Machine Learning Identifies Predictors of Revision Endoscopic Sinus Surgery

A machine learning approach identified age and several comorbidities, such as nasal polyps and asthma, as being associated with revision endoscopic sinus surgery among patients with chronic rhinosinusitis.

A machine learning approach identified age and several comorbidities, such as nasal polyps and asthma, as being associated with revision endoscopic sinus surgery (ESS) among patients with chronic rhinosinusitis (CRS). Results were published in PLoS One.

For patients with CRS whose disease burden is uncontrolled by standard-of-care intranasal corticosteroids and nasal saline irrigation, ESS has provided a cost-effective option linked with significant improvement of symptoms and health-related quality of life. However, researchers note that several risk factors have been associated with the need for revision ESS, which is estimated to occur in more than 1 in 5 patients after 5 to 10 years.

“The early identification of CRS recurrence risk following ESS is cost-effective, helping to correctly target treatment and prevent permanent tissue change,” said the study authors. “No prior research has analyzed the prediction accuracy of revision ESS at the individual level or for variables with a nonlinear association.”

Researchers conducted a retrospective follow-up study of patients 16 years and older with CRS to examine the accuracy of a personalized prediction approach for revision ESS, as well as identify the effects of predictor variables via modern machine-learning algorithms and methods.

They collected demographic and clinical variables from the electronic health records of 767 surgical patients with CRS from the Department of Otorhinolaryngology at the Hospital District of Helsinki and Uusimaa, Finland.

“The prediction accuracy of revision ESS was examined by training and validating different machine learning models, while the effects of variables were analyzed using the Shapley values and partial dependence plots,” explained the study authors.

Revision ESS was performed on 111 (14.5%) patients, of whom 88 underwent 1 revision ESS and 23 patients reported 2 or more revisions.

Analyses indicated that the logistic regression, gradient boosting, and random forest classifiers performed similarly in predicting revision ESS as shown by area under the receiving operating characteristic curve (AUROC) values of 0.744, 0.741, and 0.730, respectively, using data collected from the baseline visit until 6 months after baseline ESS.

The length of time during which data were collected was shown to improve the prediction performance, as data collection times of 0, 3, 6, and 12 months after baseline ESS exhibited AUROC values for the logistic regression of 0.682, 0.715, 0.744, and 0.784, respectively.

Several important predictors were associated with revision ESS:

  • number of visits before or after baseline ESS
  • number of days from the baseline visit to the baseline ESS
  • patient age
  • CRS with nasal polyps (CRSwNP)
  • asthma
  • nonsteroidal anti-inflammatory drug–exacerbated respiratory disease (NERD)
  • immunodeficiency or suspicion of it

Patient age and number of visits before baseline ESS were noted to carry nonlinear effects for predictions. A lower risk for revision ESS was observed among patients logging 10 to 25 clinical visits between the baseline visit and baseline ESS than patients with fewer than 10 or more than 25 clinic visits. Further assessment was cited for this association due to the small sample size.

“Although these findings require validation in other populations, our results reinforce the importance of diagnostics and the management of NERD, CRSwNP, asthma, and other comorbidities to prevent uncontrolled CRS, and carry relevancy for patient counselling specifically,” concluded researchers.


Nuutinen M, Haukka J, Virkkula P, Torkki P, Toppila-Salmi S. Using machine learning for the personalised prediction of revision endoscopic sinus surgery. PLoS One. 2022;17(4):e0267146. doi:10.1371/journal.pone.0267146