| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
EDITORIALS |
From the Department of Plastic Surgery, Indiana University School of Medicine, Indianapolis, Indiana.
Correspondence: Address correspondence to: Jeffrey D. Wagner, MD, Wagner and Associates Plastic and Reconstructive Surgery Consultants, Suite 570, 8040 Clearvista Parkway, Indianapolis, IN 46256; Fax: (317) 621-2580; e-mail: jdwagner{at}insightbb.com
Identification of melanoma patients with occult metastases is important for accurate staging and treatment planning in melanoma. This is particularly true when major therapeutic interventions, such as metastatectomy, are considered for patients with advanced disease (stage IV). Accurate identification of all sites of metastatic disease is critical, as only patients rendered completely free of disease are likely to enjoy significant disease-free intervals or a potential survival benefit from surgery.
Metabolic imaging with fluorodeoxyglucose positron emission tomography (FDG-PET) has recently become more widely applied in a variety of human malignancies. Several studies have reported FDG-PET to be a sensitive indicator of metastatic melanoma in comparison with conventional diagnostic imaging modalities,1–5 particularly in patients with recurrent disease. However, design flaws in early studies may have led to overestimation of the diagnostic accuracy of PET in this disease.6 Major methodological shortcomings included retrospective methods, lack of a valid reference test, unblinded comparisons, poor description of disease stage, use of repetitive PET scans, small sample size, lack of histopathological verification of PET findings, comparison with conventional imaging tests with worse sensitivity than PET, and inadequate clinical follow-up to confirm negative findings. These shortcomings have made it difficult to develop guidelines for the effective use of PET in patients with melanoma.
Literature shortcomings notwithstanding, there is mounting clinical evidence that FDG-PET is clinically useful in patients with advanced (stage III and IV) melanoma. Investigators in several series have reported that FDG-PET demonstrates better disease detection than conventional imaging in patients with advanced melanoma, prompting changes in clinical management decisions.7–9 In contrast, critical analysis of the collective literature demonstrates that FDG-PET is a relatively insensitive indicator of occult regional lymph node metastases in early stage melanoma.10–11 Routine FDG-PET staging/screening at the time of initial disease presentation does not appear to have a significant clinical impact on the management of patients with stage I and II melanoma.
PET sensitivity for tumor detection has been demonstrated to be dependent on a number of clinical factors, including tumor burden.12 Tumor burdens of at least 80 mm3 are required for sensitive (90%) detection by modern PET scanning equipment. Tumor burden studies show that patients with stage III or IV melanoma are more likely to have PET-detectable disease than those with early stage disease. This is the likely explanation for differences in sensitivity and specificity in the literature. PET staging at the time of recurrence or in patients with stage III or IV disease is more likely to impact clinical management algorithms because of the higher prevalence of PET-detectable tumor deposits in this population.
In this issue of Annals of Surgical Oncology, Finkelstein and colleagues report their prospective evaluation of FDG-PET for evaluation of 18 stage IV melanoma patients prior to surgical resection of stage IV melanoma.13 This thoughtful study overcomes some of the flaws of prior PET melanoma studies. While this report shows that FDG-PET imaging has similar sensitivity and specificity to conventional imaging modalities in advanced melanoma, the combination of PET and conventional imaging is more sensitive and specific than either modality alone. The authors make a rational and intuitively compelling case for use of both PET and computed tomography (CT) in the clinical management of advanced melanoma prior to aggressive surgery for metastatic disease. This study further defines the role of PET imaging in melanoma management.
Studies of this nature are very difficult to design, and bias is hard to avoid. True sensitivity and specificity are elusive because comparison to clinical follow-up and histology must be relied upon. It is difficult to verify many (i.e., brain and visceral) lesions histologically. Furthermore, differences in follow-up regimens and therapy can contribute to apparent differences in diagnostic performance. Follow-up also implies use of conventional imaging, known to be relatively unsatisfactory in detection of early disease.
In the current study, all patients had immunotherapy and other therapies, and many responded. FDG-PET is known to have decreased sensitivity in even large lesions after therapy. This could have accounted for increased false-negatives or even true-negatives in lesions that did not progress and were not histologically verified. This almost unavoidable design flaw of PET studies in patients with advanced disease is difficult to account for and could inflate the observed accuracy of PET.
Another potential source of bias is that all clinical information was available for conventional image review, whereas for PET review, the radiologist was blinded. Differences in interpreter blinding could possibly inflate the performance of conventional and or combination imaging in this study. Another problem with studies of this type is the somewhat subjective nature of PET interpretations. This is really important when one considers the problem of false-positives in PET imaging. The impact of the interpreter’s threshold for determining malignancy upon PET performance is best demonstrated by use of receiver operator curves. Last, if duration of clinical follow-up is inadequate, the true-negative rate will be overestimated and artificially inflate measures of diagnostic performance.
Perhaps the most important contribution of this work is the clear demonstration that PET and conventional imaging really are complementary, not competitive modalities. The weaknesses of metabolic imaging are improved by the strengths of anatomically based conventional imaging, and vice versa. A rational case can be made for using both staging modalities to help select patients with advanced melanoma for aggressive surgical therapy. This study helps to lay the groundwork for further prospective studies of new technologies such as PET-CT.
The next generation of imaging studies should move beyond simple measures of sensitivity and specificity, which are not always accurate measures of impact, and focus on cost-benefit and cost-effectiveness of clinical decision-making. Such studies must be prospective in nature, designed to involve homogeneous populations of patients with standardized treatment and follow-up regimens, to allow unbiased comparison of disease-free intervals/disease progression and cost analysis in a predetermined clinical decision-making schema. These types of data are necessary to determine the true utility of FDG-PET in melanoma.
Received for publication June 11, 2004. Accepted for publication June 18, 2004.
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
