Abstract Archives of the RSNA, 2014
Saurabh Jha MD, Presenter: Speaker, Toshiba Corporation
Stefan L. Zimmerman MD, Abstract Co-Author: Nothing to Disclose
Tessa S. Cook MD, PhD, Abstract Co-Author: Nothing to Disclose
Arrhythmogenic Right Ventricular Dysplasia (ARVD) is uncommon, but can be fatal if undetected and the
management, placement of an Implantable Cardioverter Defibrillator (ICD), is not trivial. The major and
minor diagnostic criteria for ARVD include findings on cardiac MRI (CMR). The potential for overtreatment
in low prevalence populations is explored.
Bayes’ inversion tree was constructed to explore the burden of unnecessary ICDs in a hypothetical cohort
of males with electrophysiological abnormalities suspected to have ARVD.
Test characteristics of CMR for diagnosis of ARVD were abstracted from literature.
Positive CMR was assumed to trigger placement of ICD; the assumption relaxed during sensitivity analysis.
The prevalence of ARVD was varied between 0.5 % and 20 %.
Tradeoff between the incremental cases of ARVD detected and unnecessary ICDs placed when using
minor over major criteria, was explored.
Sensitivity and specificity of CMR findings, for the detection of ARVD in males, representing major and
minor criteria are 76 % and 90 % and 79 % and 85 %, respectively.
The model postulates overtreatment when diagnosis is based on CMR findings. At prevalence of 5 %, use of
major criteria results in five inappropriate ICD placements for two appropriate ICDs.
In a cohort of 10, 000 males suspected of ARVD use of minor instead of major criteria detects more cases
of ARVD at a cost: at prior probability of one percent, 3 more cases of ARVD are diagnosed at the expense
of 495 additional patients receiving unnecessary ICDs; at five percent, an additional 15 cases are detected
and 475 additional ICDs unnecessarily placed; even at prior probability of twenty percent, the incremental
detection of 60 cases of ARVD comes at the price of unnecessary ICD placement in an additional 400
patients.
The specter of sudden death may lead clinicians to lower their threshold for suspicion of ARVD and request CMR for exclusion of ARVD in patients at low probability of ARVD. Imagers must be aware of the potential for overtreatment when using taskforce guidelines to rule out rather than rule in ARVD.
Overtreatment is a recognized problem in modern medicine particularly when attempting to diagnose uncommon but dangerous conditions with imperfect tests.
Jha, S,
Zimmerman, S,
Cook, T,
Using Decision Analysis to Explore Potential Overtreatment of ARVD with ICD in Low Prevalence Population. Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL.
http://archive.rsna.org/2014/14014833.html