Skip to main content
Premium Trial:

Request an Annual Quote

Cancer Researchers Detail Efforts to 'Discover, Predict, Prevent, Treat' at Opening of AACR Meeting

WASHINGTON (GenomeWeb) – The American Association for Cancer Research's annual meeting kicked off in Washington, DC, on Sunday with a tribute to past progress in research and treatment of cancer, and with some harsh words for President Donald Trump's recent proposals for deep cuts to the NIH's budget in 2017 and 2018.

AACR President Nancy Davidson called the cuts "appalling" and "short-sighted," while emphasizing the importance of the work such funding supports. She asked the assembled researchers to stand and hold up preprinted signs with slogans such as "Make NIH Funding a National Priority" and "Cancer Research Makes Economic Sense," as cameras panned the room. Davidson also urged the audience to do what they could to fight the proposed cuts, and pledged to send attendees "action alerts" throughout the conference with ideas on what to do.

The theme of the conference this year is "Discover, Predict, Prevent, Treat" — four arms of a single mission to reduce deaths and learn more about the diseases that comprise cancer.

MIT researcher Angela Belcher spoke about her efforts to discover ovarian cancer tumors half a centimeter in size — an advancement in cancer discovery that she said could cut mortality in half.

Belcher's expertise is in engineering the M13 bacteriophage to carry carbon nanotubes made of graphene in order to imprint genetic traits onto solar cells, batteries, and other materials. She was asked to work with researchers at Massachusetts General Hospital and MIT's Lincoln Laboratory to apply that expertise to detecting cancer. Because carbon nanotubes have fluorescent properties, Belcher thought of attaching them to bacteriophage that were specifically engineered to search for cancer cells and attach to them. The fluorescence of the nanotubes could potentially help a clinician detect very small or hidden tumors for surgical removal.

Belcher and her students even created a real-time imaging system, which would allow a surgeon to go from seeing with his own eyes to looking at a monitor that would show fluorescence during surgery. The true positive rate for finding tumors this way was 98 percent, Belcher said — an order of magnitude better than contrast CT.

The technology also has applications for tracking immune cells in real time in response to cancer, which could improve immunotherapy, and could help a clinician change therapies based on patient response.

Cancer Research UK researcher Carlos Caldas then spoke about his experience looking at the tumor heterogeneity landscape of breast cancer in order to help predict disease progression.

Breast cancer consists of 11 distinct genomic subtypes, which are defined by copy number aberrations layered with different mutations, he said. But each subtype in the early stages of disease already exhibits significantly varied intratumor heterogeneity (ITH). Further, he noted, tumor metastases on average don't accumulate many more mutations than what's present in early breast cancer, when the disease is still curable.

Using genomic sequencing and analysis to define the tumoral landscape of patients with lethal disease, Caldas found that most mutations and drivers are occurring either as stem or clade mutations. Further, he noted that certain new mutations — dubbed "genomic scars" — can be caused by response to treatment or as a consequence of metastasis, but that the metastases themselves seemed to have much the same mutational landscape as early tumors.

The RNA sequencing data, however, showed something a little different. That data, from the same patients with lethal disease, was used to characterize the immune tumor microenvironment (TME), and showed that the immune TME, unlike the mutational landscape, varied across the metastases.

And when looking at T-cell receptors, Caldas noticed that they appear to be organized according to which tissue the metastasis is in.

Knowing that the phylogenies of T-cell receptors and cancer genomes are correlated, and that T-cell receptors and neoantigens are significantly co-shared across metastases, Caldas speculated that this may suggest that disease and T-cell receptors co-evolve. Though the data is still preliminary, he added, this work suggests that a combined analysis of cancer genomics and T-cell receptors may advance our understanding of how to predict disease progression.

In a separate symposium on computational biology on Sunday, Hospital for Sick Kids bioinformatician Benjamin Brew detailed his group's efforts to use methylation data to accurately predict the age of cancer onset in patients with Li Fraumeni Syndrome (LFS).

LFS is a rare hereditary genetic cancer predisposition syndrome. Children who have it have an almost 100 percent chance of developing cancer over their lifetimes, and a 40 percent chance of developing cancer by the age of 25. Their survival often depends on their clinicians being able to observe them closely to catch the disease in its earliest stages.

The researchers devised an algorithm that can use methylation data gained from blood samples from each patient to determine the age at which a given child will develop a cancer. The predictive computer model was trained and tested using existing data from an LFS cohort. The first step was to use an existing algorithm called Bumphunter to analyze methylated regions in LFS patients with and without cancer, and to narrow that list down to those that could predict cancer onset. The researchers then used 70 percent of the cohort data as a training set for the computer model and saved the remaining 30 percent as test data to see if the model had learned its task.

The team found that the model was able to use methylation data values and clinical data for each patient to predict age of cancer onset in months and also to determine the probability of a diagnosis before and after that age, Brew said. The model had an 88 percent true positive rate before the age of 4, and a true positive rate of 89 percent before the age of 6.

Though the team is working to improve the rate of false positives, Brew added, the consequence of such a result is that a patient would be subjected to greater observation — not ideal, but better than the consequence for a false negative. Also, he said, the methylation profiles generated by the model can help researchers pinpoint more specifically regions of the genome that could be responsible for the development of cancer in children with LFS, which could lead to new treatments.

Illustrating the prevention theme at the opening plenary was Johns Hopkins University's Bert Vogelstein, the development of new therapies goes hand in hand with the development of new prevention techniques.

He used mice as an example of the benefits of finding and excising tumors when they are at their smallest. Though he joked about turning people into mice in order to cure cancer — and then immediately asked that no one tweet his comment lest President Trump get the wrong idea and call for even deeper cuts to the NIH's budget — Vogelstein said in reality, very small amounts of disease can often be cured by chemotherapeutic agents, and that clinicians have a big window of opportunity from the time cancer starts to when it becomes malignant to treat it.

Along with this, he added, we need new modes of prevention, and one key step in that is identifying the source of mutations for each type of cancer. One way to do that is to find improved molecular markers of disease with diagnostics such as liquid biopsies. Vogelstein called liquid biopsies the "best new class of molecular markers" in use today.

He further noted that prevention of cancer should take a page from the book of cancer treatment — that is to say, using more than one diagnostic at a time, much like therapies are often combined. Liquid biopsies, for example, look for somatic mutations, but other diagnostics can be used to find DNA methylation, cancer-specific metabolites, and many other markers.

A second combination type involves using liquid biopsies for finding ctDNA from tumors and using other diagnostics to look for localized tumor DNA. For example, pap smears can be used not only to look for HPV, but to look for ovarian or uterine tumor DNA in vaginal cells. Likewise, the oral cavity can be swabbed to look for head and neck cancer DNA. These types of localized tests in combination with liquid biopsy can amplify their diagnostic power and help clinicians prevent cancer from developing into lethal disease.

Finally, Memorial Sloan Kettering Cancer Center's Charles Sawyers took on the idea of developing better treatments for cancer through a better understanding of drug resistance.

Through his work on prostate cancer, Sawyers has found that next-generation therapies allow patients to live longer and cause fewer patients to develop resistance. Certain inhibitors, for example, can decrease the relative frequency of common mutations. Because of that, he noted, tumors develop non-genomic resistance mechanisms. But these, too, are druggable through the selective use of chromatin modifiers and other methods.

His research has shown that the early and rational application of combination therapies is the only logical path forward. And as patients that could possibly respond to such therapies can be identified through the many profiling technologies that already exist, these therapy combinations must be deployed as first, rather than last, resorts.