Almost all stem cells transfers has import risks such as dislocation of stem cells, differentiation to different kind of cells. If your study experimental you can infuse via eye' blood stream via venous. But you don't forget that yours cells may localize another part of eye and make important side effects such as tissue degeneration or transformation even cancer. In this point you must select most suitable cell line and if possible you label with GFP etc to chase where is your cells.
The short answer seems to be that the potential benefits are probably much greater than the potential risks. The current strategy involves differentiating cultured stem cells into mature retinal pigment epithelial (RPE) cells and then implanting a sheet of these cells under the retina to replace damaged RPE. The source of the stem cells can either be embryonic (Dr. Schwartz et al. in the USA) which require immunosuppression, or IPS stem cells (Dr. Takahashi at RIKEN in Japan) which are derived from the patients own fibroblasts and do not require immunosuppression. Remember that these are differentiated cells, not pluripotent cells, that are being implanted in the eye. The risk of malignancy is probably (but not certainly) low, and if anything it would be a teratoma that developed, not a metastatic cancer.
For reading I would start with these two short Nature News articles:
I have attached several papers - one by Schwartz about the US trial, and two from Japan including one specifically addressing the tumorgenicity of IPS derived cells. They have good references that you can use to read further.
You have picked a good subject to study - this is one of the most interesting and exciting areas in all of medicine.
There were some researches in using of stem cells to treat eye diseases in our Institute as well (Kazakh Eye Research Institute). However, they all had side effects and transformed into cancer. Then, I suggested the theory of using specific growth factors (for that we need deeper study the embryology of eye)
If they are embryonic stem cells then injection may well lead to tumour growth since the formation of teratomas is indeed the gold standard formal proof of pluripotency. While it is more likely that differentiated cells from embryonic stem cells (or adult stem cells) would be injected into the retina, there is still a small risk because there may be residual undifferentiated cells among those being injected.
Like any medicinal product, any stem cell derived cell product, e.g. retinal pigemental epithelial differentiated from stem cells, would need to undergo quality control procedures, which in this case would include checking for the presence of residual pluripotent stem cells.
While the application of such quality control procedures would minimise risk, still a risk might remain due to the possibility of a de-differentiation event occurring in a precursor cell type, that together with any acquired genetic anomaly, could then in theory lead to tumour progression.
However, one reason for the eye being chosen as one of the first organs to be targeted by such therapy is that in the extremely unlikely event of tumour formation (due to the quality control procedure that would be applied), the eye could simply be removed and thus removing the risk of any metastases in the individual.