Preoperative staging of liver metastases from uveal melanoma by magnetic resonance imaging (MRI) and fluorodeoxyglucose-positron emission tomography (FDG-PET)

https://doi.org/10.1016/j.ejso.2009.08.010Get rights and content

Abstract

Background

Microscopically complete (R0) resection of metastases from uveal melanoma prolongs median overall survival compared to incomplete surgery. The aim of this study was to compare the sensitivity of dynamic-enhanced magnetic resonance imaging (MRI) with fluorodeoxyglucose-positron emission tomography (FDG-PET) in the preoperative diagnosis of liver metastases from uveal melanoma.

Patients and methods

Fifteen consecutive patients (mean age: 56 years) underwent FDG-PET and liver MRI. Extrahepatic metastatic disease was excluded by whole body computed tomography and bone scintigraphy. MRI and FDG-PET were performed with a mean of 19 days (range: 1–30) before surgery. Imaging findings were compared with surgical (including intraoperative ultrasonography) and histological findings on a lesion by lesion analysis.

Results

R0 resection was performed in 12 patients. A total of 28 lesions were resected with 27 histologically proven metastases. Nine lesions were smaller than 5 mm, 7 measured 5–10 mm and 11 were larger than 10 mm. Sensitivity and positive predictive value were 67% and 95% for MRI compared to 41% and 100% for FDG-PET. The difference between the two modalities was statistically significant (p = 0.01; McNemar test). In remaining 3 patients, diffuse miliary disease (>10 capsular lesions) was discovered intraoperatively, and was suspected on preoperative MRI in 2 cases. Only one extrahepatic lesion identified by FDG-PET was falsely positive.

Conclusions

In this preliminary study, MRI was superior to FDG-PET for staging of liver metastases from uveal melanoma. Although miliary disease was suggested by MRI in some cases, preoperative confirmation remains imperfect.

Introduction

Uveal melanoma is the most frequent primary malignant intraocular tumour in adults; 500 to 600 new cases are diagnosed each year in France.1 Metastatic spread of uveal melanoma is observed in 50% of patients and the liver is the almost exclusive site of metastases after a mean interval of 3 years.2 The recommended hepatic surveillance after treatment of intraocular tumour comprises liver ultrasound (US), every 6 months for 10 years and liver computed tomography (CT) scan and/or liver magnetic resonance imaging (MRI) when an abnormality is detected on liver ultrasound. Following the detection of metastases, the median overall survival without treatment ranges from 2 to 6 months according to the study.3 The early diagnosis of liver metastases may allow R0 (microscopically complete) surgical resection in patients with limited metastatic extension. In the experience reported by the Institut Curie in 1998, based on 75 operated patients, survival was 22 months in the case of R0 resection versus 9 months in the case of R1 or R2 incomplete resection.4 These data were updated in 2009 in a series of 224 patients confirming a complete resection rate of 28% and a median survival of 22 months for R0 resected patients (13 months for the overall cohort).5 The main limitation to R0 surgery is the intraoperative discovery of miliary liver metastases that makes R0 resection impossible.

Liver imaging methods classically comprise US, CT, MRI and F[18] fluoro-desoxyglucose (FDG) positron emission tomography (FDG-PET). MRI is the investigation technique of choice, as it presents a sensitivity of detection at least equivalent to that of computed tomography with better specificity.6 In the particular diagnosis of uveal melanoma, the presence of melanin in metastatic lesions induces shortening of T1 and T2 relaxation times.7, 8 Typical lesions therefore have a high-intensity appearance on T1-weighted sequences. The hypervascular nature of most metastatic lesions also allows improved detection and characterization of abnormalities on T1-weighted dynamic sequences after injection of a gadolinium chelate.

Although FDG-PET has been evaluated in the metastatic staging of cutaneous malignant melanoma.9, 10 Few studies have evaluated FDG-PET in uveal melanoma.11, 12, 13 This whole body imaging modality presents the theoretical advantage of evaluating liver involvement while also detecting any extrahepatic one. To our knowledge, no study has compared the contribution of MRI and FDG-PET in the preoperative assessment of liver metastases from uveal melanoma.

The objective of this study was therefore to compare the performances of MRI and FDG-PET in the preoperative assessment of liver metastases from uveal melanoma by comparing them with intraoperative staging and histological examination.

Section snippets

Population

Between November 2006 and March 2009, study was conducted according to our institution's ethical guidelines and was approved by the melanoma group. Fifteen patients, 7 males, 8 females with a mean age of 56 years (range: 38–71 years) were investigated by MRI and FDG-PET for staging of preoperative assessment of liver metastases.

All patients were suspected to present liver metastases following the discovery of one or more lesions on surveillance ultrasound. Overall staging comprising

Surgical exploration

A total of 28 lesions were resected in 12 of the 15 operated patients. The number of lesions resected ranged between 1 and 6, i.e. a mean of 2 lesions per patient. All resected lesions corresponded histologically to metastases from uveal melanoma except for one lesion corresponding to an eosinophil abscess. Resection was considered to be R0 for these 12 patients.

Nine of the 27 metastatic lesions resected were smaller than 5 mm, 7 measured between 5 and 10 mm and 11 were larger than 10 mm.

For 3

Methodological aspects of the study and PET study

In our institution, the metastatic work-up in patients suffering from uveal melanoma consists of total body CT, bone scintigraphy, liver MRI. The aim of this study was to compare the performances of the last versus FDG-PET scan. For this reason we decided to add to the metastatic work-up metabolic imaging by PET scan.

The subset of patients in this study presents a highly selected group because of the surgical indications limiting number of detected metastasis less than five.5 In all 15 patients

Conclusion

In this preliminary series, we showed that MRI was superior to FDG-PET for the staging of liver metastases from uveal melanoma. This difference was particularly marked for lesions measuring less than 10 mm. Although MRI has a poor sensitivity for the detection of lesions smaller than 5 mm, it appears to be the only imaging modality sometimes able to suggest the presence of miliary metastatic disease. It is important to detect this particular form of metastases, as it compromises R0 resection and

Conflict of interest

The authors have no conflict of interest.

References (22)

  • A. Premkumar et al.

    Metastatic melanoma: correlation of MRI characteristics and histopathology

    J Magn Reson Imaging

    (1996)
  • Cited by (0)

    View full text