| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Letters to the Editor |
Tristán Associates, 4518 Union Deposit Road, Harrisburg, PA 17111. e-mail: mguenin@tristans.com
Editor:
I found myself smiling as I read the recent article by Dr Sickles and colleagues in the September issue of Radiology (1) in which the performance of general radiologists in their group was found to be, shall we say, suboptimal compared with that of the breast imaging specialists in their group with regard to interpretation of mammograms. I’ll bet the article fostered much peace and harmony within the group.
The authors did not consider what I believe to be a critical element—namely, hands-on experience in performing stereotactic and ultrasonographically (US) guided breast biopsy. It has been my experience (and anecdotally the experience of radiologists with whom I’ve discussed the issue) that performing biopsy of the breast sharpens one’s mammography and breast US interpretive skills. Specifically, the process of focusing on a lesion to perform a biopsy, critically assessing its likelihood of malignancy to judge concordance or discordance, and subsequently obtaining the near-immediate feedback of tissue diagnosis the next day is more powerful reinforcement than is viewing a series of cases at a radiology-pathology correlation conference. I would venture that high-volume experience in breast biopsy narrows or eliminates the gap in performance between the two groups of radiologists. Do the authors have any data to support or refute this?
Second, I would hesitate to make generalizations from the observations of this single group of radiologists with regard to the population of radiologists as a whole. According to table 1 in the article (1), radiologist A, a breast imaging specialist, interpreted 72% (36,196 of 50,489) of the total number of mammograms interpreted by all breast imaging specialists in that group. The authors’ conclusion that "specialist radiologists detect more cancers and more early-stage cancers, recommend more biopsies, and have lower recall rates than general radiologists" seems overly broad and far reaching, given that most of the data reflect the skills of a single radiologist. Furthermore, if radiologist A happens to be Dr Sickles, a widely published and world-renowned breast imaging specialist, then the assumption that his level of performance will be matched by all other breast imaging specialists seems even less plausible.
I am frankly a bit puzzled by the reason for this article in the first place. Articles like this are not published in a vacuum. One cannot pretend that publishing such a study will not have implications. What was the larger point behind the article? The authors make no recommendation to remedy the situation other than to suggest further study. Should only breast imaging specialists be allowed to interpret mammograms? Radiologists are not exactly tripping over themselves to read more mammograms; recruiting residents for breast imaging fellowships is difficult, and many areas are underserved by breast imaging services (2). Any attempt to restrict the population of radiologists deemed worthy of interpreting mammograms will ultimately exacerbate those shortages. That would certainly make the early detection rate plummet.
In summary, the field of radiology is not theoretical mathematics or physics; publications should not merely serve to increase knowledge but should instead be of practical value to our patients. Other than being divisive and a bit self-congratulatory, I’m not sure what purpose this article has served.
REFERENCES
Department of Radiology, University of California, San Francisco School of Medicine, Box 1667, San Francisco, CA 94143-1667
*Current address: Department of Radiology, University of Washington Medical Center, Seattle.
We thank Dr Guenin for giving us the opportunity to emphasize some of the major points of our article (1) and to comment on the possible clinical implications that it may have.
We have no data on the specific contribution to radiologist performance in mammography interpretation of experience in performing and supervising stereotactic and US-guided breast biopsy. Although this particular type of experience was shared by all of our specialist radiologists and by none of our general diagnostic radiologists, so were the several other types of advanced training and experience described in our article (breast imaging fellowship training and teaching; subspecialization in breast imaging; interpretation of more than 5,000 mammography examinations per year, at least six times the amount of federally required continuing medical education in mammography; and regular participation in a weekly radiology-pathology correlation conference). We share Dr Guenin’s opinion that experience in performance and supervision of percutaneous breast biopsy provides useful feedback to the radiologist.
Furthermore, to the extent that caseload is important, we agree that high-volume experience with percutaneous breast biopsy will indeed contribute to improved individual performance, as long as the radiologist also maintains high-volume experience in interpretation of screening and diagnostic mammograms. However, our data do not permit us to attribute the improved performance that we observed for specialists over generalists to any one particular type of advanced training or experience. Actually, we suspect that a combination of factors, involving caseload, outcomes feedback, and perhaps others, contributed to our results.
With regard to Dr Guenin’s specific comments on percutaneous biopsy experience, we propose a theoretic advantage to regular participation in a practice-wide radiology-pathology correlation conference over simple hands-on experience: the ability to learn from the outcomes of all interventional cases in the practice (including preoperative wire localizations), rather than those limited to the individual radiologist. We have found the learning experience to be enhanced by presenting each case as an "unknown," in effect asking all participating radiologists to estimate the likelihood of malignancy before the pathologic diagnosis is revealed. This converts passive learning into a more active and participatory process.
Dr Guenin notes correctly that more than two-thirds of the mammograms in our study were interpreted by one specialist radiologist. He then questions whether it is appropriate to make general conclusions when most of the data are based on the performance of a single radiologist. However, careful review of all the tables in our article clearly shows that the improved performance of our highest-volume radiologist was shared to a remarkably equal degree by our other two specialist radiologists, each of whom contributed more than 4,500 interpretations to the study. It is this triplication of improved performance (in contrast to different results from each of seven general diagnostic radiologists) that prompted us to generalize our conclusions. Thus, one should consider the near identical performance of each of our three specialists, which was significantly superior to the similar performance of each of our generalists (except as affected by multiple readings), to be important evidence that justifies the grouping of our results. We urge other practices that conduct comprehensive mammography audits to publish on this subject to provide even more convincing evidence that experience and expertise do affect performance in mammogram interpretation substantially.
Finally, Dr Guenin muses on our motives in publishing the article. He is correct in concluding that our results may have far-reaching implications, especially if they are corroborated by others. He should also be aware that there have been several other reports documenting considerable variability in radiologist performance in interpreting mammograms (2–7), and that those who oppose the widespread use of screening mammography are beginning to cite these articles in support of their position. Rather than ignore the issue, we have chosen to address it directly. We believe that our results provide a preliminary indication of the magnitude of the difference in performance at the two ends of the spectrum of radiologist experience and expertise in interpreting mammograms (breast imaging specialists who received training or who train fellows versus general diagnostic radiologists who meet the minimum requirements mandated by federal regulation).
We also recognize the important need for reliable data involving radiologists in between the two ends of this wide spectrum and hope that our article will spur other mammography practices with appropriate mixes of experience and expertise to publish results to fill in the gap. One must also take note that the performance of our generalist radiologists, although significantly inferior to that of our specialists, was well within the performance guidelines cited by the American College of Radiology in its Breast Imaging Reporting and Data System, or BI-RADS, atlas (8,9). Thus, we do not show that current federal regulations permit substandard mammography interpretive performance by "setting the bar" too low, but rather that it may be possible to raise the bar by concentrating the mammography workload among those radiologists who have the most experience and expertise.
Dr Guenin is correct in describing the current shortage in breast imaging specialists in the United States and elsewhere as a limiting factor. However, we prefer to "push the envelope" by challenging the status quo that relies heavily on a single reading of mammograms by generalists who meet minimum training and/or experience requirements. If convincing evidence is produced that increased experience and expertise significantly improves interpretive performance, especially if we can identify a specific point to which the bar should be raised, then we might reasonably expect support from the general public (and the government) to provide the incentives needed to eliminate the current shortage in breast imaging specialists.
It would be counterproductive to raise the bar without first eliminating the shortage, but efforts to eliminate the shortage appear to require more than just the collective say-so of most radiologists. Hence, the ultimate rationale behind publishing our article was to contribute to the process of identifying a convincing reason (scientific evidence) to justify widespread support for overcoming the shortage to achieve the final goal of improved mammographic interpretive services. We believe that Jane (and John) Q. Public will strongly support this approach.
Are there any conclusions that are immediately applicable to the individual mammography practice that can be derived from the data in our article? One obvious conclusion is that practices should consider the advantages of improved overall performance in deciding the degree to which they concentrate their mammography workload among those radiologists in the group who have the highest levels of performance, as shown with periodic auditing. At the University of California at San Francisco, we have recently taken this decision to the extreme by restricting mammography interpretation to our breast imaging specialists. However, this extreme approach likely will not be applicable to almost all other practices for which outcomes may well be different and for which there may not be the luxury of already having a critical mass of breast imaging specialists on staff.
REFERENCES
This article has been cited by other articles:
|
|
 |
|
  J. W. T. Leung, F. R. Margolin, K. E. Dee, R. P. Jacobs, S. R. Denny, and John. D. Schrumpf Performance Parameters for Screening and Diagnostic Mammography in a Community Practice: Are There Differences Between Specialists and General Radiologists? Am. J. Roentgenol., January 1, 2007; 188(1): 236 - 241. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. A. Sickles, D. L. Miglioretti, R. Ballard-Barbash, B. M. Geller, J. W. T. Leung, R. D. Rosenberg, R. Smith-Bindman, and B. C. Yankaskas Performance Benchmarks for Diagnostic Mammography Radiology, June 1, 2005; 235(3): 775 - 790. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
