How it works
What these tests are looking for
A clear scan, a blood test, and a sequencing map, in ordinary language. The details under “For clinicians” use the same facts with the usual terms. Nothing here names a product or tells a person what treatment to take.
What does “no evidence of disease” really mean?
“No evidence of disease” means the scan did not show a tumor. It does not mean every cancer cell is gone. A scan has to see a lump. Cells far smaller than that lump can still be there.
“To the human eye, a clear scan can still hide around 200 million cancer cells.”
That gap is a blind spot in oncology. MRD testing, especially ultrasensitive MRD, looks in the blood for tumor DNA the scan did not show.
For clinicians
NED is a radiographic and clinical statement, not a molecular one. The figure of about 200 million cells is Dr. Tan’s illustration of the distance between the resolution of CT and residual disease. It is not a validated cutoff, and it is not taken from the paper cited later on this page. Ultrasensitive MRD assays are built to detect a smaller fraction of tumor-derived DNA. This site does not rank assays.
Whole exome and whole genome sequencing
Think of the genome as a long city street. Whole exome sequencing reads only the protein-coding exons, about 1–2% of the genome. Whole genome sequencing reads the whole street. A test built from the tumor can only follow variants that were found when the tumor was sequenced. More variants to follow means more chances to catch a faint signal in the blood. This is the idea, not a score for any named test.
Whole exome sequencing reads the protein-coding exons, the teal blocks, about 1–2% of the genome.
Whole genome sequencing reads the whole strand, so it can find far more tumor-specific variants to track.
For clinicians
Exome capture targets coding sequence, conventionally described as about 1–2% of the genome. Whole-genome sequencing also covers noncoding sequence, so a tumor-informed panel built from it can include variants an exome would miss. The chance of detecting a trace of ctDNA rises, conceptually, as the number of independently informative tumor-specific variants increases. That statement is not a limit of detection.
Why the number of variants matters
Alan Tan, MD, uses a sidewalk in light rain. With 16 sensors, most drops miss, and it can look dry. With 1,800 or more sensors, standing in for 1,800 or more tumor variants, a miss becomes rare. The 16 and the 1,800 are parts of that picture. They are not the specifications of a test.
Detected less often in this study
Less than 50% of patients. In the picture, a light drizzle.
- Kidney
- Prostate
- Thyroid
- Primary brain
Detected more often in this study
More than 75% of patients with advanced disease. In the picture, a downpour.
- Pancreas
- Ovary
- Colorectal
- Bladder
- Gastroesophageal
- Breast
- Melanoma
- Liver
- Head and neck
Those two groups are the advanced cancers named by Bettegowda and colleagues. In their study, ctDNA was detectable in more than 75% of patients with advanced pancreatic, ovarian, colorectal, bladder, gastroesophageal, breast, melanoma, hepatocellular, and head and neck cancers, and in less than 50% of primary brain, renal, prostate, or thyroid cancers. The paper also notes that some tumor types had few patients, so this is not a ranked list inside either group. Kidney cancer is the renal group. Lung adenocarcinoma and sarcoma are not placed here, because that paper does not report a detection rate for them.
Bettegowda C, et al. Sci Transl Med. 2014;6(224):224ra24. PMID 24553385. doi 10.1126/scitranslmed.3007094
In the picture, a light drizzle is a cancer that releases less tumor DNA into the blood, and a downpour is one that releases more. Tracking the most variants is the best way to catch the low shedders. “Low shedders” here means the illustration’s name for the cancers detected less often in that study.
For clinicians
Bettegowda et al. used digital PCR after tumor genotyping in 640 patients and reported detection frequency, not a universal shedding rate. The abstract’s advanced-disease split is the only ordering used above. The same paper reported ctDNA in 73%, 57%, 48%, and 50% of patients with localized colorectal, gastroesophageal, pancreatic, and breast adenocarcinoma, respectively. Those localized figures are a stage comparison, not part of the drizzle-and-downpour drawing. The sensor counts, 16 and 1,800 or more, are Dr. Tan’s analogy for variant tracking. They are not assay specifications.
Video transcript
A person walks through New York City in light rain.
With only 16 sensors on the sidewalk, most raindrops fall through the cracks. It can look as if it is not raining.
With 1,800 or more sensors, meaning 1,800 or more tumor variants to track, missing that signal becomes very rare.
Bettegowda and colleagues, Science Translational Medicine, 2014, PMID 24553385: tumor DNA was found in the blood of more than 75 percent of people with several advanced cancers, and in less than half of people with kidney, prostate, thyroid, or primary brain cancers.
A light drizzle stands for a cancer detected less often. A downpour stands for one detected more often. Tracking the most variants is the best way to catch the low shedders.
This is an illustration of a concept, not a description of any specific test.
A urine test in bladder cancer
After treatment meant to remove all the cancer that doctors can see — surgery, chemotherapy, immunotherapy, radiation, or a combination — scans and cystoscopy can look clear. MRD (minimal residual disease, sometimes called molecular residual disease) means a laboratory test still found traces of cancer that are too small for a scan to show. In bladder and urothelial cancer, that test is often done on a blood sample (circulating tumor DNA, or ctDNA) or on a urine sample (urine tumor DNA, or utDNA). A positive result means cancer DNA was found. In several studies, people with a positive result were more likely to have cancer left behind or to have it come back. It does not automatically mean the cancer is already visible again, and it does not by itself decide the next treatment. A negative result means the test did not find cancer DNA at that time. In several studies that was reassuring, but it does not prove that every last cancer cell is gone. No urine or blood test has replaced cystoscopy or scans as the sole way to watch for bladder cancer. If you have a result to discuss, it should be interpreted with your clinician, in the context of your diagnosis and the specific test that was used.
A urine DNA result is not a treatment decision by itself, and it has not replaced cystoscopy or scans. The longer explanation, including field cancerization, is on urine tumor DNA.
From a tube of blood to a result
The same steps, in plain language. Open a card for the clinical wording.
A blood draw
A tube of blood is taken. The test uses the liquid part of the blood.
For clinicians
After a tumor-informed assay has been designed, later tests use plasma. They do not require a new biopsy at every time point.
DNA in the blood
Tiny pieces of DNA float in that liquid. Some can come from cancer cells. Those pieces are called circulating tumor DNA.
For clinicians
Cell-free DNA is the total circulating DNA. ctDNA is the tumor-derived fraction. Most cell-free DNA is not from the tumor.
Two ways to look
One test is built from the cancer already removed or biopsied, so it knows which changes to follow. Another looks for a standard list of changes and does not need that tissue.
For clinicians
Tumor-informed assays track variants found in that person’s tumor. Tumor-naive assays use a fixed panel or pattern and do not need a new tumor sequence.
Found, or not found
The report can say a signal was found, or that it was not found. Not found does not prove the cancer is gone.
For clinicians
A positive call clears the assay’s threshold. A negative call does not. A result below that threshold is not proof that every cancer cell is gone.
What a study may do next
Some studies keep watching. Some change treatment after a found or not-found result. This drawing is not a treatment plan.
For clinicians
Protocols differ. Some continue surveillance. Some escalate or de-escalate treatment. The choice is defined by the protocol and the clinical team.
When studies usually draw blood
Timing changes with the cancer and the study. This line is a sketch of windows that often show up in papers. It is not a schedule to follow.
Before treatment
Some studies take blood at diagnosis, to see whether any tumor DNA is detectable.
Soon after treatment
A common research window is a few weeks after surgery or the end of curative therapy, before a decision about more treatment.
During treatment
Repeat blood draws can show whether a signal that was present has fallen below the test’s cutoff.
Later checkups
Later draws are often months apart. The spacing is set by that study, not by this site.
For clinicians
Papers often label the post-therapy window a landmark draw. Later draws are surveillance time points. Intervals are protocol-specific. A sketch of those windows is not a guideline and not a recommendation for an individual patient.