A retrospective two-center cohort study has developed and internally validated an integer-based clinical score to predict conversion to surgery among patients with acute appendicitis initially managed nonoperatively. The prediction model specifically evaluates patients during a reassessment intended to take place approximately 24 hours after starting nonoperative treatment.
The study analyzed records from 376 patients across two medical centers who began care with nonoperative management. From this initial group, researchers established an analytic cohort of 280 eligible patients who reached the approximately 24-hour reassessment mark without having undergone an appendectomy prior to that evaluation and who met all required eligibility and data-completeness criteria.
To select candidate predictors from patient data, investigators employed least absolute shrinkage and selection operator regression combined with 10-fold cross-validation. Cutoff values for continuous variables were established using receiver operating characteristic curve analysis and the Youden index. The team then derived an integer-based clinical score using multivariable logistic regression models.
During the remainder of the index hospitalization, conversion to surgery occurred in 43 of the 280 patients, representing an overall rate of 15.4 percent. The resulting scoring system assigns points to three clinical parameters measured at reassessment: elevated body temperature accounts for two points, while white blood cell count and C-reactive protein concentration contribute one point each.
Model performance evaluation showed an apparent area under the receiver operating characteristic curve of 0.835 with a 95 percent confidence interval ranging from 0.774 to 0.896. Internal validation using 1,000 bootstrap resamples with complete model redevelopment demonstrated an optimism-corrected area under the curve of 0.781 with a 95 percent bootstrap percentile interval between 0.718 and 0.838.
Internal validation metrics also revealed a Brier score of 0.120 with a 95 percent bootstrap percentile interval of 0.095 to 0.145, alongside a calibration slope of 0.749 with an interval spanning 0.459 to 1.009. These validation steps accounted for optimism across the entire model derivation process.
Observed conversion rates aligned progressively with cumulative score values across the patient cohort. Patients scoring zero points experienced a conversion-to-surgery rate of 1.9 percent, while scores of one, two, three, and four corresponded to observed surgical conversion rates of 6.7 percent, 14.9 percent, 40.9 percent, and 47.7 percent, respectively.
The investigators concluded that the integer-based score may support observed-risk stratification and structured reassessment in selected patients continuing nonoperative management. However, the authors noted that the score should not serve as a stand-alone indication for surgery, emphasizing that external validation and prospective evaluations of clinical utility are necessary before clinical adoption.
The prediction framework specifically addresses patient outcomes across the remainder of the index hospitalization after initiating nonoperative care. Furthermore, repeating the complete model derivation process across all one thousand bootstrap resamples accounted for statistical optimism during both predictor selection and threshold determination.
Clinically, this risk stratification applies exclusively to patients who remain candidates for continuing nonoperative management following that structured reassessment. The authors concluded that prospective studies evaluating clinical utility and external validation across independent patient cohorts are required before adopting the scoring system into routine medical practice.
Subject of Research: Clinical prediction score for surgical conversion in acute appendicitis
Article Title: Development and internal validation of a clinical score to predict conversion to surgery after an approximately 24-hour reassessment among patients with acute appendicitis initially managed nonoperatively: a retrospective two-center cohort study
Article References: Mori, M., Shuto, K., Kosugi, C., Narushima, K., Hirano, A., Usui, A., Oka, Y., Suzuki, D., Fujino, T., Yamazaki, K., & Shimizu, H. (2026). Development and internal validation of a clinical score to predict conversion to surgery after an approximately 24-hour reassessment among patients with acute appendicitis initially managed nonoperatively: a retrospective two-center cohort study. BMC Gastroenterology. https://doi.org/10.1186/s12876-026-05615-6
Image Credits: AI Generated
DOI: 10.1186/s12876-026-05615-6
Keywords: acute appendicitis, nonoperative management, clinical score, surgical conversion, appendectomy, internal validation, bootstrap resampling, receiver operating characteristic, body temperature, white blood cell count, C-reactive protein, LASSO regression, logistic regression, risk stratification, cohort study, Brier score, calibration slope
News Source: Ophelia Keating. (November 21, 2026). Clinical Score Predicts Surgical Conversion in Nonoperatively Managed Acute Appendicitis. Scienmag.



