Back to 2025 Abstracts
Predicting Cancer-related Lymphedema: A Decade Of Longitudinal Outcomes From A Tertiary Cancer Center
Stav Brown, MD, Yizhuo Shen, BS, Alexzandra Mattia, BS, Bohdan Pomahac, MD, Siba Haykal, MD PhD.
Yale Cancer Center, Yale School of Medicine, New Haven, CT, USA.
Purpose: No clinically applicable method currently exists to predict which patients will develop lymphedema after axillary lymph node dissection (ALND) or when it will occur. This decade-long study aimed to identify lymphedema predictors and create an algorithm estimating both risk and timing of onset.
Methods: We conducted a longitudinal-study of patients who underwent ALND at a tertiary-cancer-center from 2013 to 2024. Demographic and clinical data, along with lymphedema onset, were recorded. Two multivariate-regression models were developed to predict 1) risk of lymphedema and 2) time to first diagnosis.
Results: Among 15,666 patients, 14.9% (2,345) developed lymphedema, with an average onset of 20.5 months post-ALND. Key predictors included chemotherapy (OR=2.3, 95% CI 2.1-2.5, p<0.0001), radiation (OR=1.9, 95% CI 1.8-2.1, p<0.0001), black race (OR=1.4, 95% CI 1.3-1.7, p<0.0001), BMI>30 (OR=1.4, 95% CI 1.3-1.5, p<0.0001), and diabetes (OR=1.2, 95% CI 1.1-1.3, p<0.0001). Rapid-onset predictors included black race (β=-4.2, 95% CI -6.8 to -1.4, p=0.002), radiation (β=-4.5, 95% CI -6.3 to -2.6, p<0.0001), and diabetes (β=-2.8, 95% CI -5.0 to -0.6, p=0.01). Hemoglobin-A1C was an independent predictive factor for both lymphedema development (OR=1.2, 95% CI 1.1-1.3, p=0.04) and rapid disease development (β=-3.3, 95% CI -6.1 to -1.2, p=0.003). Model accuracy was 79.65% (AUC=0.72, 95% CI 0.71-0.73, p<0.0001).
Conclusion: This is the largest longitudinal study to provide clinically applicable prediction algorithms for individualized risk and time-to-diagnosis predictions, and the first to highlight Hb-A1C as a novel independent predictive marker. These findings offer a foundation for individualized screening protocols and early preventative measures in high-risk patients.
Back to 2025 Abstracts