 Can computers learn to diagnose cancer? An international team of scientists has created a method that teaches computers to spot patterns and images of ovarian tumors through a process called Texture Analysis. With this computer vision, doctors can tell a lot about how fatal a tumor is and how the cancer will progress. This is particularly important because ovarian cancer is highly lethal and is rarely detected before it has spread throughout the body. The group previously used Texture Analysis to look at two-dimensional images of ovarian tumors to learn how advanced the cancer had become. 2D images can be stitched together to create 3D images, which provide scientists with a more detailed look at tissues. However, even when combined, these images still leave out a lot of potentially valuable information. To capture these crucial additional details, the team gathered 3D images from the start using a technique called 3D second harmonic generation microscopy. They then wrote an algorithm that could perform Texture Analysis of that data. Because these images created a more complete picture to begin with, this approach improved the accuracy of cancer staging by up to 8%, allowing about 90% of tumors to be accurately diagnosed. Using the algorithm, the scientists trained computers to separate normal tissue from benign tumors, low-grade tumors, and high-grade tumors. They even predicted when growths were at high risk of becoming cancer. They identified features corresponding to these different tumor types and found that they were consistent among different patients. This means that 3D Texture Analysis could provide a universal way of telling healthy tissue from potentially life-threatening tissue. As a result, Computer Vision offers a powerful method for examining cancer that can complement other techniques such as genetic screening. When combined with microscopic imaging, this new way of analyzing growths in the ovaries offers a minimally invasive approach for screening women at risk of developing ovarian cancer. By offering a consistent way to measure not only cancer stage but also cancer risk, ovarian cancer can be caught earlier, allowing doctors to treat tumors before they become deadly.