"Blood test for cancer" female version of Jobs convicted, is liquid biopsy reliable? NEJM review

Early this morning, Elizabeth Holmes, the founder of Theranos, a US in vitro testing company that was valued at $9 billion, was found guilty of four fraud charges by a jury and could face up to 20 years in prison
.

Holmes is known as the "female version of Jobs" because of her charisma and a pullover dress similar to Apple founder Jobs
.

 In fact, although the myth of “blood test for cancer” has long been shattered, liquid biopsy, especially liquid biopsy involving free DNA (cfDNA) in plasma, is rapidly becoming an important minimally invasive aid to standard tumor biopsy, and in some cases even It is a potential alternative
.

Liquid biopsy is becoming a useful tool for molecular testing, gaining new insights about tumor heterogeneity, and cancer detection and surveillance
.

 "New England Journal of Medicine" (NEJM) once published a review "Application of Cell-Free DNA Analysis in Cancer Treatment", summarizing the application of cfDNA analysis in cancer diagnosis and treatment
.

We introduce its main content here
.

To read the full text translation, please visit the "NEJM Medical Frontier" APP, official website or click on the picture of the WeChat applet
.


Liquid biopsy Although liquid biopsy usually refers to the analysis of cfDNA in peripheral blood, the term also includes the separation and analysis of tumor-derived materials (such as DNA, RNA, and even intact cells) from blood or other body fluids (Figure 1)
.

For example, intact circulating tumor cells enter the bloodstream at a low frequency (in patients with metastatic cancer, there are usually <10 circulating tumor cells per milliliter of blood)
.

Tumor cells also release subcellular particles called exosomes, or vesicles wrapped in the outer cell membrane, into the bloodstream.
Exosomes contain tumor-specific proteins and nucleic acids
.

Free nucleic acids (including free RNA [which is less stable than DNA] and cfDNA [focus of this review]) are also released into the circulating blood
.

Blood is not the only body fluid available for liquid biopsy
.

Urine, stool, cerebrospinal fluid, saliva, pleural fluid, and ascites are all potential sources from which tumor-derived substances (including cfDNA) can be obtained
.

 Figure 1.
Overview of liquid biopsy methods.
Plasma-derived cfDNA is currently believed to be released into the blood through apoptosis or necrosis, and cfDNA is usually a double-stranded fragment with a length of 150 to 200 base pairs, corresponding to nucleosome-associated DNA10
.

CfDNA molecules in the circulation are quickly eliminated, with a half-life of 1 hour or less
.

CfDNA from normal cells is present at low levels in the plasma of healthy people (about 10-15 ng/mL on average), and under tissue stress conditions (including exercise, inflammation, surgery, or tissue damage), cfDNA levels may increase
.

 In cancer patients, cfDNA released from tumor cells is usually called circulating tumor DNA (ctDNA), which constitutes only a part of the entire cfDNA
.

The proportion of ctDNA in the entire cfDNA of cancer patients may vary greatly, from less than 0.
1% to more than 90%
.

Although the proportion of ctDNA in individual patients is often parallel to the tumor burden, significant variation has been observed between different patients with the same type of cancer, which may reflect the biological differences of individual tumors or differences in cell mortality
.

In addition, patients with different types of tumors have significant variations in the frequency of detectable ctDNA
.

Therefore, it is a big problem to detect and analyze ctDNA in the background of normal germline cfDNA
.

 Analysis of cfDNA The proportion of ctDNA in the background of normal cfDNA in cancer patients is usually very small and varies greatly from patient to patient
.

Therefore, an ultra-sensitive method is needed to detect mutations, copy number changes, or other changes in cfDNA with a very low variant-allele frequency (Figure 2)
.

In order to detect mutations at each point, mutation-specific techniques based on PCR analysis (such as BEAMing or droplet digital PCR [ddPCR] analysis) can identify and quantify changes in cfDNA with an allele frequency of ≤0.
01%
.

Next-generation sequencing methods are also customized for cfDNA, ranging from whole-genome or whole-exome sequencing to targeted sequencing of limited genomes
.

However, the sensitivity and specificity of these methods are limited by DNA polymerase error rates and sequencing reactions
.

Therefore, an improved method combining deep sequencing coverage, molecular barcode methods (in which a unique nucleotide barcode is used to label each input template DNA fragment), and error suppression algorithms improves the detection limit
.

 Figure 2.
Isolation and analysis of cfDNA clinical application cfDNA analysis is minimally invasive, providing a molecular profiling method for tumors that are difficult or unsafe to biopsy, and a practical method that can continuously monitor tumor DNA over time.
There are no risks and potential complications of standard tumor biopsy (Figure 3)
.

In addition, compared with needle aspiration biopsy of a single tumor lesion, cfDNA analysis can better detect the molecular heterogeneity of multiple different clonal populations in a patient's tumor
.

Finally, cfDNA analysis provides the possibility of tumor detection or monitoring for patients without clinically obvious diseases
.

 Figure 3.
Clinical application of cfDNA analysis.
Diagnosis and molecular profiling.
Tumor molecular profiling for selecting treatments has become a basic examination in cancer medicine
.

No invasive biopsy is required, and the molecular profile of a patient’s tumor can be assessed through a simple blood draw.
This possibility makes cfDNA analysis an attractive tool
.

However, a key initial question is whether the mutation profile established by cfDNA testing can reliably reproduce the mutation profile obtained from direct tumor biopsy
.

Early studies based on a small number of patient samples suggest that there is low consistency between DNA changes detected in tumor and plasma samples from the same patient
.

However, the validity of these studies is affected by the following shortcomings: tumor and plasma samples are usually not collected at the same time, and the molecular evolution of tumors can make potential differences
.

In addition, many plasma samples have poor collection times, such as during treatment, when ctDNA is at its lowest level
.

Inconsistent key changes are most common in patients with low ctDNA levels, and low levels make changes more difficult to detect
.

In addition, how to obtain true positive or true negative results in patients with low cfDNA levels remains an important challenge
.

 Although these data suggest that cfDNA testing may be a potential alternative or auxiliary to the standard operation of tumor biopsy, tumor biopsy is still the standard for initial pathological diagnosis and molecular testing
.

Tumor biopsy can provide histological interpretation and evaluation based on non-DNA changes (such as the expression of hormone receptors or other proteins), which is important for diagnosis and treatment decisions
.

In addition, a large series of studies have shown that about 15% of patients with metastatic cancer may not have enough ctDNA to obtain mutation profiles from plasma.
These values ​​vary with tumor types and tumor burden
.

Although improving the blood draw time for cfDNA analysis (for example, before treatment starts) can increase the output of cfDNA testing, as the ctDNA ratio approaches the detection limit of current technology, confidence in the existence of key changes decreases
.

This factor must be considered when interpreting clinical cfDNA testing
.

 Nevertheless, cfDNA testing can still play an important role in initial molecular testing, especially for patients whose standard tumor biopsy produces insufficient material for clinical sequencing, and as many as 20% to 25% of needle aspiration biopsies may have insufficient material
.

In this case, cfDNA testing for treatment selection is increasingly used as an alternative to repeated invasive biopsy, and may reveal treatment-instructive mutations to guide treatment decisions for these patients
.

Technological advances may make the relatively rare abnormal cfDNA test faster and cheaper, to identify mutations that are useful for treatment and predict the likelihood of remission (for example, the presence of microsatellite instability can predict T cell checkpoints) Suppress the response)
.

 For repeated or continuous detection after first-line or multi-line treatment, minimally invasive cfDNA analysis has many obvious advantages compared with repeated invasive tumor biopsy.
Repeated invasive tumor biopsy is less practical, less safe, and more cost-effective.
Low
.

In particular, liquid biopsy can identify new genetic changes that lead to acquired resistance, and then use a new generation of therapies for targeted therapy
.

 Treatment response tracking Although the ctDNA level of different patients may vary greatly, over time, the ctDNA level of an individual patient is closely related to changes in tumor burden and treatment response
.

Unlike many standard serum tumor markers in clinical applications (such as carcinoembryonic antigen [CEA] and cancer antigen 125 [CA 12-5], which have a half-life of several days to several weeks), the half-life of cfDNA in the blood circulation is relatively short (about 1 hour), which is very useful for determining the real-time tumor burden in response to treatment
.

Many standard clinical tumor markers have limited sensitivity and specificity, while tumor-specific clonal changes (that is, changes that exist in the original tumor clone and therefore exist in all tumor cells) can be monitored in plasma with high sensitivity and are patient tumors Peculiar
.

Some studies suggest that ctDNA levels may actually rise briefly after treatment begins, because tumor cell death leads to increased ctDNA release
.

However, within 1 to 2 weeks after the start of treatment, the level of ctDNA in patients who responded to treatment dropped sharply
.

 In fact, some studies suggest that changes in ctDNA may be better than standard tumor markers in predicting treatment response
.

In addition, elevated ctDNA levels may occur weeks to months before radiological progress
.

Therefore, due to the increasing popularity and cost-effectiveness of cfDNA monitoring technologies, their potential in detecting early evidence of response or progress may become very important for clinical management
.

 Monitoring of drug resistance and tumor heterogeneity The clinical benefits of precision cancer treatment are limited by the eventual emergence of acquired drug resistance
.

Generally speaking, acquired drug resistance comes from one or more tumor subclones that carry pre-existing changes in drug resistance, and changes in drug resistance appear under the selective pressure of treatment methods
.

For molecular changes that lead to clinical resistance, cfDNA analysis has become an effective tool for early detection and identification
.

A key advantage of cfDNA is the ability to detect molecular heterogeneity associated with drug resistance (Figure 4)
.

 Figure 4.
Tumor heterogeneity and cfDNA therapy may have cytotoxic effects on most tumor cells, but the growth of drug-resistant subgroups may occur, leading to dynamic changes in clonal abundance, and ultimately leading to disease progression
.

A single tumor biopsy specimen obtained during disease progression may only show a subset of drug-resistant clones (ie only green).
Follow-up treatments against this drug-resistant mechanism may lead to mixed clinical responses and the growth of other co-existing clones.
The treatment failed
.

cfDNA analysis has the potential to identify multiple coexisting resistance mechanisms and monitor the cloning dynamics during treatment
.

Clone changes (present in the original tumor clone and therefore in all cells) are shown in gray, and drug-resistant subclone changes are shown in other colors
.

 In fact, studies combining multiple tumor biopsy or autopsy specimens have shown that multiple unique drug resistance changes often coexist at different metastatic sites in individual patients, but they can all be detected in the cfDNA of a single plasma sample
.

Therefore, based on the results of a single tumor biopsy, targeting a single drug resistance mechanism may lead to a mixed clinical response, which is due to the growth of uncollected drug-resistant subclones in the biopsy specimen, and cfDNA testing may help guide treatment decisions
.

 In addition to being an important discovery tool, cfDNA analysis can also be used clinically to manage drug resistance
.

For example, cfDNA analysis can identify the coexistence of EGFR T790M with other resistance changes (such as MET amplification), and patients with such coexisting changes may be treated with third-generation EGFR inhibitors (such as osimertinib).
The benefit is less
.

 cfDNA analysis has also been used to track the clonal dynamics of different drug-resistant subclones during sequential treatment, and even after treatment is stopped
.

Incorporating real-time cfDNA analysis into clinical trials, and eventually into standard clinical management, may become a valuable tool for precision medicine
.

 The detection of residual lesions after surgery is one of the most transformative potential applications of cfDNA analysis.
One possibility is that it can detect the presence of tumors in patients without clinically obvious diseases—for example, as a screening tool for new cancers.
Early detection, or for the detection of tumor recurrence after surgery or adjuvant treatment
.

Usually, the main method that can cure solid tumors is surgical resection
.

If there are tumor cells remaining after the operation, it can lead to the eventual recurrence of the tumor
.

In high-risk patients, adjuvant chemotherapy can reduce the risk of recurrence
.

However, it is currently impossible to determine immediately after surgery which patients have residual disease and which patients have cured their disease
.

We do not routinely give adjuvant therapy in certain situations (such as stage II colorectal cancer patients with low clinical risk and no evidence of lymph node or distant metastasis), although about 15% of these patients may eventually relapse
.

Effective methods for detecting postoperative residual disease can prevent cured patients from receiving potentially toxic adjuvant chemotherapy, or can identify patients with residual disease who may benefit from adjuvant therapy (Figure 5)
.

 Figure 5.
Using cfDNA to detect residual lesions and postoperative treatment management.
Some key studies have shown that the detection of tumor-specific mutations in cfDNA after surgery can predict residual lesions and tumor recurrence in breast, lung, and pancreatic cancer
.

This method may become a key tool for postoperative management of cancer patients, but it needs to be tested in prospective clinical trials to evaluate the utility of postoperative residual ctDNA testing in guiding adjuvant chemotherapy (Figure 5)
.

 The advantage of liquid biopsy for early cancer detection is that it is possible to detect early cancer by performing simple blood tests on otherwise healthy, asymptomatic people
.

There is currently no mature technology to achieve this goal, but this method needs to have high sensitivity (to detect trace amounts of cfDNA or other substances released by precancerous lesions or early cancers), and it also needs to have high Specificity (to reduce false positive results in large-scale unaffected populations undergoing screening)
.

 Other difficulties complicate the work
.

First, since many types of cancer have common mutations in genes such as KRAS, BRAF, or TP53, it may be difficult to locate the cancer in a specific organ after a positive liquid biopsy result is obtained
.

In addition, benign lesions may have the same mutations as cancer, and the cfDNA of benign lesions may also fall off and enter the blood circulation
.

In fact, benign moles may have the same BRAF mutations seen in advanced cancers
.

Detectable mutations in cfDNA can also originate from abnormal and usually benign clonal populations in the bone marrow, produced through a process called CHIP (clonal hematopoiesis of uncertain potential)
.

The frequency of CHIP increases exponentially with age, and the incidence in people over 70 years old exceeds 10%
.

Therefore, CHIP is a common source of multiple detectable mutations in cfDNA, which poses a major challenge to this method
.

Therefore, other methods besides pure cfDNA mutation detection are being explored-including tumor-associated virus sequences (for example, in tumors associated with Epstein-Barr virus [EBV] infection and human papillomavirus infection) and DNA methylation changes- Early detection
.

 Recent studies have shown that liquid biopsy can be used to detect cancer early
.

Cohen et al.
developed a screening method that combines the detection of mutations in cfDNA with circulating protein markers, called CancerSEEK
.

Among 1,005 patients with non-metastatic, clinically detectable tumors with 8 common types of tumors, the median proportion of patients with positive test results was 70%, and the sensitivity ranged from 69% to 98%
.

The overall specificity is more than 99%, and only 7 of the 812 healthy controls tested positive
.

In addition, using supervised machine learning, in 83% of cases, researchers used detected mutations and protein profiles to locate the cancer to its primary site
.

 Although these results are encouraging, it is important to evaluate in this more representative screening population of asymptomatic patients
.

Although the liquid biopsy screening method needs to be further improved and validated in prospective clinical trials, cfDNA analysis as an early detection tool provides potentially transformative advances in cancer medicine
.

Reference Corcoran RB, Chabner BA.
Application of cell-free DNA analysis to cancertreatment.
N Engl J Med 2018;379:1754-65.