TO THE EDITOR:
I read with great interest the research work by Crippa et al. (1). The report detailed clinical experience with 38 melanoma patients examined for nodal metastases by 18F-FDG PET. The purpose of the study was to establish the accuracy of 18F-FDG PET in the diagnosis of nodal metastases and to determine the smallest detectable volume of disease. The study population apparently comprised melanoma patients with documented (clinically known) lymph node metastases demonstrated by a variety of conventional techniques (physical examination, CT, or ultrasound). Patients underwent preoperative 18F-FDG PET followed by lymphadenectomy. PET images were interpreted in a masked fashion and were compared with lymph node histology. The study found a sensitivity of 95%, a specificity of 84%, and an overall accuracy of 91% for the detection of melanoma lymph node metastases. PET detected all metastases ≥ 10 mm in diameter, 83% of metastases 6–10 mm in diameter, and 23% of metastases ≤ 5 mm in diameter. The authors concluded that PET could detect small amounts of macroscopic disease but did not have acceptable sensitivity in the detection of microscopic disease.
The report of 18F-FDG PET sensitivity for detection of nodal metastases of various sizes is a useful contribution to the literature, and the data are consistent with multiple prior reports on 18F-FDG PET for melanoma staging. However, the authors conclude that their findings “encourage the use of FDG PET in melanoma patients with doubtful lymph node involvement.” This conclusion is not supported by the data because the population studied had clinically obvious lymph node metastases. The statement seems misleading and is probably incorrect if it is extrapolated to patients with clinically normal lymph nodes.
The performance of 18F-FDG PET as a diagnostic test will vary depending on the prevalence of detectable disease in the population to which it is applied. This principle is illustrated in a study (not referenced by the authors) addressing the use of PET for the primary staging of patients with clinically localized melanoma (2). In a prospective controlled study of 70 melanoma patients with cutaneous melanoma > 1.0 mm in Breslow depth localized to the skin, the Indiana University group found that PET had a sensitivity of 17% and sensitivity of 96%, compared with sentinel lymph node histology, in 89 nonpalpable lymph node basins. Half of all positive scans were false-positive, and false-negative scans were noted 5 times more often than true-positives. Furthermore, for no patient were the PET findings confirmed to show distant metastatic disease at the time of the scan.
The authors’ conclusions cannot be extrapolated to patients with clinically doubtful (i.e., nonpalpable) lymph nodes because these patients have a lower prevalence of detectable disease. We have reported typical lymph node tumor volumes that exist in melanoma patients who are staged by sentinel lymph node biopsy (3). About 20% of these patients harbor occult lymph node metastases. Although the range of tumor burdens encountered in this population is broad (0.1–3,618 mm3), the median tumor volume is only about 5 mm3. Sixty-five percent of all nodal tumor burdens observed were less than 10 mm3, and 73% were less than 30 mm3. By comparison, the spatial resolution of modern PET scanning equipment in common use today is about 5.5–6 mm, which corresponds to tumor volumes in the 87- to 113-mm3 range. This volume mismatch and the low prevalence of disease (especially disease within the range of PET detectability) largely explain the poor performance of 18F-FDG PET observed in patients with early-stage melanoma. Even with avid 18F-FDG uptake, most lymph node tumor nodules in this population are far too small to be visualized by today’s imaging technology.
To be defined for any clinical situation, the clinical utility of 18F-FDG PET must be prospectively studied in precise clinical scenarios. 18F-FDG PET is a useful staging tool for selected melanoma patients with recurrent or metastatic disease (American Joint Committee on Cancer [AJCC] stages III and IV). However, critical analysis of the collective PET literature reveals little evidence that 18F-FDG PET can be recommended for lymph node staging in clinically localized melanoma (AJCC stages I and II). With a sensitivity of at least 95%, sentinel node biopsy is the gold standard for nodal staging in early-stage melanoma. 18F-FDG PET should not be considered an acceptable alternative.