Artificial intelligence project to identify skin cancer based on machine learning algorithm works as good as a doctor

Universal access to health care was on the minds of computer scientists at Stanford when they set out to create an artificially intelligent diagnosis algorithm for skin cancer. They made a database of nearly 130,000 skin disease images and trained their algorithm to visually diagnose potential cancer. From the very first test, it performed with inspiring accuracy.

This is a dermatologist using a dermatoscope. Credit: Matt Young

“We realized it was feasible, not just to do something well, but as well as a human dermatologist,” said Sebastian Thrun, an adjunct professor in the Stanford Artificial Intelligence Laboratory. “That’s when our thinking changed. That’s when we said, ‘Look, this is not just a class project for students, this is an opportunity to do something great for humanity.'”

The final product, the subject of a paper in Nature, was tested against 21 board-certified dermatologists. In its diagnoses of skin lesions, which represented the most common and deadliest skin cancers, the algorithm matched the performance of dermatologists.

Every year there are about 5.4 million new cases of skin cancer in the United States, and while the five-year survival rate for melanoma detected in its earliest states is around 97 percent, that drops to approximately 14 percent if it’s detected in its latest stages. Early detection could likely have an enormous impact on skin cancer outcomes.

Diagnosing skin cancer begins with a visual examination. A dermatologist usually looks at the suspicious lesion with the naked eye and with the aid of a dermatoscope, which is a handheld microscope that provides low-level magnification of the skin. If these methods are inconclusive or lead the dermatologist to believe the lesion is cancerous, a biopsy is the next step.

Bringing this algorithm into the examination process follows a trend in computing that combines visual processing with deep learning, a type of artificial intelligence modeled after neural networks in the brain. Deep learning has a decades-long history in computer science but it only recently has been applied to visual processing tasks, with great success. The essence of machine learning, including deep learning, is that a computer is trained to figure out a problem rather than having the answers programmed into it.

“We made a very powerful machine learning algorithm that learns from data,” said Andre Esteva, co-lead author of the paper and a graduate student in the Thrun lab. “Instead of writing into computer code exactly what to look for, you let the algorithm figure it out.”

The algorithm was fed each image as raw pixels with an associated disease label. Compared to other methods for training algorithms, this one requires very little processing or sorting of the images prior to classification, allowing the algorithm to work off a wider variety of data.

Rather than building an algorithm from scratch, the researchers began with an algorithm developed by Google that was already trained to identify 1.28 million images from 1,000 object categories. While it was primed to be able to differentiate cats from dogs, the researchers needed it to know a malignant carcinoma from a benign seborrheic keratosis.

“There’s no huge dataset of skin cancer that we can just train our algorithms on, so we had to make our own,” said Brett Kuprel, co-lead author of the paper and a graduate student in the Thrun lab. “We gathered images from the internet and worked with the medical school to create a nice taxonomy out of data that was very messy – the labels alone were in several languages, including German, Arabic and Latin.”

After going through the necessary translations, the researchers collaborated with dermatologists at Stanford Medicine, as well as Helen M. Blau, professor of microbiology and immunology at Stanford and co-author of the paper. Together, this interdisciplinary team worked to classify the hodgepodge of internet images. Many of these, unlike those taken by medical professionals, were varied in terms of angle, zoom and lighting. In the end, they amassed about 130,000 images of skin lesions representing over 2,000 different diseases.

During testing, the researchers used only high-quality, biopsy-confirmed images provided by the University of Edinburgh and the International Skin Imaging Collaboration Project that represented the most common and deadliest skin cancers malignant carcinomas and malignant melanomas. The 21 dermatologists were asked whether, based on each image, they would proceed with biopsy or treatment, or reassure the patient. The researchers evaluated success by how well the dermatologists were able to correctly diagnose both cancerous and non-cancerous lesions in over 370 images.

The algorithm’s performance was measured through the creation of a sensitivity-specificity curve, where sensitivity represented its ability to correctly identify malignant lesions and specificity represented its ability to correctly identify benign lesions. It was assessed through three key diagnostic tasks: keratinocyte carcinoma classification, melanoma classification, and melanoma classification when viewed using dermoscopy. In all three tasks, the algorithm matched the performance of the dermatologists with the area under the sensitivity-specificity curve amounting to at least 91 percent of the total area of the graph.

An added advantage of the algorithm is that, unlike a person, the algorithm can be made more or less sensitive, allowing the researchers to tune its response depending on what they want it to assess. This ability to alter the sensitivity hints at the depth and complexity of this algorithm. The underlying architecture of seemingly irrelevant photos — including cats and dogs — helps it better evaluate the skin lesion images.

Although this algorithm currently exists on a computer, the team would like to make it smartphone compatible in the near future, bringing reliable skin cancer diagnoses to our fingertips.

“My main eureka moment was when I realized just how ubiquitous smartphones will be,” said Esteva. “Everyone will have a supercomputer in their pockets with a number of sensors in it, including a camera. What if we could use it to visually screen for skin cancer? Or other ailments?”

The team believes it will be relatively easy to transition the algorithm to mobile devices but there still needs to be further testing in a real-world clinical setting.

“Advances in computer-aided classification of benign versus malignant skin lesions could greatly assist dermatologists in improved diagnosis for challenging lesions and provide better management options for patients,” said Susan Swetter, professor of dermatology and director of the Pigmented Lesion and Melanoma Program at the Stanford Cancer Institute, and co-author of the paper. “However, rigorous prospective validation of the algorithm is necessary before it can be implemented in clinical practice, by practitioners and patients alike.”

Even in light of the challenges ahead, the researchers are hopeful that deep learning could someday contribute to visual diagnosis in many medical fields.

Citation: Andre Esteva, Brett Kuprel, Roberto A. Novoa, Justin Ko, Susan M. Swetter, Helen M. Blau and Sebastian Thrun. “Dermatologist-level classification of skin cancer with deep neural networks.” Nature 2017.
DOI: 10.1038/nature21056
Adapted from press release by Stanford University.

Protective skin bacteria

There are more and more examples of the ways in which we can benefit from our bacteria. According to researcher Rolf Lood from Lund University in Sweden, this is true for the skin as well. He has shown that the most common bacteria on human skin secrete a protein which protects us from the reactive oxygen species thought to contribute to several skin diseases. The skin bacterium is called Propionibacterium acnes.

Propionibacterium acnes. Credit: Matthias Mörgelin, Lund University

He has discovered that the “acne bacterium” secretes a protein called RoxP. This protein protects against what is known as oxidative stress, a condition in which reactive oxygen species damage cells. A common cause of oxidative stress on the skin is UV radiation from the sun.

“This protein is important for the bacterium’s very survival on our skin. The bacterium improves its living environment by secreting RoxP, but in doing so it also benefits us”, explains Rolf Lood.
Oxidative stress is considered to be a contributing factor in several skin diseases, including atopic dermatitis, psoriasis and skin cancer.

Since Propionibacterium acnes is so common, it is present in both healthy individuals and people with skin diseases. According to Rolf Lood, however, people have different amounts of the bacterium on their skin, and it can also produce more or less of the protective protein RoxP.

This will now be further investigated in both patients and laboratory animals by Lood and his team. The human study will compare patients with basal cell carcinoma, a pre-cancerous condition called actinic keratosis and a healthy control group. The study will be able to show whether there is any connection between the degree of illness and the amount of RoxP on the patient’s skin.

The study on laboratory animals will also examine whether RoxP also f
unctions as protection. Here, mice who have been given RoxP and others who have not will be exposed to UV radiation. The researchers will then observe whether the RoxP mice have a better outcome than those who were not given the protective protein.

“If the study results are positive, they could lead to the inclusion of RoxP in sunscreens and its use in the treatment of psoriasis and atopic dermatitis”, hopes Rolf Lood. His research findings have recently been published in an article in the Nature journal Scientific Reports.

Citation: Allhorn, Maria, Sabine Arve, Holger Brüggemann, and Rolf Lood. “A novel enzyme with antioxidant capacity produced by the ubiquitous skin colonizer Propionibacterium acnes.” Scientific Reports 6 (2016): 36412.
DOI: http://dx.doi.org/10.1038/srep36412
Adapted from press release by Lund University.

Synthetic binding protein called "NS1 Monobody" found to inhibit common cancer causing (RAS) mutation

Monobody NS1 binds to H-RAS or K-RAS protein and blocks
RAS function by disrupting the protein’s ability to form active
 molecular pairs. Credit: John P. O’Bryan, et al. 

Researchers at the University of Illinois at Chicago have identified a new way to block the action of genetic mutations found in nearly 30 percent of all cancers. Mutations in genes for the RAS family of proteins are present in nearly 90 percent of pancreatic cancers and are also highly prevalent in colon cancer, lung cancer and melanoma, the most dangerous kind of skin cancer. The group of proteins include three members, K-RAS, H-RAS and N-RAS.

The prevalence of RAS mutations in human cancers and the dependence of tumors on RAS for survival has made a RAS a prime target for cancer research and drug discovery. Scientists and drug developers have long studied RAS oncogenes hoping to find a new treatment for cancer, but they have not yet been able to identify drugs that safely inhibit the oncogene’s activity.

John O’Bryan, associate professor of pharmacology in the UIC College of Medicine, led a team of researchers that took a different approach to studying RAS, and discovered that a synthetic binding protein they call “NS1 monobody,” which they created in the lab, can block the activity of the RAS proteins.

“We did not look for a drug or specifically for an inhibitor,” said O’Bryan, who is also a member of the University of Illinois Cancer Center and holds an appointment at the Jesse Brown VA Medical Center in Chicago. “We used monobody technology, a type of protein-engineering technology, to identify regions of RAS that are critical for its function.” Unlike conventional antibodies, monobodies are not dependent on their environment and can be readily used as genetically encoded inhibitors, O’Bryan said. “The beauty of the technology is that when a monobody binds a protein, it usually works as an inhibitor of that protein,” he said.

Monobodies were developed by Shohei Koide, a co-author on the study who is professor of biochemistry and molecular pharmacology at New York University. They have been used to target a diverse array of proteins that include enzymes and receptors.

The researchers found that the NS1 monobody binds to an area of the RAS protein molecule that was not previously known to be important for its oncogenic activity. NS1 strongly inhibits oncogenic K-RAS and H-RAS function by blocking the ability of the protein to interact with an identical one to form a molecular pair. NS1 does not affect N-RAS.

O’Bryan says the findings, published in the journal Nature Chemical Biology, provide important insight into long-standing questions about how RAS proteins function in cells. These insights may help guide the development of new therapeutic approaches to treating cancer by interfering with mutant RAS function in cancer cells.

“Development of effective RAS inhibitors represents a ‘holy grail’ in cancer biology,” O’Bryan said. “We now have a powerful tool we can use to further probe RAS function. While future studies and trials are needed before these findings can be leveraged outside the lab, this study provides new insight into how we can potentially inhibit RAS to slow tumor growth.”

Citation: Russell Spencer-Smith, Akiko Koide, Yong Zhou, Raphael R Eguchi, Fern Sha, Priyanka Gajwani, Dianicha Santana, Ankit Gupta, Miranda Jacobs, Erika Herrero-Garcia, Jacqueline Cobbert, Hugo Lavoie, Matthew Smith, Thanashan Rajakulendran, Evan Dowdell, Mustafa Nazir Okur, Irina Dementieva, Frank Sicheri, Marc Therrien, John F Hancock, Mitsuhiko Ikura, Shohei Koide & John P O’Bryan. “Inhibition of RAS function through targeting an allosteric regulatory site” Nature Chemical Biology 2016
DOI: http://dx.doi.org/10.1038/nchembio.2231
Research funding: Chicago Biomedical Consortium, Searle Funds at the Chicago Community Trust, Department of Veterans Affairs, National Institutes of Health
Adapted from press release by University of Illinois at Chicago

Synthetic binding protein called “NS1 Monobody” found to inhibit common cancer causing (RAS) mutation

Monobody NS1 binds to H-RAS or K-RAS protein and blocks
RAS function by disrupting the protein’s ability to form active
 molecular pairs. Credit: John P. O’Bryan, et al. 

Researchers at the University of Illinois at Chicago have identified a new way to block the action of genetic mutations found in nearly 30 percent of all cancers. Mutations in genes for the RAS family of proteins are present in nearly 90 percent of pancreatic cancers and are also highly prevalent in colon cancer, lung cancer and melanoma, the most dangerous kind of skin cancer. The group of proteins include three members, K-RAS, H-RAS and N-RAS.

The prevalence of RAS mutations in human cancers and the dependence of tumors on RAS for survival has made a RAS a prime target for cancer research and drug discovery. Scientists and drug developers have long studied RAS oncogenes hoping to find a new treatment for cancer, but they have not yet been able to identify drugs that safely inhibit the oncogene’s activity.

John O’Bryan, associate professor of pharmacology in the UIC College of Medicine, led a team of researchers that took a different approach to studying RAS, and discovered that a synthetic binding protein they call “NS1 monobody,” which they created in the lab, can block the activity of the RAS proteins.

“We did not look for a drug or specifically for an inhibitor,” said O’Bryan, who is also a member of the University of Illinois Cancer Center and holds an appointment at the Jesse Brown VA Medical Center in Chicago. “We used monobody technology, a type of protein-engineering technology, to identify regions of RAS that are critical for its function.” Unlike conventional antibodies, monobodies are not dependent on their environment and can be readily used as genetically encoded inhibitors, O’Bryan said. “The beauty of the technology is that when a monobody binds a protein, it usually works as an inhibitor of that protein,” he said.

Monobodies were developed by Shohei Koide, a co-author on the study who is professor of biochemistry and molecular pharmacology at New York University. They have been used to target a diverse array of proteins that include enzymes and receptors.

The researchers found that the NS1 monobody binds to an area of the RAS protein molecule that was not previously known to be important for its oncogenic activity. NS1 strongly inhibits oncogenic K-RAS and H-RAS function by blocking the ability of the protein to interact with an identical one to form a molecular pair. NS1 does not affect N-RAS.

O’Bryan says the findings, published in the journal Nature Chemical Biology, provide important insight into long-standing questions about how RAS proteins function in cells. These insights may help guide the development of new therapeutic approaches to treating cancer by interfering with mutant RAS function in cancer cells.

“Development of effective RAS inhibitors represents a ‘holy grail’ in cancer biology,” O’Bryan said. “We now have a powerful tool we can use to further probe RAS function. While future studies and trials are needed before these findings can be leveraged outside the lab, this study provides new insight into how we can potentially inhibit RAS to slow tumor growth.”

Citation: Russell Spencer-Smith, Akiko Koide, Yong Zhou, Raphael R Eguchi, Fern Sha, Priyanka Gajwani, Dianicha Santana, Ankit Gupta, Miranda Jacobs, Erika Herrero-Garcia, Jacqueline Cobbert, Hugo Lavoie, Matthew Smith, Thanashan Rajakulendran, Evan Dowdell, Mustafa Nazir Okur, Irina Dementieva, Frank Sicheri, Marc Therrien, John F Hancock, Mitsuhiko Ikura, Shohei Koide & John P O’Bryan. “Inhibition of RAS function through targeting an allosteric regulatory site” Nature Chemical Biology 2016
DOI: http://dx.doi.org/10.1038/nchembio.2231
Research funding: Chicago Biomedical Consortium, Searle Funds at the Chicago Community Trust, Department of Veterans Affairs, National Institutes of Health
Adapted from press release by University of Illinois at Chicago