Liquid Biopsies and ctDNA Analysis

Slide 1 · Liquid Biopsies and

ctDNA Analysis

Molecular principles, laboratory workflow, and clinical interpretation

Dr Abu Suraih Sakhri

MBBS, MD (Pathology)

www.abusuraihsakhri.com

Slide 2 · Learning objectives

Distinguish ctDNA from other circulating tumor-derived analytes.

Explain why rare tumor alleles can be detected amid normal cfDNA.

Outline preanalytic, analytic, and interpretive vulnerabilities.

Appraise clinical indications without conflating validity and utility.

Slide 3 · The central diagnostic idea

Tumor

Genetic or epigenetic abnormalities

Blood

Tumor-derived material enters circulation

Assay

Isolate and interrogate analytes

Decision

Interpret in clinical context

A blood draw can provide molecular information that otherwise may require tumor tissue.

Liquid biopsy complements—not universally replaces—histopathologic sampling.

Slide 4 · What constitutes a liquid biopsy?

Definition

Relatively noninvasive sampling of tumor-associated material from a body fluid.

Plasma is the principal specimen discussed here for ctDNA.

Not equivalent to tissue

Does not directly show tissue architecture, grade, or invasion.

Analyte yield varies with disease state and assay design.

Slide 5 · Tumor-derived analytes in circulation

Analyte

Biological form

Principal analytical information

ctDNA

Fragmented tumor-derived cfDNA

Sequence variants; methylation signatures

Circulating tumor cells

Intact malignant cells

Cell-associated nucleic acid / phenotype

Extracellular vesicles

Membrane-enclosed material

Molecular cargo under investigation

Trace proteins

Soluble tumor-associated products

Complementary biomarker signals

Slide 6 · Cell-free DNA is not synonymous with ctDNA

CfDNA

DNA fragments from multiple normal and abnormal cellular sources.

Most circulating DNA in plasma may be nontumor-derived.

CtDNA

The tumor-derived fraction of the cfDNA pool.

Identification depends on tumor-associated molecular features.

Slide 7 · How tumor DNA enters blood

Living tumor

Genomic alterations in neoplastic cells

Cell death

DNA released by dying cells

Circulation

Fragments mix with background cfDNA

Plasma assay

Detect cancer-associated patterns

Tumor-derived DNA can also be carried in circulating tumor cells.

The assay sees a molecular mixture, not a pure tumor genome.

Slide 8 · Circulating cells and free DNA are distinct

Circulating tumor cells

Whole malignant cells detached from tumor sites.

Potentially enable cellular and phenotypic studies.

Circulating tumor DNA

Free tumor-derived nucleic acid fragments.

Suitable for highly sensitive sequence and methylation analyses.

Slide 9 · Biological shedding is heterogeneous

Tumor presence does not guarantee enough ctDNA in a particular blood sample.

Burden, anatomical site, biology, and treatment state affect shedding.

Low tumor fraction produces biologically driven false-negative risk.

A negative plasma finding is not proof that a lesion is absent.

Slide 10 · Why the background matters

Normal tissues

Background cfDNA fragments

Malignant tissue

Rare variant-bearing fragments

Plasma sample

Mixed molecular population

Sensitive assay

Discriminate tumor from noise

Analytic sensitivity must be considered relative to the fraction of tumor DNA in the input.

Sample processing can further increase the normal-DNA background.

Slide 11 · Spatial tumor heterogeneity

Single tissue biopsy

Samples one lesion or region.

May not reflect every metastatic clone.

Plasma sampling

May contain DNA shed from more than one tumor site.

Cannot ensure every lesion is represented.

Slide 12 · Temporal tumor heterogeneity

Pretreatment

Dominant baseline clones

Therapy

Differential selective pressures

Residual disease

Surviving populations

Progression

Potential resistance-associated variants

Serial liquid biopsies can sample molecular change over time.

Observed changes require consideration of tumor fraction and assay variability.

Slide 13 · Why early-stage detection is difficult

Small or low-shedding tumors may contribute very little ctDNA.

Tumor-associated signals can be below the assay’s detection capability.

Population screening amplifies the consequences of false positives.

Clinical benefit must be demonstrated separately from detection alone.

Slide 14 · Genomic signals of tumor origin

Tumor signal

Illustrative meaning

Interpretive caution

Recurrent driver mutation

Known cancer-associated alteration

Not inherently specific to one tumor or tissue

Patient-specific passenger mutations

Individualized tumor molecular signature

Requires convincing tumor linkage

Epigenetic signature

Cancer-associated methylation pattern

Cell-type origin depends on validated classifier

Slide 15 · Driver versus passenger variants

Driver alterations

Selected because they confer a neoplastic advantage.

May recur across independent tumors.

Passenger alterations

Acquired during clonal evolution without necessarily driving malignancy.

Collectively can identify one patient’s tumor lineage.

Slide 16 · Recurrent point mutations

Tumor clone

Carries a recurrent alteration

DNA shedding

Mutated fragments enter plasma

Targeted testing

Assay interrogates specified locus

Interpretation

Relate finding to tumor and phenotype

A recurrent alteration can be analytically recognizable in a background of wild-type sequences.

Tumor specificity is not automatically assured by the variant name alone.

Slide 17 · Patient-specific mutation signatures

Tumor tissue

Identify somatic mutation profile

Assay design

Select informative patient variants

Plasma sampling

Search for these molecular traces

Longitudinal test

Assess persistence or reappearance

A bespoke signature can be more specific to the known tumor lineage.

A tumor-informed assay depends on suitable tissue and reliable baseline profiling.

Slide 18 · Epigenetic alterations as biomarkers

DNA methylation can change without altering nucleotide sequence.

Cancer-associated methylation patterns may serve as molecular signatures.

Cell-type dependence creates a possible tissue-of-origin signal.

A validated classifier is needed before inferring primary tumor site.

Slide 19 · Methylation and tissue-of-origin inference

Cell lineage

Characteristic methylation landscape

Malignancy

Altered epigenetic pattern

CtDNA

Fragments preserve some patterns

Classifier

Estimate biological origin

This is an inference from molecular patterns, not direct morphologic identification.

Sensitivity and specificity depend on assay development and validation.

Slide 20 · Complementary molecular analytes

Proteins and vesicles

Tumor-associated proteins and extracellular vesicles can enter blood.

May capture biology not represented by DNA alone.

CtDNA strengths

Directly interrogates mutation or methylation signatures.

Can support patient-specific molecular tracking.

Slide 21 · Why NGS changed the field

Massively parallel sequencing reads many DNA molecules independently.

Mixtures of normal and abnormal molecules can be analyzed.

High depth increases the opportunity to detect rare variant alleles.

Input and error characteristics still determine real sensitivity.

Slide 22 · Core NGS mechanism

Molecules

Fragmented DNA templates

Separation

Individual spatial templates

Amplification

Generate detectable local signal

Sequencing

Read many templates in parallel

The output comprises reads that must be assigned to genomic locations.

Bioinformatics distinguishes candidate variants from noise.

Slide 23 · Targeted panels concentrate sequencing effort

Targeted gene panel

Enriches selected genomic intervals.

Provides deep coverage of chosen targets.

Broad genomic approach

Covers more regions but distributes resources.

May be less efficient for extremely rare alleles without specialized design.

Slide 24 · A sensitive assay begins with a clear question

Clinical question

Appropriate target concept

Key limitation

Identify actionable variant

Relevant mutation panel

Alteration must be covered and detectable

Track a known tumor

Patient-specific variants

Requires prior tumor profile

Infer tumor origin

Epigenetic patterns

Needs validated origin classifier

Explore heterogeneous disease

Broader genomic profiling

Lower shed and complex alterations remain difficult

Slide 25 · Variant allele frequency (VAF)

VAF expresses variant-bearing reads relative to total reads covering a locus.

Low VAF may reflect a small tumor-derived contribution to cfDNA.

VAF does not equal the percentage of malignant cells in the patient.

Copy number, zygosity, clonality, and biology complicate translation.

Slide 26 · Depth is not the same as sensitivity

Sequencing depth

Number of reads covering a locus.

More reads can improve sampling of rare alleles.

Actual detection capability

Also depends on molecules, errors, assay design, and background.

High nominal depth cannot recover absent tumor fragments.

Slide 27 · Input molecules create a hard limit

Blood sample

Finite plasma volume

Extraction

Finite recoverable cfDNA

Library

Subset of molecules represented

Analysis

No call possible if mutant template absent

Extra PCR copies are not new independent biological molecules.

Sampling variability becomes important near the detection limit.

Slide 28 · Sequencing artifacts can mimic rare variants

Base-calling, DNA damage, and amplification introduce technical noise.

Rare true events must be separated from systematic artifacts.

Error suppression is especially important for low-frequency ctDNA.

Every platform needs a validated detection threshold and quality controls.

Slide 29 · Molecular barcodes and error suppression

Original DNA

Assign molecular identifier

Amplification

Copies share the identifier

Sequence reads

Group reads by molecule

Consensus

Reduce amplification / sequencing errors

This is a conceptual explanation of a common error-suppression approach.

Barcoding does not correct biological confounding such as clonal hematopoiesis.

Slide 30 · Digital PCR and NGS answer different questions

Feature

Targeted digital PCR

Targeted NGS

Core strategy

Interrogate defined alterations

Sequence multiple selected regions

Best-fit question

Is this known variant present?

What variants are represented in the panel?

Tradeoff

Limited multiplexed discovery

More complex informatics and error modeling

Interpretation

Known target coverage

Report only within validated assay scope

Slide 31 · Tumor-informed versus tumor-naïve designs

Tumor-informed

Initial tumor sequencing identifies personalized variants.

Useful concept for highly specific longitudinal tracking.

Tumor-naïve

Plasma assay does not require bespoke tumor sequencing.

May interrogate recurrent variants or epigenetic patterns.

Slide 32 · Preanalytical quality starts before sequencing

Phlebotomy

Correct tube and specimen identification

Transport

Control time and handling

Separation

Avoid leukocyte contamination

Extraction

Assess cfDNA quantity and quality

Leukocyte lysis adds non-tumor DNA and can dilute the ctDNA fraction.

Preanalytical parameters must follow a validated laboratory protocol.

Slide 33 · Plasma is generally preferred to serum

Plasma

Obtained from anticoagulated blood.

Reduces contamination from DNA released during clotting.

Serum

Clotting can release leukocyte-derived DNA.

Additional background reduces relative tumor fraction.

Slide 34 · Tube choice and processing delay

Routine EDTA tubes require prompt plasma separation.

Stabilizing tubes can support delayed processing within validated conditions.

Delay and agitation can elevate normal cfDNA through cellular breakdown.

Do not substitute generic transport assumptions for assay instructions.

Slide 35 · Plasma separation is a quality-critical step

Whole blood

Anticoagulated sample

First spin

Separate plasma from cells

Second spin

Reduce residual cellular material

Aliquot

Protect plasma for extraction

Avoid carrying over the buffy coat during pipetting.

Specific centrifugal conditions must be laboratory-validated.

Slide 36 · Preanalytic failure pathways

Failure

Likely effect

Preventive control

Leukocyte lysis

High normal DNA background

Timely separation / stabilizing tube

Buffy-coat carryover

Cellular DNA contamination

Validated two-step separation

Insufficient plasma

Low unique molecule count

Volume sufficiency checks

Mislabeled time point

Incorrect kinetic interpretation

Accurate collection documentation

Slide 37 · Sequence analysis is a pipeline

Library reads

Quality and artifact review

Alignment

Map reads to reference

Variant calling

Identify candidate changes

Annotation

Assign clinical significance

Each step has separate quality thresholds and opportunities for errors.

Clinical reporting follows laboratory review of annotated calls.

Slide 38 · Alignment establishes genomic context

Short reads must be mapped to a reference genome.

Mismapping or repetitive regions can create spurious variant calls.

Coverage uniformity matters as much as global read count.

Alignment quality is a prerequisite for reliable interpretation.

Slide 39 · Variant calling requires calibrated thresholds

Candidate alleles are compared with reference sequence.

Read support, base quality, strand pattern, and artifact profile matter.

Low-frequency calls must satisfy assay-specific validation rules.

A reportable variant is not automatically a clinically actionable variant.

Slide 40 · Variant annotation has multiple layers

Layer

Question answered

Example of caution

Identity

What change and gene?

Correct transcript and variant nomenclature

Pathobiology

Tumor driver or incidental finding?

Tumor source not proven by plasma alone

Clinical evidence

Diagnostic / prognostic / predictive?

Cancer type and therapy context differ

Actionability

Does the finding alter management?

Requires current indication / evidence review

Slide 41 · Clonal hematopoiesis is a major confounder

Nonmalignant hematopoietic clones can carry somatic mutations.

DNA shed from these clones contributes to plasma cfDNA.

A biologically real plasma mutation may be unrelated to the solid tumor.

Matched leukocyte testing can distinguish blood-cell and tumor origin.

Slide 42 · Why clonal hematopoiesis mimics cancer

Hematopoietic clone

Somatic mutation in blood-cell lineage

Cell turnover

Mutated fragments enter plasma

Plasma sequencing

Variant detected alongside cfDNA

Source check

Leukocyte vs tumor attribution

Analytic detection can be correct while tumor attribution is wrong.

This is a clinical false positive for tumor genotyping, not necessarily a sequencing error.

Slide 43 · Clonal hematopoiesis versus true ctDNA

Characteristic

Clonal hematopoiesis

Tumor-derived ctDNA

Source

Nonmalignant hematopoietic clone

Malignant neoplastic clone

Plasma variant

Often technically real

Technically real when detectable

Matched leukocyte DNA

Can show concordant variant

Not expected for a strictly tumor-specific somatic variant

Clinical use

Do not misattribute to solid tumor

Interpret against tumor type and context

Slide 44 · Germline variants are a separate issue

Somatic tumor alteration

Acquired by malignant clone.

May be useful for tumor genotyping or monitoring.

Inherited germline variation

Constitutional variant across normal tissues.

Requires appropriate confirmatory germline testing when suspected.

Slide 45 · Why a negative plasma test can mislead

Cancer present

Actual tumor burden

Low shedding

Very few mutated molecules

Blood specimen

Variant missed by sampling

Result

Non-detected does not exclude mutation

The relevant alteration may be present in tissue despite a negative plasma result.

Consider reflex tissue testing when clinically indicated.

Slide 46 · Copy-number and fusion detection is harder

Performance varies across molecular alteration classes.

Plasma assays may be less sensitive for fusions and copy-number changes.

Tissue-based testing may still be preferred for many cancer settings.

Interpret “no alteration detected” only within the assay’s technical scope.

Slide 47 · Sensitivity, specificity, and clinical value

Term

Central question

Why it matters

Analytical validity

Does assay detect what it claims?

LOD, accuracy, precision, interference

Clinical validity

Is the result associated with disease or outcome?

Prognostic or diagnostic relationship

Clinical utility

Does result-guided care improve decisions/outcomes?

Evidence beyond association

Slide 48 · A result requires a two-axis interpretation

Tumor biology axis

Expected shedding / stage / treatment.

Chance that a real tumor alteration reached plasma.

Assay axis

Coverage / detection limit / quality.

Chance that a present tumor fragment would be called.

Slide 49 · Interpreting plasma-positive findings

Confirm the variant is within validated assay coverage and QC.

Distinguish tumor-derived from hematopoietic or germline origins.

Evaluate alteration-specific clinical evidence for this cancer.

Document whether the finding is diagnostic, prognostic, or predictive.

Slide 50 · Interpreting plasma-negative findings

Do not equate non-detection with absence of malignancy.

Review tumor type, disease burden, collection timing, and sample adequacy.

Determine whether the target and variant class are reliably assessed.

Pursue tissue testing when a negative plasma result is not informative.

Slide 51 · Plasma and tissue are complementary

Dimension

Tissue analysis

Plasma ctDNA analysis

Architecture

Histology, grade, invasion

No tissue architecture

Sampling

Local tumor region

Circulating, variable lesion representation

Repeat testing

May require invasive procedure

Relatively noninvasive serial sampling

Key failure

Limited tissue or spatial sampling

Low ctDNA or nontumor DNA confounding

Slide 52 · Advanced cancer genotyping: Supported role

Validated, sensitive plasma assays can identify actionable variants.

May be especially useful when tissue is inaccessible or speed matters.

Clinical interpretation must account for assay coverage and sensitivity.

ESMO 2022: Tissue testing may be needed after a noninformative plasma test.

Slide 53 · Why speed and accessibility can matter

Advanced cancer

Urgent molecular decision

Tissue barrier

Biopsy risk, delay, or inadequacy

Plasma assay

Obtain genotype more readily

Review

Confirm clinically actionable result

This workflow is scenario dependent, not universal “liquid-first” policy.

Histologic classification and other essential pathology may still require tissue.

Slide 54 · Molecular resistance at progression

Baseline tumor

Initial targetable alteration

Targeted therapy

Sensitive clones decline

Selective escape

Resistant clone may emerge

Repeat profiling

Investigate changed genotype

A newly detected variant may reflect resistance or subclonal selection.

Clinical action requires validated variant–drug evidence.

Slide 55 · Monitoring therapeutic response

Potential benefit

Serial ctDNA changes may reflect tumor molecular burden.

Can sometimes precede conventional detection of progression.

Why caution persists

Levels fluctuate with shedding, timing, and assay conditions.

Early ctDNA-guided regimen changes require clinical utility evidence.

Slide 56 · Molecular residual disease (MRD)

MRD refers to residual malignant material after curative-intent therapy.

CtDNA detection may indicate persistent occult disease.

A positive association with future relapse is clinical validity.

Changing therapy on the result needs separate prospective utility evidence.

Slide 57 · MRD and radiologic assessment differ

Radiologic assessment

Identifies disease above imaging detectability.

Cannot reliably exclude microscopic persistence.

Molecular assessment

Searches for tumor-specific material in plasma.

Cannot ensure a negative sample is biologically disease-free.

Slide 58 · A postoperative ctDNA result is contextual

Surgery

Macroscopic tumor removed

Time point

Define postoperative sampling window

Plasma test

Positive or non-detected result

Decision

Use validated evidence / trial context

A result must be tied to a defined collection time and intended use.

FDA 2024 addresses ctDNA biomarkers in early-stage trials; It does not establish universal care algorithms.

Slide 59 · Early detection and population screening

Screening requires careful assessment of low-prevalence populations.

False-positive workups and missed early disease can cause harm.

Tissue-of-origin classifiers need independent clinical validation.

ESMO 2022 regarded asymptomatic screening as investigational.

Slide 60 · Current evidence must be use-case specific

Use case

Evidence interpretation

Important caveat

Advanced cancer genotyping

Established role for validated assays

Noninformative result may require tissue

MRD / molecular relapse

Strong prognostic associations in many settings

Treatment benefit is indication-specific, not automatic

Response-adaptive therapy

Promising molecular dynamics

Prospective clinical utility needed

Asymptomatic screening

Investigational in ESMO 2022 framework

False-positive and false-negative consequences

Slide 61 · What an adequate report should communicate

Specimen, sampling time, platform, and genes or targets assayed.

Reported variants with nomenclature and interpretive evidence.

Limits of detection, relevant quality constraints, and non-detection caveats.

Whether results are suitable for specific clinical decisions.

Slide 62 · Case: Inadequate tissue in advanced cancer

CLINICAL VIGNETTE

A patient with advanced carcinoma needs molecular profiling, but the available tissue is exhausted.

What can a plasma assay contribute, and what if it is negative?

A validated ctDNA assay may identify actionable tumor-associated alterations.

A noninformative plasma result should not terminate tissue-based testing when feasible.

Tissue remains necessary when morphology or molecular coverage is unresolved.

Slide 63 · Case: Negative plasma in early malignancy

CLINICAL VIGNETTE

A localized tumor is confirmed histologically. The liquid biopsy reports no detectable tumor variant.

Which interpretation is scientifically defensible?

Low tumor DNA shedding or finite sampling may explain non-detection.

A negative plasma result does not overrule tissue-based diagnosis.

Consider the analytical targets, assay limit, and collection context.

Slide 64 · Case: A plasma mutation from blood cells

CLINICAL VIGNETTE

A plasma assay detects an oncogenic variant; The same variant is found in matched peripheral blood leukocytes.

What source should enter the differential?

Clonal hematopoiesis may explain the plasma finding.

A confirmed sequence change need not be derived from the solid tumor.

Avoid assigning tumor therapy based only on an unverified plasma source.

Slide 65 · Case: Detectable ctDNA after surgery

CLINICAL VIGNETTE

After curative-intent resection, a tumor-informed plasma assay detects a patient-specific mutation signature.

What can and cannot be concluded?

This can be molecular evidence of residual tumor-derived material.

Risk association is not identical to proof that changing treatment improves outcomes.

Interpret using disease-specific trials, timing, and current clinical protocols.

Slide 66 · Interpretation exercise: Four statements

Statement

Assessment

Reason

“ctDNA = all plasma cfDNA”

Incorrect

cfDNA includes substantial nontumor material

“Plasma-negative excludes mutation”

Incorrect

Low shedding and finite sampling

“CH can mimic tumor variants”

Correct

Hematopoietic clones shed mutant DNA

“MRD positivity proves benefit of adjuvant change”

Incorrect

Utility must be demonstrated

Slide 67 · Analytic safeguards across the test pathway

Before assay

Specimen suitability

During assay

Coverage / error controls

After assay

Variant source attribution

Clinical use

Evidence + alternatives

No single quality-control step solves all sources of uncertainty.

Every molecular conclusion depends on a linked preanalytic-to-clinical chain.

Slide 68 · Integrated take-home principles

CtDNA is a tumor-derived subset of circulating cell-free DNA.

Point mutations, patient-specific signatures, and methylation are complementary readouts.

NGS allows rare-allele detection, but biology and assay limits remain decisive.

A positive or negative result must be interpreted with tissue and clinical context.

Slide 69 · Clinical interpretation: Final synthesis

Separate analytically correct findings from true tumor attribution.

Distinguish predictive targets, prognostic MRD associations, and screening hypotheses.

Do not overrule tissue pathology with an uninformative plasma result.

Demand evidence of patient benefit before changing therapy from ctDNA alone.

Slide 70 · Primary and supporting references

Primary textbook

Robbins, Cotran & Kumar. Pathologic Basis of Disease. 11th ed. 2026; Ch. 5, pp. 164–166.

Clinical consensus

Pascual J, et al. Ann Oncol. 2022;33:750–768. Doi:10.1016/j.annonc.2022.05.520

Preanalytical guidance

Clinical Practice Guideline for Blood-based ctDNA Assays. 2024. PMCID: PMC10813828.

Slide 71 · Additional literature and regulatory context

Biological confounding

Razavi P, et al. Nat Med. 2019;25:1928–1937. doi:10.1038/s41591-019-0652-7

Regulatory methodology

FDA. Use of ctDNA for Early-Stage Solid Tumor Drug Development. Final Guidance, Nov 2024.

Implementation update

ESMO commentary on ctDNA standardization and implementation, 19 Jun 2026.