Models Can Predict Related Lymphedema

Preoperative and postoperative models are highly accurate for predicting breast cancer–related lymphedema (BCRL), according to a study published in JAMA Surgery. Danielle H. Rochlin, MD, and colleagues conducted a prognostic study involving 1,1882 women with breast cancer (mean age, 55.6) who underwent axillary lymph node dissection (ALND) or sentinel lymph node biopsy (SLNB) from 1999 to 2020 to create a simple and accurate prediction model for BCRL. At a mean follow-up of 3.9 years, 11.6% of patients had a BCRL diagnosis. The BCRL rate was significantly higher among Black women compared with Asian, White, and women of other races (22.1% vs 12.5%, 10.1%, and 14.8%, respectively). Accuracy for model 1 (preoperative), which included age, weight, height, race, ALND/SLNB status, any radiation therapy, and any chemotherapy, was 73.0%, sensitivity was 76.6%, specificity was 72.5%, and the area under the receiver operating characteristic curve (AUC) was 0.78 at a cutoff of 0.18. For model 2 (postoperative), which included age, weight, race, ALND/SLNB status, any chemotherapy, and patient-reported arm swelling, accuracy was 81.1%, sensitivity was 78.0%, specificity was 81.5%, and the AUC was 0.86 at a cutoff of 0.10.