MRI patterns in orbital malignant lymphoma and atypical lymphocytic infiltrates.

OBJECTIVE

To evaluate the potential of MRI in differentiating between malignant lymphomas and atypical lymphocytic infiltrates in the orbit.

MATERIALS AND METHODS

MRI, clinical and histopathological findings in 30 patients presenting with orbital lymphoproliferative diseases (malignant lymphoma and atypical lymphocytic infiltrates) were evaluated.

RESULTS

MRI detected 28 out of 30 (93%) orbital lymphoid tumors. Seven out of eight (87.5%) atypical lymphocytic infiltrates and 21/22 (95.4%) of malignant lymphomas were detected. One conjunctival malignant lymphoma and one conjunctival atypical lymphocytic infiltrate were missed. Only malignant lymphoma lesions were hyperintense compared to the extraocular muscles on precontrast and postcontrast T1-weighted images. The lacrimal duct was involved only with malignant lymphoma. Bilateral disease was more likely to be malignant lymphoma. Intraconal lesions were more likely to be associated with lesions elsewhere. No malignant lymphoma involved the extraconal fat. Only intraconal atypical lymphocytic infiltrates had indistinct margins.

CONCLUSION

In the orbit, MRI features alone may not allow clear-cut differentiation of malignant lymphomas from atypical lymphocytic infiltrates. However, certain imaging features increase the likelihood of distinguishing them. MRI may miss conjunctival disease. Both orbits should be imaged when orbital lymphoid disease is a probability. The presence of intraconal disease should prompt a search for lesions elsewhere. The use of a head coil instead of a superficial coil may be advantageous by eliminating coil shine effect and allowing evaluation of both orbits. Comparison of lesion signal to extraocular muscle signal appears to be a better alternative than cerebral gray matter or periorbital fat in differentiating malignant lymphoma from atypical lymphocytic infiltrates. Since imaging findings overlap, histopathological diagnosis is necessary in all cases.