Reader Variability: What We Can Learn From Computer-Aided Detection Experiments

Freedman, Matthew T.; Osicka, Teresa

doi:10.1016/j.jacr.2006.02.025

Cited by 18 publications

(9 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In 1998, the US Food and Drug Administration approved computed assisted detection (CAD) to assist radiologists in their interpretation of radiologic examinations. Since that time, CAD has been used primarily in mammography [68][69][70], where it has improved sensitivity by raising the level of the radiologist's suspicion for breast cancer. CAD was initially shown to be of value in reducing radiologic errors and improving interpretation in mammography [71,72].…”

Section: Computed Assisted Detection (Cad)mentioning

confidence: 99%

Radiologic errors, past, present and future

Berlin¹

2014

Diagnosis

View full text Add to dashboard Cite

During the 10-year period beginning in 1949 with publication of five articles in two radiology journals and UKs The Lancet, a California radiologist named L.H. Garland almost single-handedly shocked the entire medical and especially the radiologic community. He focused their attention on the fact now known and accepted by all, but at that time not previously recognized and acknowledged only with great reluctance, that a substantial degree of observer error was prevalent in radiologic interpretation. In the more than half-century that followed, Garland's pioneering work has been affirmed and reaffirmed by numerous researchers. Retrospective studies disclosed then and still disclose today that diagnostic errors in radiologic interpretations of plain radiographic (as well as CT, MR, ultrasound, and radionuclide) images hover in the 30% range, not too dissimilar to the error rates in clinical medicine. Seventy percent of these errors are perceptual in nature, i.e., the radiologist does not "see" the abnormality on the imaging exam, perhaps due to poor conspicuity, satisfaction of search, or simply the "inexplicable psycho-visual phenomena of human perception." The remainder are cognitive errors: the radiologist sees an abnormality but fails to render a correct diagnoses by attaching the wrong significance to what is seen, perhaps due to inadequate knowledge, or an alliterative or judgmental error. Computer-assisted detection (CAD), a technology that for the past two decades has been utilized primarily in mammographic interpretation, increases sensitivity but at the same time decreases specificity; whether it reduces errors is debatable. Efforts to reduce diagnostic radiological errors continue, but the degree to which they will be successful remains to be determined.

show abstract

Section: Computed Assisted Detection (Cad)mentioning

confidence: 99%

Radiologic errors, past, present and future

Berlin¹

2014

Diagnosis

View full text Add to dashboard Cite

show abstract

“…Studies using computer-aided detection systems have shown a significant degree of both intra- and inter-reader variability in disease detection by computed tomography, and it is in that sense that variability in interpretation has been chiefly analyzed. 26 Lessons learned from breast density estimation in digital and screen film mammography suggest that optimization of a protocol, regardless of the chosen technique, is mandatory to establish uniform measuring guidelines and criteria. 25 Variability in radiotracer injection techniques and sites translates into variability among image interpretations.…”

Section: Discussionmentioning

confidence: 99%

Intra-Individual Comparison of Lymphatic Drainage Patterns Using Subareolar and Peritumoral Isotope Injection for Breast Cancer

et al. 2009

View full text Add to dashboard Cite

Background Controversy exists in the literature regarding the optimal site for lymphatic mapping in breast cancer. This study was designed to characterize lymphatic drainage patterns within the same patient after subareolar (SA) and peritumoral (PT) radiopharmaceutical injections and examine the impact of reader interpretation on reported drainage. Methods In this prospective trial, 27 women with breast cancer underwent sequential preoperative SA and PT injections of 0.5 to 2.7 mCi of technetium-99 m filtered sulfur colloid 3 days or more apart. Patterns of radiopharmaceutical uptake were reviewed independently by two nuclear medicine physicians. Inter-reader agreement and injection success were assessed in conjunction with observed drainage patterns. Results There was near perfect inter-reader agreement observed on identification of axillary LN drainage after PT injection (P = 0.0004) and substantial agreement with SA injection (P = 0.0344). SA injection was more likely to drain to only axillary LNs, whereas PT injection appeared more likely to drain to both axillary and extra-axillary LNs, although no statistically significant differences were found. All patients with extra-axillary drainage after PT injection (n = 6 patients) had only axillary drainage after SA injection. Dual drainage was observed for six patients with PT injection and one patient with SA injection. Conclusions Our findings suggest that radiopharmaceutical injected in the SA location has a high propensity to drain to axillary LNs only. After controlling for patient factors and demonstrating inter-reader agreement, the inability to demonstrate statistically significant differences in drainage based on injection site suggests that lymphatic drainage patterns may be a function of patient and tumorspecific features.

show abstract

“…In radiology, a number of computer analysis and interpretation tools have been approved by the Food and Drug Administration and are being used clinically [90–94]. Similar tools are being developed for use with virtual pathology images and are likely to be used just as extensively to facilitate navigation through large data sets, generate “second opinion” diagnoses, and improve workflow, reader efficiency, and reader accuracy [95–98].…”

Section: Decision Support Toolsmentioning

confidence: 99%

Virtual slide telepathology workstation of the future: lessons learned from teleradiology

Krupiński

2009

Human Pathology

View full text Add to dashboard Cite

Summary The clinical reading environment for the 21st century pathologist looks very different than it did even a few short years ago. Glass slides are quickly being replaced by digital “virtual slides,” and the traditional light microscope is being replaced by the computer display. There are numerous questions that arise however when deciding exactly what this new digital display viewing environment will be like. Choosing a workstation for daily use in the interpretation of digital pathology images can be a very daunting task. Radiology went digital nearly 20 years ago and faced many of the same challenges so there are lessons to be learned from these experiences. One major lesson is that there is no “one size fits all” workstation so users must consider a variety of factors when choosing a workstation. In this article, we summarize some of the potentially critical elements in a pathology workstation and the characteristics one should be aware of and look for in the selection of one. Issues pertaining to both hardware and software aspects of medical workstations will be reviewed particularly as they may impact the interpretation process.

show abstract

Reader Variability: What We Can Learn From Computer-Aided Detection Experiments

Cited by 18 publications

References 8 publications

Radiologic errors, past, present and future

Radiologic errors, past, present and future

Intra-Individual Comparison of Lymphatic Drainage Patterns Using Subareolar and Peritumoral Isotope Injection for Breast Cancer

Virtual slide telepathology workstation of the future: lessons learned from teleradiology

Contact Info

Product

Resources

About