Back to 2025 Abstracts

---

DEVELOPMENT AND EXTERNAL VALIDATION OF A PREDICTIVE MODEL FOR NODAL METASTASES IN APPENDICEAL CANCER – WHO NEEDS RIGHT HEMICOLECTOMY?
Po Hong Tan^*1, Richard Max Miller¹, Andrew G. Saleeb¹, Hoe Yan Hor¹, Melody Pi Yin Tu¹, Arshia Pessaran¹, Mayar Alatout¹, Sanjay Bagaria², Travis E. Grotz³, Emmanuel Gabriel², Chee-Chee Stucky¹, Zhi Ven Fong¹, Yu-Hui Chang¹, Nabil Wasif^1
1Surgical Oncology, Mayo Clinic Phoenix, Mesa, AZ; ²Mayo Clinic in Florida, Jacksonville, FL; ³Mayo Clinic Minnesota, Rochester, MN

Introduction
Appendiceal cancers (AC) are rare tumors with varied histologies, each exhibiting a different propensity for nodal metastasis. The decision to perform right hemicolectomy following appendectomy for AC can be controversial. Our goal was to develop a prediction model to aid decision making in this patient population.
Methods
The National Cancer Database (NCDB) was utilized to identify patients diagnosed with AC who underwent surgery between 2010 and 2022. The histological subtypes included were neuroendocrine tumor, goblet cell carcinoma, mucinous adenocarcinoma, non-mucinous adenocarcinoma, and signet cell carcinoma. Only patients with at least one lymph node sampled, known T and N staging, grades, and documented lymphovascular invasion (LVI) were included. Multivariable logistic regression was used to develop a risk prediction model and internally validated with bootstrapping. The Brier score and Area Under the Curve (AUC) were used to measure model performance. The model was then externally validated with data from patients diagnosed with AC between 1995 and 2024 across three hospitals of an institution.
Results
Patient baseline demographics were similar for the training NCDB set (n= 18,833) and the institutional external validation set (n= 303). The most common histology was non-mucinous adenocarcinoma in the NCDB group (26.2%) and mucinous adenocarcinoma in the institution group (35.0%). The overall rate of node positive disease was 27.1% in the NCDB group and 33.3% in the validation group. On multivariable regression analysis the strongest predictor of nodal metastases was LVI (OR 6.57, 95% CI 6.05-7.13), increasing T stage (T4 OR 6.54 95% CI 5.40-7.86), and increasing tumor grade (G3 OR 4.03, 95% CI 2.60-4.83, G4 OR 3.52 95% CI 2.58-4.80). The initial risk prediction model using NCDB data comprising of patient age, sex, T stage, tumor grade and LVI showed an AUC of 0.84(95% CI 0.83-0.84) and Brier score of 0.14. External validation using institutional data showed an AUC of 0.87(95% CI 0.83-0.91) and Brier score of 0.14.
Conclusions
The risk of nodal metastases in patients with AC can be predicted with reasonable accuracy using patient age, sex, tumor T stage, grade and LVI. Our predictive model can be used as an aid in clinical decision-making, particularly in determining the need for a right hemicolectomy following appendectomy.

---

Back to 2025 Abstracts