Yesterday I wrote a short comment in the utility of DWI in the evaluation of head and neck lesions. While reading cases in the afternoon I came across a new patient with a retinoblastoma. This patient had DWI images that included the orbits and you all are able to see in the illustration that accompanies this comment, the lesion has significantly restricted diffusion as expected for such a highly cellular and malignant tumor.
Here is another nice example of the utility of DWI in head and neck lesions. In this case, the conventional post Gd T1 image shows no enhancement of the soft tissue mass in the mastoid bowl of a patient who had a canal up mastoidectomy previously for cholesteatoma. The fact that lesion does not enhance suggests recurrent/residual cholesteatoma which is confirmed by its very bright signal on DWI. For more discussion on this theme, look at the following article recently published in AJNR:
S. Thiriat, S. Riehm, S. Kremer, E. Martin, and F. Veillon. Apparent Diffusion Coefficient Values of Middle Ear Cholesteatoma Differ from Abscess and Cholesteatoma Admixed Infection. AJNR Am J Neuroradiol 2009 30: 1123-1126.
We have consistently noticed a similar degree of diffusion restriction with retinoblastomas (probably somewhere between 10-15 untreated cases over the last few years). This could theoretically help differentiate Rb from Coats’ disease, as we would expect much higher ADC with the hypocellular, exudative retinopathy seen in Coats’ (anecdotally, we have seen one case of Coats’ that fit this pattern, brought in from an outside hospital).
Another ocular problem-solving area where this may be helpful is in differentiating melanoma (particularly amelanotic) from metastases. The melanosomes and relative dense cellular packing of melanoma typically results in greater diffusion restriction than most other neoplasms. However, that can be a little trickier than the Rb/Coats differentiation or the pseudotumor/lymphoma differentiation. The gap in ADC between “diffusion restricted” neoplasms and other neoplasms is relatively small. Artifacts in the orbit make the ADC measurements a little less precise, and overlap between lesions becomes a bigger concern. Also, there is a significant effect of volume averaging with small lesions also (< 1 cm). This is an area that probably requires a technical advancement of some sort before it’s ready for everyday use.