Diagnostic value of 18F-FDG PET/CT in the follow-up of conjunctival melanoma.

OBJECTIVE

To evaluate the diagnostic performance of 18F-fluorodeoxyglucose positron emission tomography/computed tomography (18F-FDG PET/CT) in detecting metastatic conjunctival melanoma (CM).

METHODS

This retrospective study enrolled 67 patients with histopathologically confirmed CM who underwent 18F-FDG PET/CT for follow-up or suspected recurrence. Parameters including short-axis diameter of lesions, maximum standardized uptake value (SUVmax), the target-to-nontarget (T/NT) ratios of cervical lymph nodes, and bone lesion characterization (osteolytic, osteoblastic, or unchanged) were evaluated. Metastases were confirmed via either histopathology or ≥ 6-month imaging follow-up. Diagnostic sensitivity, specificity, accuracy, and metastatic patterns were analyzed at patient level.

RESULTS

A total of 16 patients were confirmed metastasis. The median interval from surgery to metastasis was 20.3 months (range 1-100 months), with 56.3% (9/16) occurring within the first postoperative year. PET/CT detected metastases in 13 patients, missed metastases in 3 patients (2 with small preauricular lymph nodes and 1 with tiny pulmonary metastases), and misdiagnosed 1 patient with parotid benign nodules as metastatic. PET/CT demonstrated a sensitivity of 81.3%, specificity of 98.0%, and accuracy of 94.0%. The most common metastatic sites included lymph nodes (62.5%), bone (37.5%), lung (31.3%), and liver (25.0%), with PET/CT demonstrating detection rates of 95.5%, 100%, 86.7%, and 100%, respectively. PET/CT also identified metastases in rare sites, including the thyroid, cerebellum, adrenal glands, pericardium, pancreas, and subcutaneous/soft tissue. Lymph node metastases in CM predominately occurred in ipsilateral regional nodes (90.1%), with rare bilateral involvement (9.1%). Metastatic lymph nodes in the cervical and submandibular regions showed significantly higher mean SUVmax (11.6 ± 10.5 vs. 2.9 ± 0.9; p = 0.005) and T/NT ratios (6.9 ± 8.2 vs. 3.0 ± 1.0; p = 0.011) compared to inflammatory lymph nodes. SUVmax of metastatic lymph nodes, lung metastases, and liver metastases showed positive correlations with lesion size (p < 0.05 for all). Among metastatic lymph nodes, 53.2% had a short-axis diameter < 10 mm, and 59.1% of bone metastases exhibited no abnormal CT density.

CONCLUSION

18F-FDG PET/CT provides high diagnostic accuracy in detecting systemic metastases of CM during follow-up assessments, particularly for small lymph nodes, early bone metastases, and uncommon sites.