Cancer researchers often find themselves flooded by the amount of data they have to sift through in their projects. Often, this data is in a form that computers find hard to process and analyze, such as pathology slides from tumor biopsies.
Cancer Research UK, finding itself loaded with a couple million slide images, decided to outsource the analysis to you, the anonymous online volunteer who always wanted to help with cancer research, but is instead an ophthalmologist. Partnering with Zooniverse, a project that allows anyone to help with looking at NASA’s data, the Cell Slider site, after a short prep, asks volunteers to count irregular cells and quantify how stained they are for every image presented. Each image is run through five unique volunteers to improve data confidence and each volunteer is rated in the background by the site based on the accuracy of previous image runs.
Each sample in the database has been stained to highlight the differences between ordinary cells, such as white blood cells, and irregular, cancerous cells. One slide has around 120 tumour samples, but these are broken down into 16 sections so the public is looking at a tiny subsection of the full dataset when they view each image on Cell Slider.
In the case of the breast cancer samples, the yellow colour represents the level of the oestrogen receptor (ER) protein present in the cell, a normal protein that in excess attracts too much oestrogen, which stimulates cancer cell division and tumour growth. Around 70 percent of breast cancers are classed as ER positive. Hormone therapies are helpful in treating this type of cancer — drugs such as tamoxifen effectively slot into the oestrogen receptors to block oestrogen, preventing it reaching and feeding the cancer. This is usually only effective if there are high ER levels, and that’s where Click to Cure comes in.
More from Wired UK: Citizen science project crowdsources identification of cancer cells
Link: Cell Slider…
