Finding cancer early can often mean the difference between life and death. Doctors stress the importance of colonoscopies, mammograms, and prostate checks, and patients are warned about the need for constant vigilance, like checking for breast lumps. Some companies now offer genetic tests to give patients a better idea of which gene mutations they carry that may influence their risk of developing certain cancers, and most oncologists say having a family history of cancer can be just as important for calculating risk as knowing whether the patient smokes or has tested positive for human papillomavirus. Yet, even armed with all that information, cancer can slip by the most vigilant doctors and patients.
"Almost any cancer is treatable successfully if you get it early enough. So there has to be, in my mind, a huge emphasis on early detection," says the University of Georgia's Michael "Hawkeye" Pierce, who conducts research on early detection biomarkers for pancreatic cancer. "If we had some blood tests that assessed risk for certain populations and got people in to be screened by more sophisticated means like PET scans, that would be tremendous. But we need to add as many weapons to our quiver as we can get, to target specific types of cells and to kill them."
Pancreatic cancer, by far one of the most aggressive cancers, has an average first-year survival rate of 24 percent. The five-year survival rate is 5 percent. And yet even a disease like pancreatic cancer can be stopped if it is caught early enough. Recent data show that most pancreatic cancers have been growing for about 20 years by the time they are diagnosed — they may be tiny and buried deep in the pancreas, and patients may be asymptomatic for many years. By the time those tumors are large enough to be found, treatment is difficult and rarely successful. "But the good news is that there is a window of time, that if we could screen people — especially people who are at risk, maybe familial groups of people with some huge risk factors, if we could focus in on them — there's good reason to believe that we could find these things a little earlier," Pierce says.
With the ultimate goal of arming clinicians with accurate, non-invasive clinical tests to detect cancer at its earliest stages, researchers are employing a number of sophisticated 'omics strategies. The National Cancer Institute recently awarded grants for a consortium dedicated to the use of genomics and proteomics in combination to find and validate cancer biomarkers. Other researchers are turning to the immune system to see if it can provide early clues to cancer development. And in one of the newest approaches to cancer research, some scientists are looking to the glycome and the signatures it leaves in the blood.
Harnessing the immune system
Many researchers say that when it comes to fighting disease — even cancer — the immune system is a patient's best friend. Even when the immune system cannot fight the disease on its own — as it usually can't with cancer — it still sends up signals that a person is infected, which can, in theory, be detected by clinicians. "Going back to the basics, we know from multiple lines of evidence that the immune system is capable of recognizing the tumor at the early stages of tumor development," says Samir Hanash from the Fred Hutchinson Cancer Research Center. Hanash published an editorial in the April issue of the journal Cancer Epidemiology, Biomarkers, & Prevention, saying that progress has been made to identify a number of antigens that could serve as detection tools for cancer.
"This is a situation that I compare to what goes on in HIV," he says. "Basically, the immune system recognizes the presence of the virus and it mounts an immune response, and we take advantage of this immune response to actually detect the HIV through seropositivity. And the analogy continues to the extent that the kind of immune response that you have in HIV does not actually get rid of the virus, but it allows you to make a diagnosis." Similarly, in cancer, he adds, the immune system recognizes that there is something wrong with the body and sends out signals in the form of antigens, though it seems unable to resolve the problem by itself.
Hanash says the question then becomes how to turn the immune response into a diagnostic. What needs to be done, he adds, is to identify the most informative antigens associated with a particular cancer, and once they are found, they can then be put on chips. "And then you can imagine having a chip which contains the antigens for lung cancer, colon cancer, breast cancer, pancreatic cancer, what have you. And eventually, as part of your routine checkup, if there's seropositivity against the breast cancer antigens, this would be an indication to want to do an imaging study to confirm or rule out that there is breast cancer," Hanash says.
At present, researchers are busy identifying panels of antigens, validating them, and performing studies to show their clinical utility, he adds. The data look promising, but there are challenges to overcome to bring this kind of test to the clinic — doing a test in the lab is one thing, but showing that it is reliable in a variable clinical setting is not so easy. "It's not enough to say that this set of markers can distinguish people with disease from people who don't have disease," Hanash says. "You have to indicate what the specifics of the clinical application are — what age group, what would be the indication that they would get the test in the first place, whether people are asymptomatic or getting the test on account of their family history, or their age or on account of their smoking behavior or something, that this test would work for them."
The search for glycans
At her glycomics lab at New York University, Margaret Huflejt is another one of those researchers looking to use the immune system as a tool in the early detection of cancer. As part of her work with NCI's Tumor Glycome Laboratories, Huflejt is seeking to determine the validity and utility of diagnostic or prognostic anti-glycan auto-antibody signatures in cancer patients, using printed glycan arrays — a high-throughput technology containing multiple probes made up of synthetic carbohydrates. "We print glycans on our glyco-chips and we then query human serum for the presence of natural, anti-glycan auto-antibodies that would be binding to the glycan-probes on the glyco-chips," Huflejt says. "Using our glyco-chips, we ultimately immuno-profile the anti-carbohydrate auto-antibodies of the individual, thus gaining information about his or her immune health status. We believe this will be fabulous technology in the service of the detection of pathology, and in following patients' responses to therapies."
The immune system knows everything and is constantly sending out signals, she adds, so paying attention to those signals is important for finding cancer biomarkers. The immune system sees the surface of the cell, and the glycans that it reads on the cell surface give it an indication of what's going on inside the cell, including any genes that are mutated, abnormally expressed, or abnormally located inside the cell.
This is a new approach to cancer detection, Huflejt says, but already she is seeing results. "While developing methods for early detection of cancer and cancer risk, we should be in parallel identifying biological targets, which suggest ways to protect those individuals whom we identify are at risk," she says. In this case, cancer should be treated almost like diabetes, in that clinicians know how to measure for glucose levels, and know what to do when those levels hit a certain limit. "In the situation where we are looking for the immune-identifiers of risk, we would be looking for the auto-antibodies whose proportions are already altered as the result of a disease — because the immune system has seen the problem and it has to do something about it," Huflejt says. "So there's going to be an imbalance among these auto-antibodies in high-risk patients as compared to healthy individuals. Persistence of these alterations would mean a physician should call back this patient and determine the best course of action; for example, starting a preventive therapy, or removing a suspicious tissue as opposed to leaving it alone because the original biopsy said it was benign."
Eventually, she adds, if researchers learn enough about the immune response to cancer and what the various anti-glycan auto-antibodies indicate, it may be possible to develop vaccines in conjunction with early detection diagnostics.
Northeastern University's William Hancock — co-chair of the Tumor Glycome Labs project — says that part of the challenge of finding cancer biomarkers in the blood for early detection is that blood is a complex fluid to analyze. His work in glycomics — which he dubs glyco-proteomics — is different than Huflejt's in that he and his team analyze proteins rather than auto-antibodies. "Until recently, this was a difficult area, and it still is," Hancock says. "People would do one of two things — they'd capture the glycoprotein and strip off the glycan, and, of course, they didn't know what the glycan was. Or the other approach is that you clean the glycans off, and throw away the proteins. The problem is you've lost the protein interaction, which can give you some idea of the tissue of origin." Hancock says he and his team decided to analyze the glycoproteins in blood as potential markers for early detection, in order to get the value of the protein and glycan information together.
"If we are clever enough, we can develop diagnostic tests or measurement tests so that we can not only find a particular protein in the blood, but a protein that has a unique glyco-signature on it," says Georgia's Pierce, another member of the Glycome Labs project. "Looking for the signature adds another dimension of specificity." Instead of just looking for a protein in the blood, he and his team are looking for specific proteins attached to specific glycans, which will not only tell a clinician if there are signs of possible cancer, but what the tissue of origin is and possibly at what stage the cancer is.
If researchers find that those markers are on the cells themselves, there may be a way to develop therapeutics by targeting the glycans with monoclonal antibodies. "What is it on these cells that's unique? Can we exploit that by finding those molecules that the cell is releasing that tells us that the cancer is there?" Pierce asks. "And then if the molecule is still on the surface, can we develop reagents to go in and direct chemotherapy to kill that cell?"
The metabolomics approach
Metabolomics might also be used to detect cancer in its early stages. In February, Robert Gardiner at the University of Queensland Centre for Clinical Research in Australia and his colleagues published a review in the Korean Journal of Urology proposing metabolomics as an early and non-invasive approach for the detection of prostate cancer.
Currently, serum prostate-specific antigen remains the standard initial test to detect prostate cancer in men, but as a biomarker, it leaves much to be desired — the test gives as many false negatives as false positives, and does not provide prognostic information. Using a metabolomic approach, Gardiner and his team are hoping to find a more sensitive and specific set of markers for the disease. They are conducting studies with nuclear magnetic resonance technology, which allows for the concurrent detection of multiple metabolites. "We have been trying to detect the presence of prostate cancer in patients using ejaculate for a number of years using molecular techniques, and our best results have been with a combination of the non-coding RNA PCA3 and Hepsin RNA, together with serum PSA," Gardiner says. "The presence of different mixtures of metabolites as a result of changed metabolism with cancers and other conditions has been known for a long time, and various metabolites have been reported in the past to be different in the ejaculate of men with prostate cancer."
In addition, Gardiner says, the "metabolic milieu" of prostate cancer is the result of different genes and proteins interacting, and NMR analysis allows for a comprehensive assessment of the changes created by those interactions, to provide a clinician with a more complete evaluation of the patient, rather than the "reductionist approach" of testing one marker at a time. "The metabonomic findings incorporating multiple discriminating parameters then can be added to the other two markers with the best track record — serum PSA and PCA3 in prostatic fluid — to produce a profile," he adds.
Although it is still experimental, Gardiner hopes his research will lead to a test that can be used in the clinic. "We are not the first to study ejaculate using NMR, but I believe that our comprehensive and long-term focused research commitment will result in the provision of a non-invasive diagnostic test," he says. "We hope that a test or tests resulting from our research will be another factor, as a more reliable indicator of prostate cancer status, to integrate with other considerations essential in formulating individual patient care."
In this zeal to find cancer earlier and earlier, however, are researchers creating a bigger problem? Over-diagnosis and false positives are already a problem for clinicians and cancer patients, with increasingly sophisticated imaging technology sometimes showing doctors things that aren't there, or benign lumps or nodules that could be left alone with no harm to the patient. These benign findings can sometimes be interpreted as being cancerous, leading to unnecessary treatments or biopsies that can be painful and expensive. And if these early detection tests are less than perfect, the worry is that they could end up adding to the problem instead of ameliorating it.
"These tests are prospective in nature. We don't have the data for a side-by-side comparison to say, 'This is better, this is worse.' But we can envision what the applications would be," says the Hutch's Hanash. "Initially it would be a test that would be used in conjunction with the standard imaging modality — whether mammography for breast cancer or CT scans for lung cancer — but where the performance is not really perfect. And so having a blood test that helps confirm cancer or backs up what you see in imaging can say, 'No, this looks OK,' or 'The test looks positive, you'd better get a biopsy done.'" Eventually, he adds, when the early detection biomarker blood test proves its utility in the clinic, doctors could forgo the current standard imaging procedures by giving their patients the blood test first and letting those results determine whether further testing or a biopsy is in order.
"Those are the same risks that are applicable right now to imaging modalities," Hanash says. "There's a lot of concern right now that for CT scans of the lung, even though there's evidence that you do save lives, a very large percentage of smokers who go to get a CT scan, they have something on their CT that most of the time is not cancer. So in that sense, there's a lot of false positives associated with CT scans. That creates a tremendous anxiety on the one hand, because it creates a need to go in and get a biopsy, or then you say, 'Wait a year, or six months,' and you can imagine what people go through knowing they have something in their lung and not knowing if it's cancer or not." These types of situations could be equally applicable to a blood test, he adds, but in the end, it's more likely that such early detection tests will actually cut down on the number of false positives because they add another method of testing to what a doctor can offer a patient.
Similarly, Northeastern's Hancock says he and his team hope their work in breast cancer one day leads to early detection blood tests for women who have had an "unsatisfactory mammography" and whose doctors are unsure if a biopsy is the next logical step. "If we're successful, they could do an actual glycoprotein test — then they really want to figure out if it's the less serious in situ form, or if it's metastasized," he says. "It really is meant as a complement to existing tests. This test does not cause a clinician to do something — it's to inform the physician to have more information and have better options to treat the patient correctly."
From talking to oncologists, Georgia's Pierce says clinicians need every tool they can get. "We expect the glyco-markers to be better than what they have, because what they have isn't very good. Early diagnosis is what we're really looking for, and it's just in the past five years that we've conceived a way to do that," he adds. "Maybe we can get to early detection, and maybe we can complement the tests that are out there."