Google-like algorithm shows that not all metastases are alike
April 5, 2013 | by Hiu Chung So
Although cancer metastases are generally associated with worse outcomes than primary tumors, some of these secondary tumors are more dangerous than others because they further spread cancerous cells, which then invade other organs. Thus, knowing how a cancer metastasizes could give oncologists better insight about where to provide targeted treatments that could slow or halt the disease’s progression.
In a Cancer Research study published online Feb. 27 ahead of print, researchers found that an algorithm similar to Google PageRank could provide just that kind of help to clinicians treating lung cancers.
In the paper, the authors applied an advanced math model known as a Markov chain to evaluate 163 autopsy reports from patients who died of lung cancer between 1913 and 1943. This period was chosen because it preceded the use of radiation or chemotherapy, allowing the researchers to observe the cancer’s natural, untreated progression.
Similar to the way in which Internet search engines rank websites by assessing how many pages link to it, the scientists used the Markov chain to assess whether certain secondary tumor sites were associated with further metastases, thereby making them “spreaders.” This is in contrast to “sponges,” metastasis sites that are unlikely to further distribute the cancer.
After analyzing the metastasis patterns, the researchers found that kidneys and adrenal gland metastases are more likely to be “spreaders,” being respectively 2.86 and 1.91 times more likely to disseminate cancer cells than they are to receive them. Meanwhile, distant lymph nodes, liver and bone are considered “sponges,” meaning they're more likely to absorb cancer cells than spread them.
Additionally, the researchers found that metastasis is not a unidirectional process from the primary tumor outward, but that there multiple interconnected pathways among metastatic sites, allowing them to seed one another and even back to the primary tumor site as well.
Assessing the findings, Karen Reckamp, M.D., M.S., co-chair of City of Hope's Lung Cancer and Thoracic Oncology Program, told San Diego Union-Tribune that “this [study] means there might be reasons to do localized therapy at a metastatic site.” Reckamp noted that in the past, localized therapy is only done for palliative purposes when a lung cancer has already metastasized.
“We now understand there is significant heterogeneity among the primary and secondary tumors ... increased knowledge on the source and spread of tumor cells will improve our use of targeted therapies for patients,” Reckamp said.
And with a better understanding of spreaders and sponges, clinicians will have the chance to be better prepared to nip the metastasis problem in the bud, before it wreaks further havoc upon the body.