NEW YORK (GenomeWeb) – Some of the same precision medicine capabilities coming into focus for full-blown cancers are starting to glint on the horizon for individuals with pre-cancerous conditions and their clinicians.
"You hear about precision medicine in the context of [cancer] treatment. But you don't hear about precision medicine in the context of early-stage disease," explained Sudhir Srivastava, chief of the National Cancer Institute (NCI) cancer prevention division's cancer biomarkers research group.
The PreCancer Atlas (PCA) — a scheme Srivastava and others first proposed several years ago — aims to change that.
The effort currently encompasses five pilot projects that were funded in September 2018 and are scheduled to run for five years apiece. Their goals include characterizing molecular, cellular, clinical, and other features over time in the lung lesions, colon polyps, altered milk duct cells, and other tissue changes at risk of sliding into cancer.
"Not all preneoplastic or precancerous lesions progress," Srivastava explained. "Most of them won't. But some will."
He and his colleagues outlined the rationale for PCA in an opinion piece published in Trends in Cancer last August.
By taking a deep, longitudinal look at still-cancer-free individuals, PCA investigators hope to get a clearer understanding of the changes that mark unfortunate cases that do progress. These insights are expected to inform future efforts to develop preventive and patient-tailored treatments and come up with earlier disease biomarkers, while making it possible to compare existing pre-cancer clinical management methods, and eventually, risk stratify individuals with pre-cancers based on a more accurate estimate of the likelihood they'll progress to cancer.
In the longer term, the data compendium coming out of PCA might move clinicians closer to another highly sought-after cancer prevention goal: vaccines for thwarting pre-cancers before they have a chance to gain traction. Srivastava noted that insights gleaned from longitudinal pre-cancer samples might help identify new antigens, or "neoepitopes" — molecular changes that alert the immune system to trouble and may rouse a response.
Understanding the events that lead to progression in different pre-cancer types might also make it possible to know when to use prophylactic vaccines, he further proposed.Â
Beyond that, data from PCA might even dial back excess screening in those at greater risk of over-treatment than of cancer itself. In a perspective paper published in Nature Reviews Cancer in late April, Srivastava and co-authors from NCI, the University of Texas MD Anderson Cancer Center, and the University of California, Davis laid out an argument suggesting that a better understanding of tumor evolution and microenvironment interactions, together with better biomarkers and improved early- or pre-cancer imaging methods, may rein in the unnecessary treatment or torpid, non-threatening tumors.Â
The PCA effort grew from the kernel of an idea from Srivastava and other project leaders who reasoned that progress being made in the wake of the Cancer Genome Atlas (TCGA) could be replicated at an earlier stage of disease. What if you could peer into tumor drivers, potential treatment targets, and tumor evolution before cancer develops?
"When you think about it, what you really want to figure out is — you want to get at the earliest stages of diseases. You want to understand best how to detect it, [and] treat it," said Michael Snyder, another PGA project leader, who chairs Stanford's genetics department and directs its Center for Genomics and Personalized Medicine. "It's very fundamental."
In 2016, Srivastava convened a think tank in Bethesda, Maryland to start germinating that idea.
Over two bright days in June, government and academic investigators from centers across the country and as far away as the UK and France descended on NCI's Shady Grove Campus to carefully outline the knowledge gaps that have delayed significant progress in cancer prevention and early disease detection before digging into the types of projects that might fill that void.
"There was a strong sense that we should do something in this field [of cancer prevention]," Srivastava said, noting that, traditionally, "prevention hasn't succeeded very much."
When Cancer Moonshot funding arrived in 2017, the PCA leaders saw an opportunity to move the concepts discussed in Bethesda beyond the hypothetical and joined forces with other Moonshot-funded tumor-profiling teams under the NCI's Human Tumor Atlas Network (HTAN).
Together, the HTAN project will follow cancers from their very earliest stages through to metastasis. On the precancer side of that continuum, the PCA projects, funded following a National Institutes of Health request for application and external peer review process, are focused on sporadic or familial conditions that can lead to colon, lung, breast, skin, or blood system cancers.
At Stanford University, for example, Snyder, James Ford, and colleagues plan to profile polyps from individuals with a hereditary colon cancer syndrome called familial adenomatous polyposis (FAP), characterized by germline mutations to the tumor suppressor gene APC.
That hereditary condition is especially amenable to the longitudinal sampling approaches being pursued by PCA, in part because patients inevitably advance to colon cancer. As a consequence of this inescapable trajectory, FAP patients get frequent colonoscopy screening and undergo surgery to remove their colon when polyps become particularly large or suspicious-looking.
"Most patients are very eager to help, especially when it comes to cancer, because they hope anything we learn might ultimately be useful to either them or the field," Snyder said. "The enrollment rate is quite high."
Snyder and his associates have already collected hundreds of colon polyps from individuals with FAP. With more than $4.2 million in Cancer Moonshot funding, they plan to subject these gut growths to a battery of tests — from single-cell sequencing and bulk genomic, epigenetic, metabolomic, methylation, transcriptome, and proteomic profiling to polyp imaging.
The team expects that its upcoming single-cell RNA sequence atlas will be particularly important, since it will provide a refined look at the gene expression shifts in progressing tumors in a range of cell types over time.
However, that group will also use other new and established technologies to assess polyps of different sizes, spanning the benign to the dysplastic, as well as matched blood samples, unaffected colon tissue, corresponding cytokine immune cells, and gut microbes.
"Technology has progressed a lot since the TCGA days," Snyder said, noting that "we hope to take this to a whole new level" and to develop new technologies along the way.
At Vanderbilt University, meanwhile, Robert Coffey and colleagues are in the process of establishing their own single-cell atlas on colorectal neoplasms and the microenvironment surrounding them for a project that has secured almost $11 million in Cancer Moonshot funding.
That work is centered on cells collected across space and time in and around the pre-cancerous adenomas that might progress to colorectal cancer, which will be tested with approaches such as single-cell RNA-seq, exome sequencing, multiplex immunofluorescence, and fluorescence in situ hybridization.
In the study's abstract, Coffey and his co-authors noted that colonoscopy may miss nearly one-quarter of the full-fledged adenomas in the colon, which has a sizeable economic and psychological impact on patients, and puts patients at risk for serious complications, such as bleeding.
"There is an unmet need to develop new preventative strategies and risk stratification models to address these and other issues," they wrote, adding that "we believe that the ability to provide the most effective precision diagnostics and preventive strategies can only be achieved with single-cell analysis."
Still other teams are studying early lesions in the hope of ultimately staving off other common cancers such as lung adenocarcinoma or lung squamous cell carcinomas (SCC), which often arise in current or former smokers.
"To us, that's a fundamental gap: we understand a lot about late-stage disease. We know very little about early, pre-malignant disease, in terms of biology," noted Avrum Spira, chief of Boston University School of Medicine's computational biomedicine division, a leader on the PCA's lung-focused project. "From there, of course, we'll hopefully get to new treatments to prevent cancer, as well as biomarkers for early detection."
With their five-year budget of nearly $6.7 million, Boston University's Spira and other members of the Lung PCA project are working on retrospective and prospective analyses on pre-cancerous lesions that lead to adenocarcinoma or SCC, along with samples from the surrounding microenvironment, when possible.
It should be possible to look at the suspicious lesions in the lung's central airways — those that can progress to SCC — both retrospectively and prospectively over time, given the accessibility of those samples with semi-invasive methods such as bronchoscopy, Spira explained.
Because adenocarcinomas typically crop up deeper in the lung, on the other hand, the precancerous lesions associated with them will not be accessible by bronchoscopy. Instead, those lesions will be profiled retrospectively from early-stage lung cancer surgical specimens that have precancerous lesions adjacent to tumors.
"The hard part is finding people that have these precancers. [The lesions are] not easy to find. That's probably the biggest limiting factor in enrollment" Spira said.
In a paper published in Nature Communications in late April, Spira and co-authors presented findings from a study of bronchial premalignant lesions in 29 high-risk individuals, uncovering four molecular subtypes and the transcriptome features associated with premalignant progression to lung SCC.
"Now with the money from NCI through the Moonshot, we're going to expand the work," Spira said. "Our hope would be to get the work done over the next three to four years."
Researchers at the University of Colorado and elsewhere are undertaking a related clinical trial that will include prospective, serial bronchoscopy on participants with pre-SCC lesions. Though the trial is not funded through the PCA, that group is collaborating with Lung PCA investigators.
Over at Duke University, meanwhile, Shelley Hwang and colleagues have nailed down nearly $10.8 million for their PCA study on sporadic breast ductal carcinoma in situ (DCIS). That project will include retrospective and prospective cohorts profiled at the DNA, RNA, methylation, imaging, single-cell, and/or immune cell levels, with accompanying clinical data.
That group intends to do extensive, multi-region analyses on samples collected over time from perhaps 10 to 30 individuals with DCIS, along with sparser profiling on a larger group of DCIS patients, noted team member Carlo Maley, a cancer, evolution, and computational biology researcher at Arizona State University.Â
Along with her work on PCA, Hwang is co-leading a clinical trial comparing standard interventions with watchful waiting for low-risk DCIS, Maley said, noting that the Breast PCA investigators will have access to some of the samples collected for that trial.
"No one knows which [DCIS precancers] will progress to cancer," Maley said. In general, roughly 20 to 30 percent of DCIS cases progress to cancer, he explained, though precancers in other tissues often have even lower progression rates.
Ideally, the PCA may help in identifying some surefire flags for progression. Conventional screening methods such as mammography have been criticized for picking up slow-growing, low-risk tumors that are at relatively low risk of progression to cancer, Maley explained, meaning increased DCIS detection may not translate into a corresponding reduction in breast cancer cases.
The US Preventive Services Task Force currently recommends mammography every two years for women between the ages of 50 and 74, but is less definitive in its recommendations for women under 50 years old. It notes that the number of DCIS in American women has jumped from six cases per 100,000 to 37 cases per 100,000 since the widespread adoption of mammography in the US, though the "natural history of DCIS — particularly screen-detected DCIS — is poorly understood."
Finally, a team led by investigators at Harvard Medical School is using more than $6.9 million in Cancer Moonshot money to put together pre-cancer atlases for both pre-melanoma and clonal hematopoeiesis events that can progress to myeloid neoplasm.
"Both [pre-malignant diseases] involve expansion of clones in normal and diseased niches as shaped by complex interactions among immune and pre-cancer cells," Harvard's Peter Sorger and other investigators involved wrote in an online abstract outlining the project. Together, they said, the two atlases they are pursing "present complementary technical challenges, avenues to scientific discovery, and opportunities for the development of precision prevention strategies and therapies."
Results from all five PCA pilot projects will be made available through a central repository, with an emphasis on data sharing and release, though the precise format for that site will become clearer in the next few months.
Although the current projects are not designed to follow pre-cancer formation in individuals with risky BRCA1/2 alleles or some other notorious cancer susceptibility variants, it should theoretically be possible to get insights into early BRCA-related disease by profiling both pre-cancer and germline samples.
It remains to be seen whether PCA can have the same impact for early-stage disease that TCGA has had in understanding cancer biology and developing targeted treatment. The future of this endeavor hangs on the success of the pilot projects already underway. But those involved are optimistic that the coordinated effort to understand early disease, combined with more powerful and affordable technologies, will reveal unprecedented pre-cancer clues.
"All things considered, it is personally very gratifying to me to see that things are moving so rapidly," Srivastava noted. "We could not have done these things 20 years ago or maybe 10 years ago."