Annals of Surgical Oncology

, Volume 19, Issue 8, pp 2580–2589

Nomograms for Predicting the Risk of Arm Lymphedema after Axillary Dissection in Breast Cancer

Authors

    • Escola Nacional de Saúde Pública/FIOCRUZ
    • Hospital Sirio-Libanes
  • Michael W. Kattan
    • Department of Quantitative Health SciencesCleveland Clinic
  • Yu Changhong
    • Department of Quantitative Health SciencesCleveland Clinic
  • Sergio Koifman
    • Escola Nacional de Saúde Pública/FIOCRUZ
  • Inês E. Mattos
    • Escola Nacional de Saúde Pública/FIOCRUZ
  • Rosalina J. Koifman
    • Escola Nacional de Saúde Pública/FIOCRUZ
  • Anke Bergmann
    • Instituto Nacional de Câncer (INCA)
    • Centro Universitario Augusto Motta-UNISUAM
Breast Oncology

DOI: 10.1245/s10434-012-2290-x

Cite this article as:
Bevilacqua, J.L.B., Kattan, M.W., Changhong, Y. et al. Ann Surg Oncol (2012) 19: 2580. doi:10.1245/s10434-012-2290-x

Abstract

Background

Lymphedema (LE) after axillary lymph node dissection (ALND) is a multifactorial, chronic, and disabling condition that currently affects an estimated 4 million people worldwide. Although several risk factors have been described, it is difficult to estimate the risk in individual patients. We therefore developed nomograms based on a large data set.

Methods

Clinicopathologic features were collected from a prospective cohort comprising 1,054 women with unilateral breast cancer undergoing ALND as part of their surgical treatment from August 2001 to November 2002. LE was defined as a volume difference of at least 200 ml between arms at 6 months or more after surgery. The cumulative incidence of LE was ascertained by the Kaplan–Meier method, and Cox proportional hazard models were used to predict the risk of developing LE on the basis of the available data at each time point: model 1, preoperatively; model 2, within 6 months from surgery; and model 3, at 6 months or later after surgery.

Results

The 5 year cumulative incidence of LE was 30.3%. Independent risk factors for LE were age, body mass index, ipsilateral arm chemotherapy infusions, level of ALND, location of radiotherapy field, development of postoperative seroma, infection, and early edema. When applied to the validation set, the concordance indices were 0.706, 0.729, and 0.736 for models 1, 2, and 3, respectively.

Conclusions

The proposed nomograms can help physicians and patients predict the 5 year probability of LE after ALND for breast cancer. Free online versions of the nomograms are available at http://www.lymphedemarisk.com/.

Supplementary material

10434_2012_2290_MOESM1_ESM.doc (83 kb)
Supplementary material 1 (DOC 83 kb)

Copyright information

© Society of Surgical Oncology 2012