Format

Send to

Choose Destination
Stat Med. 2017 Apr 30;36(9):1491-1505. doi: 10.1002/sim.7213. Epub 2017 Jan 15.

Disease severity classification using quantitative magnetic resonance imaging data of cartilage in femoroacetabular impingement.

Author information

1
Arbor Research Collaborative for Health, Ann Arbor, MI, USA.
2
Department of Biostatistics and Informatics, University of Colorado, Denver, Denver CO, USA.
3
Optum, United Health Group, Eden Prairie, MN, USA.
4
Center for Magnetic Resonance Research, Department of Radiology, University of Minnesota, Minneapolis, MN, USA.
5
Department of Orthopedic Surgery, University of Connecticut Health Center, Farmington, MN, USA.
6
Department of Applied Physics, University of Eastern Finland, Kuopio, Finland.
7
Department of Orthopaedic Surgery, University of Minnesota, Minneapolis, MN, USA.

Abstract

Femoroacetabular impingement (FAI) is a condition in which subtle deformities of the femoral head and acetabulum (hip socket) result in pathological abutment during hip motion. FAI is a common cause of hip pain and can lead to acetabular cartilage damage and osteoarthritis. For some patients with FAI, surgical intervention is indicated, and it can improve quality of life and potentially delay the onset of osteoarthritis. For other patients, however, surgery is contraindicated because significant cartilage damage has already occurred. Unfortunately, current imaging modalities (X-rays and conventional MRI) are subjective and lack the sensitivity to distinguish these two groups reliably. In this paper, we describe the pairing of T2* mapping data (an investigational, objective MRI sequence) and a spatial proportional odds model for surgically obtained ordinal outcomes (Beck's scale of cartilage damage). Each hip in the study is assigned its own spatial dependence parameter, and a Dirichlet process prior distribution permits clustering of said parameters. Using the fitted model, we produce a six-color, patient-specific predictive map of the entire acetabular cartilage. Such maps will facilitate patient education and clinical decision making.

KEYWORDS:

Bayesian composite likelihood; Dirichlet process; T2*; areal model; arthroscopy; copula; hip; osteoarthritis; personalized medicine; proportional odds

PMID:
28088837
DOI:
10.1002/sim.7213
[Indexed for MEDLINE]

Supplemental Content

Loading ...
Support Center