| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Guest Access | Sign In via User Name/Password
|
|
|
|||||||
Guest Access | Sign In via User Name/Password |
|||||||||
Hospital Vega Baja Orihuela-Alicante, Spain
Correspondence to: Eduardo Garcia-Pachon, MD, Department of Internal Medicine, Hospital Vega Baja, E-03314 Orihuela-Alicante, Spain; e-mail: egpachon{at}latinmail.com
To the Editor:
In the April 2001 issue of CHEST, Paganuzzi and colleagues1 analyzed two tumor markers (carcinoembryonic antigen [CEA], and CYFRA 21–1) in pleural fluid and concluded that elevated levels of these tumor markers suggest a diagnosis of malignant pleural effusion, and in patients with poor clinical conditions, diagnosis should be made on the basis of tumor markers alone. Can this conclusion be obtained from their results? I do not think so. In this series, CEA and CYFRA 21–1 sensitivity was 31% and 78%, respectively, and specificity was 91% and 80%. The possibility that a patient has a malignant pleural effusion if the tumor marker is "positive" depends on the pretest probability (prevalence). In that study, performed at least in part in a cancer center, the prevalence of malignancy was 68%. However, the prevalence in general hospitals is, by far, lower.2 3 4 For instance, in an epidemiologic study,2 the prevalence of effusions associated with malignancy was 24%, and in 273 consecutive patients studied by our group in a department of internal medicine,4 the prevalence was 33%. Because of the fact that in about half of the patients with malignant pleural effusion the first cytologic study finding is positive (consequently, tumor markers are not necessary), the true prevalence (effusions with a negative cytologic study finding) is still lower. In Table 1 , the posttest possibility of malignancy for several percentages of prevalence and for both tumor markers has been calculated. As can be observed, with a high prevalence the result should be considered somewhat suggestive but not diagnostic of malignancy (probability 77% for CEA, and 80% for CYFRA 21–1), and with the actual pretest probabilities in general clinical practice, the probability is purely by chance (for example, with prevalence 20% and a "positive" test finding, the probability of malignancy is 46% and 49%, respectively). Moreover, clinical data are, frequently, enough for suspecting malignancy, and in patients with pleural effusion of unknown cause malignancy is not frequent.5 Whether tumor markers are useful in this subgroup of patients (undiagnosed and without clinical suspicion of a malignant cause) needs to be demonstrated. In conclusion, tumor markers are not useful for diagnosing a malignant pleural effusion, and is a test necessary for suggesting malignancy when, in most cases, it is clinically easy to suspect?
|
References
* National Institute for Cancer Research Genoa, Italy
Correspondence to: Michela Paganuzzi, PhD, Clinical Pathology Laboratory, National Institute for Cancer Research, Largo R. Benzi 10, Genoa 16132, Italy; e-mail: patclin{at}hp380.ist.unige.it 
To the Editor:
The data published in CHEST, April 2001, were focused on mesothelioma. We and other authors believe that tumor markers may be especially useful in mesothelioma patients with poor clinical conditions, when cytology is negative or uncertain and when the performance status does not suggest a more aggressive approach.
We did not say, as reported by Pachon, that "elevated levels of CEA and CYFRA 21-1 suggest a diagnosis of malignant pleural effusions," but we said that "high CYFRA 21-1 levels with low CEA levels are suggestive of mesothelioma, whereas high levels of CEA alone or of both markers can show evidence of malignant pleural effusion, excluding mesothelioma." In our study, CYFRA 21-1 sensitivity in 19 cases of mesothelioma with negative or uncertain cytology was 89%, with no cases showing CEA values higher than the cut-off level.
Furthermore, we think that the prevalence of effusions associated with malignancy is related to different factors (incidence of tuberculosis, pollution exposure, features of the populations, and so on). For instance, in our geographic area there is intensive occupational and environmental exposure to asbestos and to other carcinogenic substances. Consequently, the incidence of malignant pleural effusion is greater than in other areas, like the region of Central Bohemia cited by Pachon.
We don’t agree with the opinion of Dr. Pachon that "clinical data are, frequently, enough for suspecting malignancy." In our experience, despite conditions of general well-being, some patients developed pleural effusion with clinical and biochemical characteristics of a parapneumonia pleural effusion. In similar cases, the increased tumor marker values could induce investigators to continue the diagnostic titer.
We thank Dr. Garcia-Pachon for his interest and hope for a future collaboration.
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
