Timely cancer detection paves the way for interventions that are less toxic, less disruptive to daily life, and substantially less costly than treatment in later stages.

By Moshe Szyf, PhD

Early cancer detection stands as a crucial cornerstone in the ongoing battle against this relentless disease. Its significance lies in the potential to drastically reduce the morbidity and mortality associated with various types of cancer, including those that afflict a vast number of individuals, such as breast, liver, and cervical cancers.

Timely cancer identification paves the way for interventions that are less toxic, less disruptive to daily life, and substantially less costly than treatment in later stages. However, despite tremendous strides in imaging technology and the discovery of cancer-specific antigens, the realization of population-wide early detection programs remains an elusive goal. This is particularly true for cancers where clinical symptoms only manifest in advanced stages, such as cervical, pancreatic, and liver cancers. Unfortunately, many cancers are still detected when they have already reached an advanced and often untreatable state, leading to dire consequences.

The Ins and Outs of Cancer Detection

Imaging techniques, notably MRI and CT scans, have proven highly effective in spotting even small lesions. However, their utility is contingent on the presence of a visible lesion and their throughput remains relatively low. These methods demand local access to specialized machines and the expertise to operate them in proximity to the screened population. The scalability of such an approach for frequent screening across diverse geographic areas, especially in underdeveloped regions, is questionable. On the other hand, blood-based tests hold the potential to be deployed across wide regions, requiring minimal expertise at the point of care, and can be sent to highly specialized labs for processing and electronic reporting back to the point of care.

Several blood-based antigen-antibody screens have been developed for early cancer detection, including alpha-fetoprotein (AFP) for liver cancer, prostate-specific antigen (PSA) for prostate cancer, CA-125 for ovarian cancer, CEA for colorectal and rectal cancers, and CA19-9 for pancreatic cancer. However, these tests suffer from limitations in terms of sensitivity, particularly in early stages, and specificity, leading to inaccuracies and high rates of false positives and false negatives. To make early cancer detection a reliable public health instrument, there is an urgent need for blood-based tests with higher accuracy and reduced false positive and false negative rates. Moreover, there is a critical need for tests that can detect cancer in its early stages.

DNA, the resilient and durable biological molecule that contains the proximal code governing physiological and pathological functions, has emerged as a promising avenue for population-wide cancer screening. DNA mutations are a hallmark of cancer, and alterations in DNA methylation, a chemical modification added to discrete positions in DNA to regulate gene activity, are dramatically altered in cancer. Importantly, changes in DNA methylation patterns occur early in cancer progression and are a hallmark of nearly all cancers. Highly accurate and sensitive methods exist for detecting changes in DNA sequence and DNA methylation. However, the challenge lies in the fact that these alterations occur within the tumor itself, making their utilization for early cancer detection challenging.

DNA and Cancer Detection

For many decades, it has been known that DNA from dying tumor cells leaks into the bloodstream, making its presence a potential marker of cancer. However, the presence of other DNA originating from normal tissues, white blood cells, and monocytes can dilute and obscure tumor-derived DNA, rendering it an inaccurate measure of cancer. Distinguishing tumor-derived DNA from other sources in the bloodstream is the key to a viable cancer detection method. Several research groups are exploring the use of cancer-specific mutations or changes in DNA methylation to detect cancer-derived cell-free DNA amidst a background of non-tumor DNA. Earlier efforts focused on cancer-specific mutations, but the variable abundance of these mutations in tumors, coupled with the excess of non-tumor DNA, reduces the sensitivity of these tests, particularly in early stages. DNA methylation alterations, on the other hand, are a common feature of most cancers and occur early in cancer progression, making them a more reliable marker for cancer DNA. Tests based on both global changes in DNA methylation and sequence-specific differences between cancer and non-cancer DNA have formed the foundation of blood-based cancer detection tests.

One approach uses a “candidate gene” strategy, targeting genes known to be differentially methylated in cancer. For example, the SEPT9 gene, known to be differentially methylated in colorectal cancer, is FDA-approved for colorectal cancer (epi proColon). Additionally, a test that combines cancer-specific genetic mutations and differentially methylated candidate genes has gained FDA approval for colorectal cancer detection (Cologuard). Both tests are designed to detect cancer DNA in stool samples and are primarily focused on specific cancers and high-risk groups.

Bioinformatic analysis of genome-wide DNA methylation data from tumor and non-tumor DNA has enabled the discovery of cancer type-specific DNA methylation profiles that can be harnessed for test development. By employing machine learning techniques, several research groups have identified a combination of DNA regions that are differentially methylated between cancer and normal tissue, allowing for the classification of cancer-derived DNA from non-cancer DNA. Such an approach has been used to develop a blood test capable of detecting over 50 types of cancer. The test has been deemed accurate enough for deployment as a multi-cancer screening tool, primarily targeting individuals at higher risk of cancer, including those aged 50 and above, who may not exhibit any symptoms. The test has exhibited high specificity, with sensitivity varying across different cancer types and stages, being lower in earlier stages but improving as cancer progresses.

One of the main challenges in detecting cancer using cell-free DNA is the variable background of non-cancer DNA, arising from technical issues (varying quality of separating white blood cells from plasma) and clinical reasons. To address this challenge, it is critical to identify positions in DNA that exhibit a categorical difference in methylation between cancer and normal tissue–being completely unmethylated in normal tissue and fully methylated in cancer. Such a “black and white” difference in DNA methylation allows for the identification of cancer DNA, even in the presence of a significant excess of unmethylated DNA, using next-generation sequencing. This approach has also been employed in the development of HKG’s epiLiver, test. This test is designed for the early detection of liver cancer. EpiLiver focuses on just five small DNA regions, which, by limiting the number of regions in the assay, allows for greater sequencing depth and the detection of even sparsely methylated molecules, potentially increasing the test’s sensitivity.

Much Promise but Challenges Remain

While “cell-free” DNA methylation markers show promise, they also present clear challenges. First, the sensitivity of the test, particularly for early-stage cancer, needs improvement, as early detection is the primary goal. Second, robustness, throughput, and cost-effectiveness are essential for ensuring widespread coverage across diverse socio-economic groups and geographic regions. Third, the question arises of whether tests targeting specific “high-risk” groups or pan-cancer tests with the potential to detect a wide range of cancers are more impactful in reducing the burden of cancer. While pan-cancer tests are appealing due to their ability to potentially detect unexpected cancers, the results may be more challenging to interpret and produce. Fourth, despite knowing that early detection of certain cancers significantly improves survival and reduces morbidity, more clinical evidence is required to demonstrate that community-wide screening and very early detection indeed reduce overall cancer mortality.

Finally, we have very limited experience and knowledge of how to treat “very early detected” cancer. The available treatments, which were primarily developed for later stages of cancer, may prove inadequate for addressing “extremely early” cancer. Hopefully, as new tools open new possibilities for early detection, they will catalyze the development of novel clinical approaches, ultimately leading to a paradigm shift in our strategies for cancer detection and treatment.

About the Author Moshe Szyf, PhD, is the Founder and CEO of HKG Epitherapeutics