Mobile Platform for Assessment, Early Detection, and Management of Breast Cancer-Related Lymphedema

Friday, 28 July 2017: 2:30 PM

Jane Armer, PhD
Sinclair School of Nursing and Ellis Fischel Cancer Center, University of Missouri, Columbia, MO, USA
Nathan C. Armer, MEd
Sinclair School of Nursing, University of Missouri, Columbia, MO, USA
Guilherme N. DeSouza, PhD
Department of Electrical and Computer Engineering and MU Informatics Institute, University of Missouri, Columbia, MO, USA

Purpose: Early detection and management of secondary lymphedema (LE) can significantly reduce the potential for symptoms and complications. Unfortunately, many patients fail to seek medical assistance at first signs of the disease. It is estimated that approximately 500,000 people in the U.S. suffer from lymphedema, and 2.4 million breast cancer survivors are at-risk for developing this chronic condition. Objectives were to: (1) Test the two completely automated measuring systems we developed (using an IR depth camera and a smart phone) for individuals to use at home or in the clinic for early detection and self-monitoring of LE; (2) Integrate and evaluate the automated smart phone application in the clinical setting.

Methods:  Over 280 data points (upper limbs) were sampled from healthy people and LE patients to exam the correlation between the mobile platform systems and the “gold standards” (water displacement and perometry); and to determine the reliability of both proposed methods. In comparison with water-displacement, 14 measurements of upper limbs from many repetitions for 7 healthy test subjects were done. Regarding the comparison with the Perometer, 34 upper limbs from 17 human subjects, including LE patients, were measured. IRB approval was received for integration of the 3D smart phone images with circumferences and symptom assessment in a prospective surveillance model with breast cancer survivors at risk of lymphedema.

Results: Pearson correlations of 0.97-0.98 with the kinect and smartphone 3D mobile platform systems and clinical limb volume measures of water displacement and perometry are reported. The two completely automated and robust systems for 3D image of human arms outperformed existing methods in many aspects, including cost, maintenance, and ease of use, while they maintained high correlations with the “gold standards.” Patients and staff are open to, and, indeed, eager to, participate in the application of the mobile device in the clinical setting.

Conclusion: The systems will accurately assist patients in managing LE through early detection and monitoring in the clinic and at home. Algorithms will be further refined to facilitate even more efficient and accurate data capture in the clinical and home setting. We will be doing further testing as we work to assess the ease of use by patients and clinicians with variable levels of comfort with technology.