Table of Contents
- Key Points
- What Is Liquid Biopsy and Why Does It Matter?
- How Was This Research Conducted?
- What Did the Researchers Find?
- How Accurate Are Liquid Biopsy Tests?
- The Quality of Evidence: A Critical Concern
- Risk of Bias in the Included Studies
- What This Means for Patients
- The Future Role of Artificial Intelligence
- Limitations of This Research
- Recommendations and Next Steps
- Frequently Asked Questions
- Source Information
Key Points
- An umbrella review of 42 systematic reviews (about 213,600 people) found 45.4% of 211 outcomes had excellent accuracy (AUC >0.90) for liquid biopsy.
- Every outcome was graded as very low quality evidence, meaning true diagnostic accuracy is likely to differ substantially from reported estimates.
- Risk of bias was very serious in 92.9% of outcomes; heterogeneity was never reported; publication bias appeared in 48.1%.
- Liquid biopsy uses a blood or urine sample and may avoid invasive tissue biopsies, but it is not yet ready for widespread clinical use.
- Patients interested in liquid biopsy should discuss the specific test, its accuracy for their condition, and its appropriateness with their healthcare provider.
What Is Liquid Biopsy and Why Does It Matter?
Liquid biopsy (LB) is a laboratory test performed on a fluid sample, such as blood or urine. The test analyzes molecular data from circulating cells and related entities, or directly from nucleic acids (the building blocks of genetic material).
Unlike traditional tissue biopsies, which require invasive procedures to remove a piece of tissue, liquid biopsy offers a minimally invasive approach. A simple venous blood sample can provide vital information about a patient's health status.
Liquid biopsy has been explored predominantly in oncology (cancer medicine) and related fields. The test holds potential diagnostic, prognostic, and predictive value. This means it may help doctors detect disease, predict how a disease will progress, and forecast how a patient will respond to treatment.
The technique has garnered significant attention across medical disciplines. Its versatility, accessibility, and potential to revolutionize precision oncology and personalized medicine have made it a major focus of research.
Liquid biopsy holds significant promise in the early detection and monitoring of various diseases, particularly cancer. Liquid biopsy detects circulating tumor cells (CTCs), cell-free DNA (cfDNA), microRNAs (miRNAs), and other molecular markers. Tumors shed these markers into the bloodstream or other bodily fluids. Liquid biopsy enables clinicians to detect cancer at its earliest stages, before the onset of clinical symptoms.
Early detection is crucial for improving patient outcomes. It allows for timely intervention and treatment strategies tailored to the specific molecular profile of the tumor.
Furthermore, liquid biopsy facilitates real-time monitoring of disease progression and response to treatment. This offers valuable insights into treatment efficacy and the emergence of drug resistance.
Liquid biopsy could bring promising innovations beyond oncology, extending its utility to various other medical specialties. In infectious diseases, it may enable the rapid detection of pathogens and monitoring of antimicrobial resistance, facilitating prompt and targeted therapeutic interventions.
In prenatal testing, the analysis of fetal DNA in maternal blood may offer a non-invasive alternative to traditional methods, reducing the risk of complications for both mother and fetus. Additionally, liquid biopsy holds promise in the field of organ transplantation, where it can aid in the early detection of graft rejection and guide personalized immunosuppressive therapies.
The significance of liquid biopsy in medicine is further underscored by its role in overcoming the limitations of conventional tissue biopsies. Tissue biopsies may be hindered by sampling bias, accessibility issues, and tumor heterogeneity (differences between cancer cells within the same tumor).
Liquid biopsy offers a comprehensive and dynamic snapshot of the entire tumor landscape. This allows for more accurate molecular characterization and treatment selection.
However, liquid biopsy may also suffer from potential issues. These include lack of standardized approaches, propensity to pre-analytical errors, and difficulties in the interpretation of data generated from multi-omics analysis (analysis of multiple types of molecular data).
In this complex scenario, this umbrella review aims to map and evaluate the evidence. An umbrella review is a systematic review of other systematic reviews on the same topic. The review covers the use of liquid biopsy in medicine across different medical specialties and conditions. The review also takes into account the multiple issues potentially emerging in this setting.
How Was This Research Conducted?
This umbrella review was conducted following the recommendations of the Cochrane handbook for systematic literature reviews. The researchers followed the Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) 2020 updated guidelines. The study protocol is freely available online.
The researchers used the PICOS model (participants, intervention, control, outcomes, study design) to define their inclusion criteria:
- Participants: Any patient population
- Intervention: Liquid biopsy, defined as a simple venous blood sample in which molecular analyses can be performed, such as the dosage of free tumoral cells, DNA, RNA, exosomes, or lysosomes
- Controls: None required
- Outcomes: All health outcomes reported in terms of accuracy (area under the curve, AUC) with sensitivity and specificity, or as precision (e.g., C-index). Accuracy data were considered when reported against the gold standard approach for a condition
- Study design: Systematic reviews of observational studies
The researchers excluded meta-analyses of intervention studies and single studies.
Several relevant bibliographic databases were comprehensively searched. These included Medline (via Ovid), Embase, and Web of Science from database inception up to October 1, 2023.
Study selection was independently carried out by four review authors, with consensus meetings to discuss studies for which divergent selection decisions were made. A third senior member of the review team was involved if necessary.
The studies selection process involved two steps. First, selection was based on title and abstracts. Then, selection of studies retrieved from this first step was based on the full-text manuscripts. The freely accessible software Rayyan was used for the title/abstract screening.
From the eligible full-text articles, the following data were extracted: first author name and affiliation, year of publication, journal name, and title of the manuscript. The extracted data also included the characteristics of the population considered, the name of the marker used in the liquid biopsy, and the outcome.
Data regarding estimates were extracted as reported in the original systematic review and reported as AUC. No studies reported data using C-indexes. These data were collected using a RedCap database. Data extraction was carried out by four authors and systematically double-checked by another author.
Four authors rated the methodological quality of the included systematic reviews. The authors used "A MeaSurement Tool to Assess systematic Reviews 2 (AMSTAR 2)." This tool ranks the quality of a meta-analysis in one of four categories. The categories range from "critically low" to "high," according to 16 predefined items. Another author double-checked this evaluation.
The analyses are proposed by estimates of accuracy (AUC), in agreement with the NICE guidelines. While no definitive threshold exists, values of AUC/C-index of 0.50 were considered to indicate accuracy or precision no better than chance. Values between 0.50 and 0.60 reflected a very poor accuracy, between 0.60 and 0.70 a poor accuracy, between 0.70 and 0.80 a good accuracy, between 0.80 and 0.90 a very good accuracy, and more than 0.90 an excellent accuracy.
The researchers used the GRADE (Grading of Recommendations, Assessment, Development, and Evaluations) assessment, adapted for prognostic and diagnostic studies. The GRADE gave four different degrees of certainty of evidence, from very low to high. Very low means there is very little confidence in the estimate, meaning the true diagnosis is likely to be substantially different from the estimate. High means we are very confident that the true diagnosis lies close to that of the estimate.
What Did the Researchers Find?
The researchers initially screened 726 articles at title/abstract level. Of these, 94 full-texts were examined, and 42 systematic reviews were finally included after eligibility assessment. The main reasons for exclusion were the lack of information about accuracy (17 studies) and the wrong study design (15 studies).
All the systematic reviews were published after the year 2017. Overall, the 42 systematic reviews included an approximate total of 213,600 individuals and a total of 1,010 studies. This means a mean of 24 studies and 5,209 participants for each systematic review.
Altogether, 211 meta-analytic outcomes were evaluated. Among the markers investigated, 121 comparisons over 211 used RNA and, in particular, microRNA (miRNA). This was followed by DNA markers (53 out of 211) and genetic biomarkers (40 out of 211).
Furthermore, eight comparisons used free tumoral cells, two used exosomes, one used nucleosomes, and one used salivary DNA biomarkers.
The large majority of the meta-analytic comparisons included cancer, at any stage, as the outcome (202 out of 211). Only nine comparisons were about prenatal diagnosis and one was about multiple sclerosis.
How Accurate Are Liquid Biopsy Tests?
When considering the AUC as an estimation of the accuracy of the liquid biopsy markers, the researchers found that 75 out of 211 (45.4%) had an excellent accuracy (AUC above 0.90).
One comparison had an AUC equal to 1, indicating a perfect correspondence between liquid biopsy and the gold standard. This was cell-free human papillomavirus DNA (cfHPV-DNA) for HPV-positive oropharyngeal squamous cell carcinoma (a type of throat cancer linked to HPV infection).
Moreover, 81 out of 211 outcomes showed an AUC between 0.80 and 0.90, thus being classified with a very good accuracy. Additionally, 39 outcomes had an AUC between 0.70 and 0.80, and 16 had an AUC between 0.60 and 0.70.
Some specific examples of high accuracy included:
- cfHPV-DNA for HPV-associated anal cancer: AUC of 0.9955 (95% CI: 0.9886 to 1), with 95% sensitivity and 100% specificity
- ALK ctDNA for lung cancer: AUC of 0.994 (95% CI: 0.953 to 1)
The Quality of Evidence: A Critical Concern
Despite the promising accuracy data, when the level of evidence using the GRADE was evaluated, all the outcomes were graded as very low. This means there is very little confidence in these estimates, and the true diagnostic accuracy is likely to be substantially different from what was reported.
Only 95 out of 211 outcomes (44.8%) reported the 95% confidence interval for AUCs. This increases the risk of imprecision in the results.
Moreover, the risk of bias was graded as very serious in almost all the outcomes included (197 out of 211, or 92.9%). The heterogeneity (inconsistency) between different studies was never reported, making inconsistency very serious in all the comparisons.
Lastly, the presence of publication bias was present in almost half of outcomes (102 out of 211, or 48.1%). Publication bias occurs when studies with positive results are more likely to be published than studies with negative results.
Risk of Bias in the Included Studies
The researchers evaluated the possible risk of bias in the 42 systematic reviews included in the study. Of these, 30 were ascertained as having a critically low quality, 11 had a low quality, and only one had a high quality.
The main reasons for this possible bias were identified as:
- Lack of sufficient information about statistical methods used for meta-analysis
- Poor consideration of the risk of bias in affecting results in original systematic reviews
What This Means for Patients
This study represents a quantitative summary of evidence related to liquid biopsy, highlighting its potential and limitations. The researchers evaluated the accuracy of several liquid biopsy markers across 42 systematic reviews published after 2017, including a total of 213,600 individuals from 1,010 studies.
A significant fraction of the outcomes considered had excellent or very good accuracy. Based on this accuracy, this study may corroborate recent evidence of the need for implementing liquid biopsy into clinical practice.
However, as expected, most systematic reviews and meta-analyses in the literature regard oncology and related fields (95.7%).
Although liquid biopsy has been mainly explored in the context of precision oncology, it should be acknowledged that it may bring crucial advantages into different clinical levels. These advantages include:
- Avoiding invasive procedures for the molecular characterization of a disease, even for patients with multiple comorbidities (other medical conditions) and/or unfit for surgery
- Allowing molecular investigation for the molecular characterization of tumors, even for stage IV or metastatic patients, where surgical resections cannot be considered and tissue biopsies are not always practicable
- Giving the possibility of overcoming intratumor heterogeneity, representing a possibility of showing a comprehensive molecular landscape of a given neoplasm, independent from tissue-related biases
These advantages are unique for liquid biopsy and related methodologies. They call for further research along those lines for implementing this promising tool into clinical practice.
The use of integrated biomarkers allows for a real-time, minimally invasive assessment of the tumor's genetic and epigenetic alterations. For instance, ctDNA can reveal mutations in key oncogenes, such as EGFR, providing valuable information for therapy selection, particularly for targeted treatments.
The inclusion of microRNAs and exosomal content further enhances the diagnostic accuracy. These components can reflect various biological processes involved in cancer progression and metastasis.
In lung cancer management, the integration of these biomarkers supports longitudinal monitoring. By analyzing changes in ctDNA levels or the presence of specific mutations over time, clinicians can adjust treatment strategies more effectively, optimizing outcomes for patients.
This approach enables personalized medicine, where treatment is tailored based on the evolving molecular profile of the tumor. This improves both the prognosis and quality of life for cancer patients.
The role of liquid biopsy in the central nervous system (CNS) is critical in improving non-invasive diagnostic and monitoring approaches. This role is specifically for detecting cell-free DNA (cfDNA) in patients with meningioma and glioblastoma. Liquid biopsy may provide a minimally invasive method to detect tumor-derived DNA in bodily fluids, offering a real-time reflection of tumor dynamics.
The Future Role of Artificial Intelligence
More recent artificial intelligence (AI) systems could significantly help to implement the use of liquid biopsy in daily clinical practice. AI can help manage big data derived from multi-omics analyses on liquid biopsy.
AI can improve the sensitivity and specificity of liquid biopsies for early cancer detection. By integrating data from multiple biomarkers and employing sophisticated algorithms, AI can distinguish between benign and malignant conditions more effectively than traditional methods. This could potentially lead to earlier diagnosis and better prognostic outcomes.
Finally, the use of AI in the liquid biopsy field may track changes in biomarker levels over time. This provides valuable insights into disease progression and patient response to treatments, allowing for personalized treatment adjustments.
Limitations of This Research
This umbrella review on existing meta-analyses highlighted some critical issues for providing quantitative estimations on the different roles of liquid biopsy.
All outcomes were graded as very low on the GRADE criteria. This means there is very little confidence in the accuracy estimates, and the true diagnostic accuracy is likely to be substantially different from what was reported.
The heterogeneity between different studies was never reported. This makes inconsistency very serious in all the comparisons.
Only 95 out of 211 outcomes (44.8%) reported the 95% confidence interval for AUCs. This increases the risk of imprecision.
The risk of bias was graded as very serious in almost all outcomes included (197 out of 211, or 92.9%).
Publication bias was present in almost half of outcomes (102 out of 211, or 48.1%).
Of the 42 systematic reviews included, 30 were ascertained as having a critically low quality, 11 had a low quality, and only one had a high quality.
The main reasons for this possible bias were the lack of sufficient information about statistical methods used for meta-analysis. Another reason was the poor consideration of the risk of bias in affecting results in original systematic reviews.
Recommendations and Next Steps
The literature about liquid biopsy is rapidly increasing, and some promising data about precision oncology are now available. However, this umbrella review on existing meta-analyses highlighted some critical issues for providing quantitative estimations on the different roles of liquid biopsy.
Based on the accuracy findings, this study may corroborate recent evidence of the need for implementing liquid biopsy into clinical practice. However, the very low quality of evidence means that stronger research is needed before liquid biopsy can be widely adopted.
For patients, this means that liquid biopsy is a promising technology that may eventually offer a less invasive alternative to traditional tissue biopsies. However, patients should be aware that the current evidence base has significant limitations.
Patients interested in liquid biopsy should discuss this option with their healthcare providers. They should ask about the specific test being considered, its accuracy for their particular condition, and whether it is appropriate for their situation.
The researchers call for further research along these lines for implementing this promising tool into clinical practice. Future studies should address the limitations identified in this review, including the need for standardized approaches, better reporting of statistical methods, and more rigorous assessment of risk of bias.
Frequently Asked Questions
What is a liquid biopsy and how is it different from a tissue biopsy?
A liquid biopsy is a laboratory test performed on a fluid sample, such as blood or urine. It analyzes molecular data from circulating cells or directly from nucleic acids. Unlike a traditional tissue biopsy, which requires an invasive procedure to remove a piece of tissue, a liquid biopsy uses a simple venous blood sample. It has been explored mainly in cancer medicine.
How accurate are liquid biopsy tests according to this research?
In an umbrella review of 42 systematic reviews covering about 213,600 people, 75 of 211 outcomes (45.4%) showed excellent accuracy, meaning an AUC above 0.90. Another 81 outcomes had very good accuracy (AUC 0.80–0.90). However, every outcome was graded as very low quality evidence, so the true accuracy may differ substantially.
Is liquid biopsy ready for routine use in patients?
Not yet. Although many outcomes showed excellent or very good accuracy, all were graded as very low quality evidence. The review concluded that stronger research is needed before liquid biopsy can be widely adopted in clinical practice. Patients interested in liquid biopsy should discuss it with their healthcare providers and ask about its accuracy for their specific condition.
What are the main limitations of the evidence on liquid biopsy?
The review found that all outcomes had very low quality evidence. Heterogeneity between studies was never reported. Only 95 of 211 outcomes (44.8%) reported confidence intervals. Risk of bias was very serious in 197 of 211 outcomes (92.9%), and publication bias was present in 102 outcomes (48.1%). Of the 42 reviews, 30 were critically low quality.
What advantages might liquid biopsy offer over tissue biopsy?
Liquid biopsy may avoid invasive procedures for molecular characterization, which could help patients with multiple other medical conditions or who are unfit for surgery. It may allow molecular investigation even for stage IV or metastatic patients where surgery is not possible. It may also overcome intratumor heterogeneity by giving a comprehensive molecular landscape of a tumor.
If I'm considering a liquid biopsy for cancer detection or monitoring, when should I get a second opinion?
Liquid biopsy accuracy estimates come from evidence graded very low, with serious risk of bias in 92.9% of outcomes and publication bias in 48.1%, so the true accuracy may differ substantially from reported figures. Before relying on a specific liquid biopsy test, ask your healthcare provider about its accuracy for your particular condition and whether it is appropriate for your situation. A second opinion can help confirm whether the test fits your diagnosis and treatment decisions. Diagnostic Detectives Network provides independent expert second opinions.
Source Information
Original Article Title: Potentialities and critical issues of liquid biopsy in clinical practice: An umbrella review.
Authors: Nicola Veronese, Claudio Luchini, Stefano Ciriminna, Katia Spinelli, Santo Fruscione, Paola Mattiolo, Miriam Belluzzo, Veronica Messina, Lee Smith, Mario Barbagallo, Walter Mazzucco
Author Affiliations:
- Department of Health Promotion, Mother and Child Care, Internal Medicine and Medical Specialties, University of Palermo, Palermo, Italy
- Department of Diagnostics and Public Health, Section of Pathology, University and Hospital Trust of Verona, Verona, Italy
- Centre for Health Performance and Wellbeing, Anglia Ruskin University, Cambridge, UK
- Division of Biostatistics & Epidemiology Research, Cincinnati Children's Hospital Medical Center, University of Cincinnati College of Medicine, United States
Journal: Translational Oncology, Volume 52 (2025), Article 102172
Publication Details: Received 20 August 2024; Received in revised form 9 October 2024; Accepted 29 October 2024; Published 2025
DOI: https://doi.org/10.1016/j.tranon.2024.102172
License: This is an open access article published under the CC BY-NC-ND license
This patient-friendly article is based on peer-reviewed research.
