T-cell lymphoma is a neoplasm defined by high invasiveness and striking molecular heterogeneity. The most common subtypes include extranodal NK/T-cell lymphoma (ENKTL), nodal T-follicular helper cell lymphoma angioimmunoblastic-type (nTFHL-AI), peripheral T-cell lymphoma not otherwise specified (PTCL, NOS), and ALK-positive or ALK-negative anaplastic large cell lymphoma (ALCL). Together these account for roughly 80% of all T-cell lymphomas. Survival outcomes remain poor across most subtypes: the five-year survival rates for PTCL, NOS and nTFHL-AI fall between 32% and 43%, while ALK-positive ALCL fares better at approximately 70%.
The tissue biopsy problem: Invasive tissue biopsy is the current gold standard for histological diagnosis and genetic subtyping, and it remains essential for treatment planning. However, surgical access is not always feasible. Patients with severe comorbidities, unresectable tumors, or anatomically inaccessible disease may be unable to undergo biopsy. Even when biopsy is possible, it captures only a single site at a single time point, missing the spatiotemporal heterogeneity that evolves as the disease progresses or responds to treatment. Tissue biopsy is also insufficiently sensitive for detecting minimal residual disease (MRD), a condition in which small numbers of cancer cells persist after therapy but fall below the threshold of imaging detection.
Liquid biopsy as an alternative: Liquid biopsy extracts biological information from body fluids, most commonly peripheral blood, but also cerebrospinal fluid, pleural effusion, urine, and saliva. The technique can analyze a range of tumor-derived components, including circulating cell-free DNA (cfDNA), circulating tumor DNA (ctDNA), circulating tumor cells (CTCs), Epstein-Barr virus (EBV) DNA, antibodies, and cytokines. Its key advantages over tissue biopsy include non-invasiveness, the ability to perform repeated serial sampling across the disease course, capture of tumor heterogeneity across multiple sites simultaneously, and real-time dynamic monitoring at the molecular level.
This 2024 review, published in Molecular Cancer by Huang and colleagues from the Sichuan Cancer Hospital and affiliated institutions, provides a comprehensive overview of liquid biopsy biomarkers in T-cell lymphoma. The paper covers the biology of each biomarker class, detection methodologies, clinical applications across diagnosis, prognosis, treatment monitoring, and MRD assessment, ongoing clinical trials, and the specific challenges that must be addressed before liquid biopsy enters routine clinical use.
ctDNA is a tumor-derived subset of cfDNA, the short fragments of double-stranded DNA shed into plasma by all cells. In healthy individuals, plasma cfDNA concentrations average approximately 30 nanograms per milliliter, while cancer patients can have levels up to 1,000 ng/mL with an average of 180 ng/mL. cfDNA fragments are generally 160-200 base pairs long; ctDNA fragments are slightly shorter at 90-150 base pairs and have a plasma half-life of roughly two hours. ctDNA carries tumor-specific genetic and epigenetic signatures, including point mutations, copy number variations, structural variations, microsatellite alterations, and methylation changes, enabling non-invasive interrogation of tumor biology.
Diagnostic utility: For nTFHL-AI, epigenetic changes in TET2, DNMT3A, and IDH2, as well as the RHOA G17V hotspot mutation, have well-established diagnostic value detectable in ctDNA. A landmark finding demonstrated that the G17V RHOA mutation in ctDNA from nTFHL-AI patients achieved 100% sensitivity and 100% specificity for predicting tumor DNA mutation status, making it an exceptionally precise non-invasive diagnostic marker. Additional novel RHOA mutations, such as c.73A>G (p.Phe25Leu) and c.48A>T (p.Cys16*) in exon 2, have been identified in peripheral T-cell lymphoma, illustrating how ctDNA can reveal tumor heterogeneity beyond what tissue biopsy captures at a single site.
Prognostic and treatment monitoring evidence: In a prospective study of ENKTL patients, targeted NGS performed on paired tumor tissue and serial longitudinal plasma cfDNA demonstrated strong concordance between genotyping results. Critically, patients with low cfDNA concentrations before treatment had significantly better one-year progression-free survival (PFS) compared to those with high concentrations (90.0% vs. 36.4%, P = 0.012). Similarly, patients achieving cfDNA clearance had a complete remission rate of 80.0% compared to 0% in those with persistently detectable ctDNA after treatment (P = 0.004). In a longitudinal study of 45 PTCL patients, post-treatment ctDNA mutation burden was strongly correlated with disease recurrence or progression, and a genomic equivalents (GE) decrease of more than 1.5 log from baseline to end of treatment was significantly associated with improved PFS and OS (P = 0.027 and P = 0.003, respectively).
MRD assessment after stem cell transplant: Herrera et al. evaluated ctDNA-based MRD in lymphoma patients undergoing allogeneic hematopoietic stem cell transplantation (HSCT). Of 19 patients, 16 (84%) had detectable ctDNA a median of 3.7 months before clinical relapse or progression. Two-year PFS for ctDNA-positive patients at three months post-HSCT was 58%, compared to 84% for ctDNA-negative patients (P = 0.033). Multivariate analysis demonstrated that detectable ctDNA was associated with a hazard ratio of 3.9 for progression or death and a hazard ratio of 10.8 for relapse or progression.
Sensitive and specific detection methods are essential for cfDNA and ctDNA analysis given their low abundance and short half-life in plasma. Sample collection typically requires only 6-10 mL of blood, but the 2-hour plasma half-life of ctDNA demands prompt separation from whole blood in EDTA tubes to prevent contamination from genomic DNA released by peripheral blood mononuclear cells. Commercial cfDNA preservative tubes such as Streck Cell-Free DNA BCT, PAXgene Blood cfDNA, and Roche Cell-Free DNA Collection tubes stabilize samples for 3-7 days at room temperature without affecting DNA yield or mutation background levels.
PCR-based methods: PCR techniques are the most widely deployed for cfDNA detection due to their high sensitivity and specificity for single molecular abnormalities. Droplet digital PCR (ddPCR) can detect genomic material at levels as low as 0.01-1.0%, making it valuable for rare mutation identification and copy number variation quantification. Most commercial digital PCR platforms accommodate only 2-6 fluorescence channels per reaction, limiting simultaneous detection to 2-6 targets, which restricts ddPCR to hotspot mutation assessment rather than comprehensive profiling. BEAMing (beads, emulsion, amplification, and magnetics), which combines PCR with flow cytometry, achieves similar sensitivity of 0.01% with good concordance to tissue testing and is considered relatively cost-effective for known mutation surveillance.
NGS-based methods: Next-generation sequencing enables detection of diverse molecular abnormalities from a single sample. Targeted sequencing approaches include tagged-amplicon deep sequencing (TAm-Seq) and cancer personalized profiling by deep sequencing (CAPP-Seq). Untargeted methods include whole-exome sequencing (WES), whole-genome sequencing (WGS), and whole-genome bisulfite sequencing (WGBS-Seq). WGS evaluates the entire tumor genome but carries higher cost and computational demands. WES reduces cost by focusing on exon regions, at some expense of sensitivity compared to WGS. WGBS-Seq is the gold standard for DNA methylation analysis, with high accuracy for cytosine-level measurements, though susceptible to sensitivity loss from DNA degradation.
Open chromatin region analysis: A newer bioinformatics approach, OCRDetector, analyzes cfDNA fragment characteristics to infer gene expression and predict transcription factor binding sites by detecting open chromatin regions (OCRs) from whole-genome cfDNA sequencing data. The method calculates window protection score (WPS) waveforms and cfDNA sequencing coverage to map chromatin accessibility genome-wide. While no studies have yet directly linked OCRs to T-cell lymphoma specifically, this approach represents a potential expansion of cfDNA's diagnostic and monitoring capabilities beyond mutation detection into epigenomic profiling.
Circulating tumor cells (CTCs) are tumor cells shed into the bloodstream from primary or metastatic sites. Their plasma half-life is only 1-2.4 hours, and most patients with metastatic cancer have fewer than 10 CTCs per milliliter of blood, making detection technically demanding. Despite these challenges, CTCs provide direct genomic and immunophenotypic information about the primary tumor, including gene expression profiles and genetic alterations, and their levels have been associated with OS and prognosis across multiple cancer types.
Diagnostic applications: Distinguishing malignant lymphoma from reactive lymphoproliferative lesions can be impossible by conventional histology alone. T-cell receptor (TCR) gene rearrangements serve as a defining molecular marker: while normal reactive T cells show polyclonal TCR rearrangements, malignant T-cell lymphoma cells carry monoclonal TCR rearrangements. Flow cytometric immunophenotyping using TCR-V-beta repertoire analysis (TCR-Vbeta-R) has demonstrated utility for detecting abnormal T-cell populations and their immunophenotypes in liquid samples. Research from Qiyao Pu's team using multiparameter flow cytometry (mpFC) with 15 immunophenotypic markers successfully identified and diagnosed multiple PTCL subtypes, with subtype-specific markers including PD-1 positive cells greater than 38.01% and CD10 positivity greater than 7.46% for nTFHL-AI, CD56 for ENKTL, and CD30 plus HLA-DR for ALCL.
Prognostic stratification in ALK-positive ALCL: Among 180 patients with ALK-positive ALCL undergoing second-line therapy, minimal disseminated disease (MDD) status based on CTCs defined three risk groups. The high-risk group (MDD-positive plus NPM::ALK transcript positive) had a 10-year PFS of 40%, the intermediate group had 75%, and the low-risk group (MDD-negative) had 86% (P less than 0.0001). Five-year survival rates followed a similar gradient: MDD-negative patients achieved 91%, MDD-positive/MRD-negative achieved 92%, and MDD-positive/MRD-positive achieved only 65% (P less than 0.001). The cumulative incidence of relapse (CIR) for MDD-positive/MRD-positive patients was 81% versus 31% for MDD-positive/MRD-negative and 15% for MDD-negative patients.
T-LBL and pediatric data: In a study of 99 children with T-cell lymphoblastic lymphoma (T-LBL), two-year event-free survival was significantly affected by CTC levels at presentation: patients with 1% or more T-LBL cells in bone marrow had EFS of 68.1% plus or minus 11.1%, compared to 90.7% plus or minus 4.4% for those with less than 1% (P = 0.031). Detection of MRD during remission induction therapy identified patients with slower disease clearance and was associated with worse prognosis. High-throughput sequencing (HTS) of TCR beta and gamma CDR3 regions identified 31 out of 43 patients (72%) by TCR-beta HTS and 27 out of 43 (63%) by TCR-gamma HTS, outperforming conventional multiparameter flow cytometry for MRD detection (25 out of 35 vs. 13 out of 35 patients).
Epstein-Barr virus DNA refers to EBV genomic fragments circulating in body fluids, either actively released from viable infected cells or passively shed during apoptosis and necrosis. EBV DNA is an established tumor biomarker for nasopharyngeal carcinoma, and accumulating evidence supports its clinical utility in EBV-related T-cell lymphomas, particularly ENKTL. Nearly all ENKTL tumors have a strong association with EBV infection, making EBV DNA quantification a particularly specific and practical liquid biopsy target for this subtype.
Diagnostic performance: In one study, quantitative real-time PCR (qRT-PCR) detected plasma EBV DNA in 17 of 18 patients with NK/T-cell lymphoma (94.4%) but in 0 of 35 healthy controls (P less than 0.0001), establishing high diagnostic specificity. Among 39 EBV-positive lymphoma patients, EBV DNA levels showed dynamic treatment-related changes, while 34 EBV-negative lymphoma cases had no EBV DNA detected at any disease stage. However, a critical challenge remains: distinguishing tumor-derived EBV DNA from non-tumor EBV DNA arising from latent viral infection in healthy B cells requires careful assay design, and the clinical utility of EBV DNA detection may depend on accurately identifying PTCL-specific viral sequences.
Prognostic cutoffs and survival data: Pre-treatment EBV DNA levels are a well-validated prognostic factor in ENKTL. Patients with pre-treatment EBV DNA at or above 500 copies/mL had a 3-year PFS of 12.2% compared to 48.2% for those below 500 copies/mL (P = 0.001), and 3-year OS of 42.5% vs. 66.1% (P = 0.003). In a separate study, patients with high baseline EBV DNA (above 6.1 x 10^7 copies/mL) had significantly inferior disease-free survival, and patients with EBV DNA positivity during treatment had inferior OS. Normalization of EBV DNA after first-line chemotherapy in ENKTL was associated with an objective response rate of 81.3% vs. 22.2% (P = 0.014), PFS of 12.0 months vs. 3.7 months (P = 0.011), and OS of 37.9 months vs. 7.8 months (P = 0.012).
MRD and treatment monitoring: Serial EBV DNA measurement during P-GEMOX treatment correlated with clinical treatment responses in ENKTL. In EBV-DNA-positive patients before treatment, achieving a negative EBV DNA status after treatment was significantly associated with better PFS and OS, while persistent EBV DNA positivity after treatment predicted higher rates of relapse and progression. These findings position plasma EBV DNA as a precise surrogate for tumor burden in ENKTL that can guide post-treatment assessment and surveillance. Interim-positive and post-treatment-positive EBV DNA had 3-year OS of 53.8% and 40.6%, respectively, compared to 99.1% and 91.8% for interim-negative and post-treatment-negative patients (P less than 0.001 for both).
Beyond nucleic acid biomarkers, the immune system itself generates detectable signals in T-cell lymphoma. Tumor-associated genetic alterations alter "self-antigen" expression, triggering autoantibody production that can be detected in peripheral blood. Separately, cytokines secreted by tumor cells and the tumor microenvironment appear at measurable concentrations in serum and plasma, providing indirect information about disease biology and drug resistance.
ALK antibodies in ALCL: ALK-positive ALCL accounts for approximately 15% of pediatric NHL and is characterized by the t(2;5)(p23;q35) translocation producing the NPM1-ALK fusion protein, a potent oncogenic tyrosine kinase. Patients with ALK-positive ALCL generate circulating anti-ALK autoantibodies. Mussolin et al. detected these antibodies in 25 of 28 patients at diagnosis (89%), while NPM::ALK transcripts were found in peripheral blood in only 15 of 28 (54%), demonstrating that antibody detection has broader sensitivity. An antibody titer cutoff of 1:2250 predicted relapse with 75% sensitivity and 55% specificity. In a larger cohort of 128 patients, combining MDD status with antibody titers defined three risk groups: high-risk patients (MDD-positive plus antibody titer 1:750 or below) had PFS of 28% and OS of 71%; low-risk patients (MDD-negative plus titer above 1:750) had PFS of 98% and OS of 98%; and intermediate-risk patients had PFS of 68% and OS of 83% (P = 0.02).
Cytokines and drug resistance in ENKTL: NK/T-cell lymphoma is plagued by high relapse rates and poor post-chemotherapy survival, primarily due to chemotherapy resistance. Two cytokines have been specifically implicated in this resistance phenotype. Ni et al. demonstrated that interleukin-13 (IL-13) was highly expressed in NK/T-cell lymphoma cells and promoted ABCC4 expression, which in turn increased resistance to asparaginase. Huo et al. showed that IL-10 similarly enhanced ABCC4-mediated gemcitabine resistance in lymphoma cells. These findings position serum cytokine profiling as a window into active resistance mechanisms, not merely a marker of disease burden.
Prognostic cytokine data: In 117 patients with peripheral T-cell NHL, soluble TNF receptor II (sTNFRII) serum levels at or above 2.16 ng/mL were associated with a 2.07-fold increased relative risk for shorter OS (univariate P = 0.0034) and a 2.49-fold higher risk for shorter event-free survival (multivariate HR 2.49, 95% CI 1.22-5.08, P = 0.012). Cytokine detection methods include ELISA and PCR for single-analyte measurement, and Luminex multiplex immunoassay for simultaneous quantification of multiple cytokines using fluorescent microbeads. Mass cytometry (CyTOF), which combines flow cytometry with mass spectrometry, enables simultaneous measurement of over 40 cellular parameters at single-cell resolution but requires distinct metal isotope probes and higher cell input.
One of the review's most practically valuable contributions is a subtype-by-subtype framework recommending the most appropriate liquid biopsy assays based on known pathological characteristics and molecular genetic profiles. This approach reflects the reality that no single biomarker is optimal across all T-cell lymphoma subtypes: the most informative liquid biopsy signal differs depending on which genetic events drive each disease entity.
nTFHL-AI: This subtype is characterized by TFH cell markers (PD-1, ICOS, CXCL13, CD10, BCL6) and a recurrent mutational spectrum featuring TET2 loss-of-function in approximately 80% of cases, DNMT3A in 30-40%, and IDH2 R172 or RHOA G17V in a subset. TCR clonal rearrangements occur in 75-90% of cases. The authors assign cfDNA/ctDNA as the top-priority assay (three stars), CTCs as acceptable (two stars), and EBV DNA as situational (one star). For ALK-positive ALCL, characterized by NPM1-ALK fusion and recurrent mutations in NOTCH1, TP53, and epigenetic modulators KMT2D and EP300, antibodies receive top priority (three stars), with CTCs and cfDNA/ctDNA both rated acceptable (two stars). For ENKTL, defined by its near-universal EBV association and JAK/STAT pathway mutations (STAT3, JAK3, STAT5B), EBV DNA receives the highest priority (two stars) with cfDNA/ctDNA situational (one star).
Clinical trials overview: More than 25 active clinical trials incorporate liquid biopsy endpoints for T-cell lymphoma, registered across NCT and ChiCTR databases. The most commonly investigated biomarkers are MRD (assessed via Ig/TCR rearrangement tracking), cfDNA, ctDNA, and quantitative EBV DNA copy numbers. A smaller number of trials examine cytokine profiling, SNP signatures, ALK variant levels, epigenetic changes, and T-cell-specific neoantigens. Trials span ENKTL (e.g., NCT04676789 directly incorporating plasma EBV DNA into design), PTCL (e.g., ChiCTR2200060450 for dynamic ctDNA monitoring in PTCL patients in China), pediatric ALCL (CCCG-ALCL-2020 for high-risk disease), and cutaneous T-cell lymphoma subtypes including mycosis fungoides and Sezary syndrome. Most trials use single biomarker designs, though a small subset explores comprehensive multi-dimensional profiling.
PTCL, NOS and ALK-negative ALCL: PTCL, NOS is a diagnosis of exclusion representing a biologically heterogeneous group. Its two biological subtypes, PTCL-GATA3 and PTCL-TBX21, differ in genetic complexity: GATA3 subtype features TP53, CDKN2A/B, and PTEN losses with STAT3 and MYC amplifications, while TBX21 subtype shows low genomic complexity and few recurrent alterations. Both cfDNA/ctDNA (two stars) and CTCs (one star) are recommended. ALK-negative ALCL, characterized by JAK-STAT signaling activation in 60% of cases via JAK1, JAK3, and STAT3 mutations, receives comparable recommendations of CTCs and cfDNA/ctDNA as acceptable assays (two stars each), given the absence of a single defining molecular target analogous to NPM1-ALK in ALK-positive disease.
Standardization of biomarker detection: The most fundamental barrier to routine clinical use is the absence of standardized, validated detection protocols. Different PCR platforms for EBV DNA measurement use different amplicon regions, quantification standards, and positivity thresholds, generating variability in sensitivity, specificity, and detection limits that makes cross-institutional comparison unreliable. For cfDNA/ctDNA, a specific challenge involves CHIP (clonal hematopoiesis of indeterminate potential): age-related somatic mutations in hematopoietic cells can be released into plasma as cfDNA and mistaken for tumor-derived ctDNA. High-depth paired leukocyte sequencing has been proposed to classify and quantify cfDNA variant sources, but this adds cost and analytical complexity.
Diagnostic threshold determination: Liquid biopsy cannot yet fully replace tissue biopsy as the primary diagnostic tool because establishing tumor origin from circulating fragments and defining validated diagnostic thresholds remain unresolved problems. Systematic analysis of cfDNA methylation profiles for tissue-of-origin tracing represents one promising approach. The integration of DNA sequencing with methylation profiling has improved the positive predictive value of plasma EBV DNA for lymphoma diagnosis, but larger cohort validation is required before these approaches reach guideline-level evidence.
Prognostic and monitoring cutoffs: Determining the optimal threshold values for prognostic scoring and treatment response assessment is an active challenge. Distinguishing genuine disease progression from pseudoprogression in patients on immunotherapy or novel agents remains difficult even with serial ctDNA monitoring, though longitudinal ctDNA profiling can enable early differentiation between these outcomes. The appropriate sampling frequency for continuous liquid biopsy monitoring has not been established and requires both clinical validation and ethical regulatory frameworks governing repeated patient sampling.
Scale of clinical evidence: Compared to solid tumors such as lung cancer, colorectal cancer, and breast cancer where liquid biopsy has advanced to clinical-grade applications, the evidence base for T-cell lymphoma is substantially smaller due to the rarity of individual subtypes. Most published studies involve small single-institution cohorts. Large-scale, multi-center prospective studies are needed to validate liquid biopsy as a clinical tool across the full spectrum of T-cell lymphoma subtypes. Interdisciplinary collaboration involving oncologists, pathologists, molecular biologists, and bioinformaticians is required to develop the analytical pipelines, reference standards, and interpretive frameworks that will enable clinical-grade liquid biopsy in this field.