Home // eTELEMED 2013, The Fifth International Conference on eHealth, Telemedicine, and Social Medicine // View article

Prediction of Metastatic Events in Patients With Cutaneous Melanoma

Authors:
Christian Scheibboeck
Patrick Huber
Stefanie Weber
Kaan Harmankaya
Romina Nemecek
Jessika Weingast
Michael Binder
Thomas Mehl
Christian Schuh
Stephan Dreiseitl

Keywords: cutaneous melanoma; TNM classification; artificial intelligence; decision support; knowledge-based system

Abstract:
Cutaneous melanoma, one of the most aggressive malignant tumors, potentially leads to widespread metastasis. The prediction of early metastatic events by using clinical information and data from specific tumor markers could substantially augment the quality of diagnostic and treatment decisions. To predict potential metastatic events during follow-up in patients with cutaneous melanoma, a knowledge-based system will be used during clinical routine by interpreting data from clinical history of the patient in combination with data from tumor markers. Specifically, data will be sent to an expert system including a rule engine which offers the physician a risk assessment and decision support. The interpretation of the tumor markers (n=493) resulted in a prediction sensitivity and specificity of 77.80% and 69.55% while using the multivariate combination of MIA, S100β and LDH. Additionally, the risk of metastasis was calculated based on fitted survival functions and was integrated into our system. Currently this knowledge-based system will calculate the individual likelihood for metastatic events based on the risk of the primary tumor, the duration of observation since the primary event and the recent values of tumor markers. The system aims to produce results that are compatible with medical expert’s opinion.

Pages: 220 to 223

Copyright: Copyright (c) IARIA, 2013

Publication date: February 24, 2013

Published in: conference

ISSN: 2308-4359

ISBN: 978-1-61208-252-3

Location: Nice, France

Dates: from February 24, 2013 to March 1, 2013