Developing and validating a nomogram to predict myasthenia gravis exacerbation in patients with postoperative thymoma recurrence

Abstract

Background: Myasthenia gravis (MG) is one of the most common accessory syndromes for thymoma patients. To reduce MG exacerbation and guide clinical practice, we developed a nomogram for predicting MG exacerbation in patients with postoperative thymoma recurrence.

Methods: Retrospective study of 176 patients with recurrence following thymoma resection who were admitted to the People's Hospital of Shijiazhuang's Center of Treatment of Myasthenia Gravis between 2013 and 2021. Among them, 120 patients with recurrent thymoma from 2013 to 2020 were selected as the training cohort, and 56 patients in 2021 as the validation cohort. Univariate and multivariate logical regression was used to determine the risk factors and draw the nomogram, and the parameters in the nomogram were proportionally assigned from 0 to 100 points. Finally, the performance of the model is evaluated by internal inspection and external inspection.

Results: Multivariate analysis revealed that postoperative treatment plan and the pathologic classification of the thymoma were independent predictors of MG exacerbation in the training cohort (n=120), so they were used to create the nomogram, which had a well-fit calibration curve and good concordance index of 0.77 [95% confidence interval (CI): 0.69-0.86] for the training cohort and 0.74 (95% CI: 0.58-0.91) for the validation cohort, respectively. Calculations were made to determine the nomogram's sensitivity, specificity, positive predictive values (PPV) and negative predictive values (NPV). The training cohort were 75.7% (95% CI: 66.1-83.4%), 64.7% (95% CI: 38.6-84.7%), 92.9% (95% CI: 84.5-97.1%) and 30.6% (95% CI: 16.9-48.3%) respectively, while the corresponding validation cohort were 84.1% (95% CI: 69.3-92.8%), 66.7% (95% CI: 35.4-88.7%), 90.2% (95% CI: 75.9-96.8%) and 53.3% (95% CI: 27.4-77.7%) respectively.

Conclusions: We identified the risk factors for MG exacerbation in patients with postoperative recurrence of thymoma and drew a nomogram, which can be used to calculate the probability of MG exacerbation and guide clinicians to choose post-operative treatment.

Keywords: Exacerbation; myasthenia gravis (MG); prediction model; recurrence; thymoma.

Figure 1

Flow chart of patient enrollment.

Figure 2

The probability of MG exacerbation was predicted by the nomogram based on pathology and therapy (A). The pathology and therapy values are found on the points axis, and the sum of the two values is found on the total points axis. The corresponding exacerbation rate of this value is used to predict the possibility of MG exacerbation. Calibration curve of (B) the Training Cohort and (C) the Validation Cohort. An x-axis represents the predictive potential for MG exacerbation and an y-axis represents the actual incidence of it. Perfect prediction based on an idealized model is represented by the diagonal dotted line. The ideal performance of the nomogram is represented by the solid line, and the closer the solid line fits the dotted line, the better the prediction will be. RC, synchronous chemoradiotherapy; OH, high-dose steroid pulse therapy only; RH, radiotherapy combined with high-dose steroid pulse therapy; OR, radiotherapy combined with other treatments; OC, chemotherapy only; OO, other treatments; ROC, receiver operating characteristic; MG, myasthenia gravis.