HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Services Email this article to a friend Similar articles in this journal Alert me to new issues of the journal Download to citation manager Google Scholar Articles by Bonta, D. Articles by Wahl, R. PubMed Articles by Bonta, D. Articles by Wahl, R.

Is a "yes"-"no" answer adequate in reporting nuclear medicine diagnostic imaging results?

¹ Radiology, Johns Hopkins Medical Institutions, Baltimore, Maryland

1162

Objectives: To (1) illustrate the limitations of 2 category (positive/negative) reporting when a binary diagnostic question is posed, (2) propose a 5 category reporting scale to address the weaknesses of 2 category reporting, and (3) apply and compare the 2 category and the 5 category scales for a sample of PET imaging results.

Methods: "The tangent at a point on the ROC curve corresponds to the likelihood ratio (LR) for a single test value represented by that point" (Choi 1998). Thus, 2 category reporting: 1: Assigns patients with findings adjacent to the cutoff point on the ROC curve to opposite categories, despite nearly identical likelihood ratios (LRs) and 2: Mixes patients with findings corresponding to a (possibly) wide range of LRs in the same reporting category. We propose a 5 category scale to report the results of a binary diagnostic inquiry: 1. Definitely Positive (LR~10), 2. Probably Positive (LR~3), 3. Equivocal (LR~1), 4. Probably Negative (LR~1/3), 5. Definitely Negative (LR~1/10). We have applied such an approach to clinical data interpretation.

Results: We used the data from a clinical study of PET detection of axillary metastases of breast cancer. Five category reporting recovered 96% of the area between the ROC curve and the main diagonal (the line equivalent to guessing), while 2 category reporting recovered a maximum of 76% of the same area.

Conclusions: The proposed 5 category scale informs the referring physician whether the test results supported or dismissed the diagnosis sought, whether the evidence is strong or weak, or whether test results were non-informative. This scale addresses obvious weaknesses of the 2 category reporting paradigm and suggests that multiple category reporting, or perhaps a continous read out, may be more informative.

This Article Services Email this article to a friend Similar articles in this journal Alert me to new issues of the journal Download to citation manager Google Scholar Articles by Bonta, D. Articles by Wahl, R. PubMed Articles by Bonta, D. Articles by Wahl, R.