The University Network

Method Revolutionizes Tracking The Spread Of Cancer

A team of researchers has developed a new method to track the spread of cancer cells, yielding a clearer understanding of cancer migration than ever before.

The spread of tumor cells to different locations in the body, known as metastasis, is the most dangerous element of cancer. Metastatic disease causes close to 90 percent of cancer deaths from solid tumors.

Understanding what causes metastasis is crucial to developing treatments capable of halting the spread of cancer.

In May, the researchers presented an algorithm called MACHINA that can track the spread of cancer cells by combining DNA sequence data with information on the location of the cells in the body.

MACHINA stands for “metastatic and clonal history integrative analysis.”

Recently, they have made improvements to MACHINA and are using the algorithm to develop a clearer understanding of metastasis.

“The main contribution of our paper is that we introduced an algorithm that infers patterns of metastasis from sequencing data,” said Mohammed El-Kebir, an assistant professor of computer science at the University of Illinois at Urbana-Champaign who co-authored the study while he was a post-doctoral researcher at Princeton University.

“In contrast to previous methods that have been developed for use in species evolution, our algorithm MACHINA incorporates an evolutionary model that is tailored to cancer.”


Some of the information gathered through the use of MACHINA suggests cancer-cell migration patterns that don’t match the current understanding of cancer biology.

In the study, MACHINA simultaneously traced the mutations and movements of cells to prove that metastatic disease can result from fewer cell migrations than previously assumed.

In a study of a breast cancer patient, MACHINA suggested that a secondary tumor in the lung caused metastatic disease through five cell migrations. A previously conducted analysis suggested it was caused by 14 migrations.  

“We decided to develop MACHINA after realizing that current analyses of metastasis might be incorrect due to inappropriate assumptions in the used algorithms, which were initially developed for use in species evolution,” said El-Kebir.

The researchers also used their algorithm to analyze metastasis in patients with melanoma, prostate and ovarian cancers.  

They added multiple features to improve MACHINA’s accuracy.

Because it has been proven that tumor cells can travel in clusters, the algorithm includes a model for the co-migration of cells.

The algorithm also is set to recognize the uncertainty in DNA data resulting from the mixtures of healthy cells and tumor cells.

A paper describing the team’s full efforts is published in the journal Nature Genetics.

Gryte Satas, a doctoral student at Princeton, was also a co-author of the study.

Implications of MACHINA

Through the use of MACHINA, researchers could uncover key patterns and mutations that cause the spread of cancer.

“A better algorithm is like a better microscope,” Ben Raphael, a professor of computer science at Princeton and senior author of the research, said in a statement.

“When you look at nature with a magnifying glass, you may miss important details. If you look with a microscope you can see much more.”

What’s next?

The next step in this development is to apply MACHINA to a large amount of matched primary and metastasis samples, explained El-Kebir.

“This will enable one to identify common patterns of metastatic progression, including the mutations that drive metastasis,” he said.

Before the technology can be actively used by doctors to track the spread of cancer, the researchers wish to improve the accuracy of MACHINA by advancing the sequencing technology with longer reads and less errors, explained El-Kebir.

Additionally, the researchers plan to improve their method by incorporating data from tumor cells and DNA in the bloodstream, and they want to recognize data regarding reversible chemical modifications of DNA.   

Altogether, the development is promising and imperative in the fight against cancer.

“I predict this new method will be of widespread use to the genomic community and will shed new light on the most deadly phase of cancer evolution,” Andrea Sottoriva, the Chris Rokos Fellow in Evolution and Cancer at The Institute of Cancer Research in London, said in a statement.