If cancer is one of the leading causes of death in the world (approximately 1 in 6 deaths), early diagnosis increases the chances of a cure. Researchers from the Koch Institute for Integrative Cancer Research at MIT and Massachusetts General Hospital (MGH) used the deep learning and built a developmental multi-layer perceptron classifier (D-MLP) to identify the origin of cancer. Their study entitled “Developmental Deconvolution for Classification of Cancer Origin” was published at the end of August by Cancer Discovery.
Cancer of Unknown Primary (CPI or CPU in English) is cancer that has already spread to other organs in the body (metastasized), but doctors have not found the original tumour. Generally small in size, it is however very aggressive, so oncologists must quickly implement non-targeted treatments, which are often toxic for the patient.
This new approach based on deep learning could help classify unknown primary cancers by taking a closer look at gene expression programs related to early cell development and differentiation.
Salil Garg, Charles W. (1955) and Jennifer C. Johnson Clinical Investigator at the Koch Institute and Pathologist at the MGH, lead author of the study, explains:
“Sometimes you can apply all the tools pathologists have to offer, and you’re still left with no answers. Machine learning tools like this could allow oncologists to choose more effective treatments and give more advice to their patients. »
A study based on gene expression and deep learning
Cancer cells look and behave very differently from normal cells, in part because of significant alterations in the way their genes are expressed. Advances in single-cell profiling and efforts to catalog different cell expression patterns in cell atlases have provided many data with clues to the origin of different cancers. the deep learning is an ideal technology to exploit this data.
To make their model more efficient, the researchers had to reduce the number of features while extracting the most relevant information, and focused the model on signs of altered developmental pathways in cancer cells.
As an embryo develops, undifferentiated cells specialize in various organs, many pathways direct how cells divide, grow, change shape and migrate. As the tumor grows, cancer cells lose many of the specialized traits of a mature cell. They can be compared to embryonic cells in some aspects, as they have the ability to proliferate, transform and metastasize.
The researchers compared two atlases of large cells, identifying correlations between tumor and embryonic cells:
- the Cancer Genome Atlas (TCGA), which contains gene expression data for 33 tumor types;
- the Mouse Organogenesis Cell Atlas (MOCA), which describes 56 distinct trajectories of embryonic cells as they develop and differentiate.
Enrico Moiso, postdoctoral fellow at MIT, also lead author of the study, explains:
“Single-cell resolution tools have radically changed the way we study cancer biology, but how we make this revolution impactful for patients is another question. With the emergence of developmental cell atlases, especially those that focus on early phases of organogenesis such as MOCA, we can extend our tools beyond histological and genomic information and open the door to new ways profile and identify tumors and develop new treatments. »
The researchers broke down the gene expression of TCGA tumor samples into individual components corresponding to a specific point in a developmental trajectory and assigned each a mathematical value.
They then constructed a model of deep learninga developmental multilayer perceptron (D-MLP), which scores a tumor for its developmental components and then predicts its origin.
Classification of tumors
After the training, D-MLP was applied to 52 new samples of cancers among the most difficult cases encountered at the MGH from 2017 to 2020, which had not been able to be diagnosed. The model classified tumors into four categories and provided predictions and other information that could guide the diagnosis and treatment of these patients.
One of these 52 samples came from a patient with a history of breast cancer who showed signs of aggressive cancer in the fluid spaces around the abdomen. D-MLP strongly predicted ovarian cancer, and indeed a mass was found in the ovary six months later that caused this cancer.
The results of this study have provided an atlas of the origins of tumor development, a tool for diagnostic pathology, and suggest that developmental classification may be a useful approach for patient tumors.
For their next work, the researchers plan to increase the predictive power of their model by adding other types of data, including information collected in radiology, microscopy and other types of tumor imaging.
Salil Garg concludes:
“Developmental gene expression represents only a small part of all the factors that could be used to diagnose and treat cancers The integration of radiology, pathology and gene expression information is the real next step in personalized medicine for cancer patients. »
Sources of the article:
- Enrico Moiso, Koch Institute for Integrative Cancer Research, Massachusetts Institute of TechnologyCambridge MA and Harvard-MIT Broad InstituteCambridge MA;
- Alexander Farahani, Department of Pathology, Massachusetts General Hospital, Harvard Medical SchoolBoston MA;
- Hetal D. Marble, Department of Pathology, Massachusetts General Hospital, Harvard Medical SchoolBoston MA;
- Austin Hendricks,Koch Institute for Integrative Cancer Research, Massachusetts Institute of TechnologyCambridge MA;
- Samuel Mildrum, Koch Institute for Integrative Cancer Research, Massachusetts Institute of TechnologyCambridge MA;
- Stuart Levin, Koch Institute for Integrative Cancer Research, Massachusetts Institute of TechnologyCambridge MA;
- Jochen K. Lennerz, Department of Pathology, Massachusetts General Hospital, Harvard Medical SchoolBoston MA;
- Salil Garg, Koch Institute for Integrative Cancer Research, Massachusetts Institute of TechnologyCambridge MA.