Researchers from MIT and ETH Zurich developed an AI model to identify stages of ductal carcinoma in situ (DCIS) from breast tissue images, aiming to reduce overtreatment.
Connect with technology leaders today!
Ductal carcinoma in situ (DCIS) is a type of preinvasive tumor that sometimes progresses to a highly deadly form of breast cancer. It accounts for about 25 percent of all breast cancer diagnoses. Because it is difficult for clinicians to determine the type and stage of DCIS, patients with DCIS are often overtreated.
To address this, an interdisciplinary team of researchers from MIT and ETH Zurich developed an AI model that can identify the different stages of DCIS from a cheap and easy-to-obtain breast tissue image. Their model shows that both the state and arrangement of cells in a tissue sample are important for determining the stage of DCIS.
Because such tissue images are so easy to obtain, the researchers were able to build one of the largest datasets of its kind, which they used to train and test their model. When they compared its predictions to conclusions of a pathologist, they found clear agreement in many instances.
In the future, the model could be used as a tool to help clinicians streamline the diagnosis of simpler cases without the need for labor-intensive tests, giving them more time to evaluate cases where it is less clear if DCIS will become invasive.
“We took the first step in understanding that we should be looking at the spatial organization of cells when diagnosing DCIS, and now we have developed a technique that is scalable. From here, we really need a prospective study. Working with a hospital and getting this all the way to the clinic will be an important step forward,” says Caroline Uhler, a professor in the Department of Electrical Engineering and Computer Science (EECS) and the Institute for Data, Systems, and Society (IDSS), who is also director of the Eric and Wendy Schmidt Center at the Broad Institute of MIT and Harvard and a researcher at MIT’s Laboratory for Information and Decision Systems (LIDS).
Uhler, co-corresponding author of a paper on this research, is joined by lead author Xinyi Zhang, a graduate student in EECS and the Eric and Wendy Schmidt Center; co-corresponding author GV Shivashankar, professor of mechogenomics at ETH Zurich jointly with the Paul Scherrer Institute; and others at MIT, ETH Zurich, and the University of Palermo in Italy. The open-access research was published July 20 in Nature Communications.
Combining Imaging with AI
Between 30 and 50 percent of patients with DCIS develop a highly invasive stage of cancer, but researchers don’t know the biomarkers that could tell a clinician which tumors will progress.
Researchers have developed a novel artificial intelligence (AI) model that may accurately predict whether patients with cancer will respond to certain therapies, according to a recent study published by Sinha et al in Nature Cancer. The findings indicated that single-cell RNA sequencing data may be used to help physicians more precisely match patients with the most effective drugs for their cancer type.
Current strategies used to match patients with cancer drugs rely on bulk sequencing of tumor DNA and RNA, which takes an average of all the cells in a tumor sample. However, tumors contain many different types of subpopulations of cells and include clones that may respond differently to specific cancer drugs or develop resistance to them.
In contrast to bulk sequencing, single-cell RNA sequencing provides much higher resolution data down to the single-cell level. Previous research has suggested that using this approach to identify and target individual clones may lead to more lasting drug responses. Nonetheless, single-cell gene-expression data could be much more costly to generate than bulk gene-expression data and are not yet widely available in clinical settings.
In the recent proof-of-concept study, the researchers used widely available bulk RNA sequencing data and a machine learning technique known as transfer learning to train an AI model to predict responses to certain cancer drugs. They then fine-tuned the model with single-cell RNA sequencing data and used the novel approach on published cell-line data from large-scale drug screens to build AI models for 44 U.S. Food and Drug Administration–approved cancer drugs. The researchers noted that the AI models accurately predicted how individual cells would respond to both cancer monotherapies and combination therapies.
Further, the researchers tested the novel approach on published data involving 41 patients with multiple myeloma who received a combination of four drugs and 33 patients with breast cancer who received a combination of two drugs. They discovered that if just one of the clones was resistant to a particular cancer drug, the patient would not respond to therapy—even if all of the other clones responded. In addition, the novel AI model successfully predicted the development of treatment resistance in published data from 24 patients with non–small cell lung cancer who received targeted therapies.
The researchers emphasized that the accuracy of this novel approach could improve if single-cell RNA sequencing data become more widely available. They stated that as a result of their findings, they have developed a research website and a guide for how to use the novel AI model (called PERCEPTION) with new data sets.
For more insights on AI and healthcare, visit our Rapid Innovation Blogs.