关键词: 吞咽造影; 卒中后吞咽障碍; 转归; 列线图

Abstract: Objective  To construct a nomogram prediction model for poor short-term outcome of post-stroke dysphagia (PSD) based on video fluoroscopic swallowing study temporal and kinematic parameters, aiming to guide clinical nutritional support decisions. 
Methods  A total of 281 patients with PSD admitted to Cangzhou People’s Hospital from January 2021 to February 2025 were consecutively enrolled as the research subjects. Clinical data potentially associated with dysphagia were collected and statistically analyzed. Patients were divided into a training set (196 cases) and a validation set (85 cases) in a 7∶3 ratio by complete randomization. The proportion of patients with persistent dysphagia at 2 weeks after stroke (poor short-term outcome) in the training set was recorded, and its relationship with temporal and kinematic parameters of video fluoroscopic swallowing study, as well as other clinical data, was analyzed. Multivariate logistic stepwise regression analysis was used to identify independent influencing factors for poor short-term outcome of PSD, and a nomogram prediction model was constructed. ROC curves and calibration curves were plotted to evaluate the predictive efficacy of the model. Decision curve analysis was applied to assess its application value in clinical nutritional support decisions. The nomogram was validated in the validation set.
Results  Among the 196 PSD patients in the training set, 86 (43.88%) had poor short-term outcomes, and 40 (47.06%) of the 85 PSD patients in the validation set showed poor short-term outcomes. There were no statistically significant differences in the rate of poor short-term outcomes and other clinical data between the training set and the validation set. Advanced age, higher NIHSS score at 24 hours post-admission, higher functional dysphagia scale score, higher penetration-aspiration scale score, prolonged oral transit time, prolonged cricopharyngeal opening time, lower pharyngeal contraction ratio, and shorter hyoid anterior displacement were independent risk factors for poor short-term outcome of PSD (all P<0.05). The nomogram prediction model based on video fluoroscopic swallowing study temporal and kinematic parameters showed good calibration and goodness-of-fit in both the training and validation sets (mean absolute errors between the predicted and actual values were 0.016 and 0.030, and the P-values of the Hosmer-Lemeshow test were 0.735 and 0.245, respectively). ROC curves showed that the AUCs of the nomogram prediction model for poor short-term outcome of PSD based on temporal and kinematic parameters were 0.945 (95%CI 0.915-0.975) in the training set and 0.944 (95%CI 0.896-0.991) in the validation set, outperforming the nomogram model based on conventional influencing factors [AUCs 0.893 (95%CI 0.846-0.939) and 0.881 (95%CI 0.805-0.956), respectively]. Decision curve analysis showed that when the threshold probability ranged from 0.05 to 0.97, the net benefit of the nomogram prediction model for poor short-term outcome of PSD (based on video fluoroscopic swallowing study temporal and kinematic parameters) was higher than that of the two extreme strategies: assuming all patients would experience poor short-term outcome of PSD or assuming no patients would experience poor short-term outcome of PSD before intervention.
Conclusions  The nomogram prediction model for poor short-term outcome of PSD, constructed based on the temporal and kinematic parameters of video fluoroscopic swallowing study, is helpful for prediction of the risk of poor short-term outcome of PSD, guiding the clinical formulation of appropriate enteral feeding decisions, and optimizing the management of medical resources.

Key words: Video fluoroscopic swallowing; Post-stroke dysphagia; Outcome; Nomogram