US researchers have recreated the process by which ovarian cancer forms in the lab, providing solid evidence that the tumors start in the fallopian tubes, not the ovaries.

The finding could provide clues on how to attack ovarian cancer, which often causes no early symptoms and by the time it is found has spread so much that the tumors are impossible to stop.

Ovarian cancer is the fifth deadliest cancer among women, affecting 200,000 women worldwide annually and killing 115,000 women on average each year.

Several studies have theorized that the cancer may originate elsewhere, but the latest research by scientists at the Dana-Farber Cancer Institute in Boston shows how the cancer takes root first in fallopian tissue.

The fallopian tubes are the pathways by which a woman’s egg travels from the ovary to the uterus as part of her reproductive cycle.

Ronny Drapkin, senior author of the study in the Proceedings of the National Academy of Sciences, said previous examinations of fallopian tissue taken from women genetically predisposed to ovarian cancer have shown "patches of cells that were predecessors of serious cancers."

So they decided to try and replicate the process of cancer formation in the lab.

Researchers took fallopian cells and altered their genetic programming so they would divide much like cancer cells.

"Like true tumor cells, these ’artificial’ cancer cells proliferated rapidly and were able to leave their home tissue and grow elsewhere," said the study released on Monday.

"When implanted in laboratory animals, they also gave rise to tumors that were structurally, behaviorally, and genomically similar to human HGSOC (high-grade serious ovarian cancer)."

Drapkin said the findings demonstrate that fallopian cells are the source of ovarian cancer, and offer clues for future treatment.

"Such studies will help us identify different types of high-grade serous ovarian cancer, as well as possibly discover biomarkers - proteins in the blood - that signal the presence of the disease," said Drapkin, an assistant professor at Harvard Medical School.

"Ultimately, the model will enable us to test potential therapies to determine which work best in each type of the disease."