Cancer Diagnosis — It’s in Your Blood
The benefits of detecting and treating cancer at early developmental stages have long been observed by clinical oncologists. Colonoscopies as well as other methods for early detection have been estimated to prevent as many as 60% of deaths attributable to colorectal cancer by identifying and removing precancerous growths1. More broadly, patients with early stage cancer have experienced a fivefold to tenfold higher rate of survival over those with late stage cancers, despite having undergone the very same treatment regimens2. Nevertheless, diagnostic tests for early detection of cancer are available only for a small minority of cancer types.
Two teams of researchers are angling to change this by engineering diagnostic technologies that can simultaneously detect multiple cancer types at early developmental stages. Their strategy revolves around a fundamental property of tumor growth. As new tumor cells grow, older tumor cells burst open and spill their DNA into the surrounding tissue. Some of the DNA diffuses into the bloodstream, where it can be readily extracted, sequenced and analyzed. This so-called circulating tumor DNA (ctDNA) can bear mutations that are highly prevalent in cancer but absent in healthy tissue, and that can — if observed — signal the presence of tumors.
The principal barrier to detection of cancer in its early stage of development is its seeming innocuous stature before progression to more advanced forms. Tumors in early stage cancers are typically too small for conventional imaging-based or biochemical-based diagnostics to detect. However, researchers can leverage time-tested methods for exponentially generating many copies of a trace amount of existing DNA in a blood draw to amplify the signal to noise and confidently examine whether ctDNA is or is not present in the blood.
GRAIL — the First on the Scene
GRAIL, a clinical-stage diagnostics company based in California, is the first company solely committed to the development and commercialization of a blood-based multicancer early detection platform. It was initially spun out of the well-known next-generation sequencing company Illumina to become an independent company in 2016. GRAIL quickly raised more than $900 million — backed by the likes of Bill Gates and Jeff Bezos — in one of the largest equity financing rounds in the history of the medical technology industry. They soon initiated the Circulating Cell-free Genome Atlas project, a 15,000-person clinical trial in which they planned to identify a panel of mutations that could most readily distinguish ctDNA in the blood of patients with cancer from free-floating DNA shed by healthy tissues.
At the 2019 American Society of Clinical Oncology annual meeting in Chicago, GRAIL revealed the performance of its diagnostic tool based on an interim analysis of 1,422 patients with cancer — representing more than 20 cancer types — in addition to 879 undiagnosed participants. The platform was able to correctly detect 76% of all cancers, representing nearly two-thirds of all cancer deaths in the U.S., according to the company3. Its ability to detect different stages of cancer ranged from 34% of stage I (earliest) cancers to 92% of stage IV (latest) cancers. Interestingly, it was able to correctly recognize 77% of stage II (early) cancers. Furthermore, it correctly identified the tissue from which the ctDNA originated (liver, lung, pancreas, etc.) in 90% of cases.
Thrive — A Cancer SEEKer
Thrive, a Massachusetts-based diagnostics company, officially launched earlier this year, on May 30, as it secured $110 million in its first significant venture capital financing round4. The company was co-founded by a team of pioneering cancer genomics researchers whose labs are at the Johns Hopkins University School of Medicine: Bert Vogelstein, Kenneth Kinzler, and Nickolas Papadopoulos. Their platform, called CancerSEEK, leverages both ctDNA and abnormal proteins found in the blood to determine the presence or absence of cancer and track its tissue of origin. A 2018 publication in Science5 revealed its performance, based on an analysis of 1,005 patients with cancer representing eight cancer types that are common in Western populations and that lack routinely used blood-based early detection tests. The ability to detect different stages of cancer ranged from about 43% of stage I (earliest) cancers to 78% of stage III (latest) cancers. Similar to GRAIL’s platform, CancerSEEK was able to accurately detect 73% of stage II (early) cancers. The ability to correctly identify tissue of origin ranged from 44% for liver cancers to 84% of colorectal cancers. Notably, CancerSEEK detected about 72% of pancreatic cancers and was able to determine its tissue of origin with 81% accuracy. This result is highly encouraging, given that the very poor prognosis of pancreatic cancer is largely attributed to a late stage diagnosis in locally advanced or metastatic forms. Thrive is currently testing CancerSEEK in a 10,000-person long-term clinical trial called DETECT, to gauge its predictive value in previously undiagnosed patients and to fine tune guidelines for defining clinically actionable insights.
Taken together, these technologies represent nascent forms of minimally invasive pan-cancer diagnostics that bridge a wide chasm between the time when cancer is most treatable and the much later time when cancers are typically recognized. Though recent findings in thousand-person cohorts point to the high sensitivity and specificity of the platforms mentioned here, completion of ongoing clinical trials on much larger cohorts is required to determine their true false positive (overdiagnosis of cancer-free patients) and false negative (underdiagnosis of patients with cancer) rate. If successful, GRAIL’s and Thrive’s early detection platforms could drastically improve disease management regimes for patients with cancer in the coming years.
- He, J. and Efron, J.E., 2011. Screening for colorectal cancer. Advances in Surgery, 45, pp. 31–44.
- Aravanis, A.M., Lee, M. and Klausner, R.D., 2017. Next-generation sequencing of circulating tumor DNA for early cancer detection. Cell, 168(4), pp. 571–574.
- Cohen, J.D., Li, L., Wang, Y., Thoburn, C., Afsari, B., Danilova, L., Douville, C., Javed, A.A., Wong, F., Mattox, A. and Hruban, R.H., 2018. Detection and localization of surgically resectable cancers with a multi-analyte blood test. Science, 359(6,378), pp. 926–930.
- A New Blood Test Tries to Detect Cancer Sooner
- A Blood Test for Alzheimer’s Disease? Breakthroughs and Limitations
- Engineering Magic Bullets for Pancreatic Cancer
Want to read more from the Johns Hopkins School of Medicine? Subscribe to the Biomedical Odyssey blog and receive new posts directly in your inbox.