecessary surgeries.

Endoscopic vs transfacial resection of sinonasal adenocarcinomas.

OBJECTIVES: Compare the survival and the complication rates of 2 comparable groups of patients who underwent either endoscopic or transfacial resection of a sinonasal adenocarcinoma. STUDY DESIGN: Historical cohort study. SETTING: Tertiary referral center. SUBJECTS AND METHODS: Twenty-four patients with sinonasal adenocarcinomas who recently underwent an endoscopy (the ENDO group, 1999-2009) were compared with 24 patients who underwent a lateral rhinotomy procedure (the LR group, 1993-2007). The present retrospective study included patients who would have had an endoscopic resection in 2011. RESULTS: The groups were comparable in terms of age and sex (P = .49), as well as tumor stages; 43 of the 48 patients had undergone postoperative radiotherapy. There were 13 T1-T2 and 11 T3-T4 tumors in the ENDO group compared with 12 T1-T2 and 12 T3-T4 tumors in the LR group (P = .77). The mean length of follow-up was 38 months for the ENDO group and 89 for the LR group. The overall survival and recurrence-free rates were not significantly different (P = .3 and P = .87, respectively). The median duration of hospitalization was significantly shorter in the ENDO group than in the LR group (4 vs 8 days, P < .0001). The rate of early complications was identical in both groups (12.5%). CONCLUSION: The endoscopic approach is a safe and effective treatment in selected sinonasal adenocarcinoma cases. The early oncological outcome and morbidity associated with the endoscopic approach were comparable with a transfacial approach. Hospitalization was significantly reduced by the endoscopic approach. Although the endoscopic approach is less invasive, it requires an optimal preoperative imaging protocol and an experienced surgical team.

Respiratory complications after supracricoid partial laryngectomy.

The objective of the study was to analyze the incidence, treatment, and prevention of early and late respiratory complications in a series of patients who had supracricoid partial laryngectomies with either cricohyoidoepiglottopexy or cricohyoidopexy. From medical charts, we retrospectively reviewed 101 patients who underwent supracricoid partial laryngectomies, from 1980 to 2006, for laryngeal squamous cell carcinoma, and recorded the various postoperative complications and the time of decannulation. The mortality rate was 3.96%. Early complications included broncho-pulmonary infections and laryngeal stenoses which occurred in 9.9%. Univariate analysis showed a statistically significant relationship between the pulmonary complications and neck dissections (p < 0.04). Later, they were due to laryngeal obstruction (neolaryngeal mucosal flap, residual false vocal cord fold or arytenoid edema). The median decannulation time was 8 days, and there was a significant relationship between the decannulation delay and the pulmonary complications. Only two patients had a later definitive tracheotomy. Respiratory complications after supracricoid partial laryngectomy are frequent, but can be easily managed in most cases. A preoperative pulmonary assessment is necessary to select patients. During surgery, a precise impaction of the hyoid bone with the cricoid cartilage and a repositioning of an arytenoid can avoid some postoperative stenoses.