| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Weill Medical College of Cornell University New York, NY 10021
In his Perspective on CT screening for lung cancer, Swensen [1] reports some remarkable results from his own ongoing study. I wish to express my counter perspective on two of his results and on his conclusion regarding the research that is now needed.
On the basis of the experience up to now, Swensen writes, "We expect that nearly 99% of all the lung nodules we identified were benign and therefore were false-positive findings." I ask, correspondingly, what is a positive finding and then what proportion of the screenees who had this positive finding had a false-positive finding on baseline screening for one and on repeated screening for another? Swensen answers only the first of these questions, but even that in a manner with which our Early Lung Cancer Action Project [2, 3] group vehemently disagrees. In particular, we disagree that in reasonable terms, the finding of "one or more uncalcified radiologically indeterminate nodules [however small]" constitutes a positive finding from the initial test at baseline. With the smallest nodules ignored, the frequency of false-positive findings is appreciably reduced without creating false-negative results. In fact, our understanding is that this is being done at the Mayo Clinic, and surprisingly Swensen presents results that are unjustifiably alarming. The concept of a positive test result must be confined to findings that justify diagnostic workup. Thus, the smallest nodules, which do not give rise to diagnostic workup, should not be viewed as positive test results; the patient merely returns for the scheduled cycle of routine screening.
Swensen writes that "eight of our patients have had surgery for removal of a benign nodule, approximately 20% of the operations that were performed." At the Mayo Clinic, what constitutes a presurgical rule-in diagnosis of lung cancer in the context of CT screening? In our ELCAP experience, the proportion is much lower.
Whether it is defining the nodules that need further diagnostic workup or the nodules that call for surgery, a rational approach, a well-defined algorithm that limits excessive diagnostic workups and unnecessary surgeries, is needed.
Swensen concludes that "a randomized, controlled trial is the best way to address the controversy..." [1]. I disagree. The issues Swensen raises are not at all addressed in a randomized controlled trial comparing screening with no screening. A random controlled trial provides no light on the issues related to the management of small nodules or on any of the other issues raised by Swensen. The answers he seeks come only from research on subjects who are actually undergoing screening.
References
Mayo Clinic, Rochester, MN 55905
Dr. Yankelevitz raised four interesting questions and comments to which I am pleased to respond.
"What is a positive finding?"
Our research team classifies all uncalcified lung nodules as positive findings on a CT screening test for lung cancer. That classification was the study design for our National Institutes of Health protocol and it is our practice today [1]. Although the cancer rate is low in the positive finding of 1- to 4-mm nodules, earlier detection offers the best prospect for a cure today. We are hoping to show in our research and in our care for patients that earlier detection makes a difference in outcome.
"What proportion of screenees who had this positive finding had a false-positive finding on baseline screening for one and on repeated screening for another?"
Fifty-one percent of participants had one or more lung nodules at baseline; 9–14% had new nodules at the three subsequent annual screening examinations. To date, 1.2% of baseline lung nodules identified have been shown to be lung cancer, and 4.3% of incident or new lung nodules (not present in retrospect) have been shown to be lung cancer. Approximately 98% of uncalcified lung nodules identified are false-positive results. This rate is a very high but not an insurmountable hurdle for a screening test.
In the setting of studying a new technology for lung cancer screening, we think it is inappropriate to ignore any lung nodules. In fact, five of the lung cancers that we have identified to date have been less than 5 mm in diameter. To ignore these smallest nodules would reduce the frequency of false-positive findings but, in doing so, would create false-negative findings and ignore the very cancers we are trying to detect—the earliest and smallest. So unlike your "understanding," both in our National Institutes of Health protocol and in our clinical practice today, we consider all radiologically indeterminate uncalcified lung nodules to be positive findings. We recommend follow-up or intervention for every lung nodule. It may be most appropriate to manage the smallest lung nodules with an annual follow-up. This is our current recommendation in both our National Institutes of Health protocol and in clinical practice.
"At the Mayo Clinic, what constitutes a presurgical rule-in diagnosis of lung cancer in the context of CT screening?"
To date, we have identified 40 patients with lung cancer. Eight participants underwent surgery for removal of benign nodules. That 17% rate (8/48) is dramatically lower than national averages from multicenter studies in both Europe and the United States of 46–52% [2, 3]. Five of those eight patients had nodules that increased in size, which is one of our criteria for high likelihood of lung cancer in a high-risk group. Although we do not rigidly protocol our workup of nodules, a large proportion of patients with nodules greater than 8 mm and those nodules seen as growing undergo positron emission tomography–CT before surgery for nodule diagnosis and staging. A smaller percentage undergo transthoracic needle aspiration biopsy. Most patients with radiologically indeterminate nodules are treated with radiologic follow-up. If nodules grow, a more aggressive mode of diagnostic workup that often leads to surgery is used. If nodules remain stable, we continue to observe them for at least 2 years. The problem remains: screening and subsequent radiologic follow-up will identify small nodules that are growing; and these nodules may be too small or in a location too difficult to evaluate further with CT enhancement, positron emission tomography-CT, or transthoracic needle biopsy.
"Swensen concludes that `a randomized controlled trial is the best way to address the controversy...' [1]. I disagree."
We are a participant in the recently launched National Lung Screening Trial that is funded by the National Institutes of Health. This trial has been well thought out. In the eyes of expert scientists at the National Institutes of Health and throughout the United States, it is an appropriate use of up to $200 million. It is designed to definitively answer the question of whether or not CT screening for lung cancer can reduce mortality rates from a deadly disease, lung cancer. Thank you for your inquiry and interest.
References
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
