The Mutation Is Not Necessarily the Lymphoma: Tumour-Infiltrating Clonal Haematopoiesis, Shared Clonal Ancestry and Cellular Attribution in Lymphoid Tissue Sequencing

1. Introduction: The Specimen is Not the Malignant Clone

Contemporary lymphoma classification rests upon the integration of tissue architecture, cytomorphology, immunophenotype, molecular findings and clinical context; no isolated sequence variant can supplant that assessment [1,2]. Clonal haematopoiesis (CH), meanwhile, denotes the expansion of genetically altered haematopoietic populations without independently establishing malignancy, while recognition of lymphoid-associated CH has broadened understanding of the relationship between acquired mutations and lymphomagenesis [3,4,5].

Misattribution of CH-associated variants during tumour sequencing is well recognised in solid malignancies, and a recent intended-use framework addresses variant provenance, host biomarkers and clinical actionability in that setting [6,7]. The lymphoma liquid-biopsy literature has likewise acknowledged the need to distinguish tumour-derived molecular signals from non-malignant sources [8]. The unresolved difficulty is therefore not sequencing interference itself, but an evolutionary ambiguity intrinsic to haematological malignancy: malignant and non-malignant populations may inherit the same alteration from a common haematopoietic progenitor [9,10,11].

Accordingly, detection in blood cannot automatically disqualify a variant from the malignant lineage, just as its detection within lymphoma tissue cannot establish malignant-cell ownership [9,10,11]. These opposing attribution errors expose the inadequacy of a binary tumour-versus-normal interpretation that conflates cellular localisation, evolutionary ancestry and pathogenic significance [7,9,10,11].

This review addresses a more exacting question: what constitutes sufficient evidence to attribute a lymphoid tissue variant to malignant cells, a shared ancestor or a biologically consequential process [7,9,10,11]? The frequency of clinically significant attribution errors remains undetermined; biological ambiguity must not be mistaken for demonstrated diagnostic misclassification [9,11].

2. Scope and Evidence Selection

A targeted narrative literature search was undertaken through 28 October 2026, drawing upon PubMed-indexed publications and accessible records from Springer Nature, the American Society of Hematology, the American Association for Cancer Research and Wiley. Search combinations encompassed CH, lymphoma, shared ancestry, tumour infiltration, molecular attribution, haematopoietic progenitors, compartment-resolved sequencing and lymphoid clonality.

Priority was given to primary investigations establishing cellular or evolutionary relationships, contemporary lymphoma classifications, molecular-testing guidance and publications presenting competing interpretations of CH-associated findings [1,2,7,8,9,10,11]. This selective critical synthesis neither claims exhaustive retrieval nor derives pooled prevalence estimates; methodological differences, discordant observations and limitations of inference are retained rather than obscured [9,11].

3. The Shared-Ancestry Paradox: Two Errors, One Mutation

A variant detected in lymphoma tissue may reside within malignant lymphocytes, infiltrating non-malignant haematopoietic cells or both populations [9,11]. Even unequivocal cellular localisation does not establish whether the alteration preceded malignant transformation, accompanied lineage divergence or contributed directly to oncogenesis [9,10,11].

Venanzi and colleagues examined 40 cases of classical Hodgkin lymphoma (cHL), identifying CH in five and extensive microenvironmental involvement in three [9]. Notably, one case demonstrated a DNMT3A/TET2-mutated population contributing to the malignant lineage, whereas another exhibited widespread DNMT3A-mutated haematopoiesis without evidence that the Hodgkin–Reed–Sternberg clone arose from that population [9]. Thus, apparently comparable infiltration by mutated haematopoietic cells may conceal fundamentally different evolutionary relationships [9].

Tiacci and colleagues established a shared ancestral relationship between angioimmunoblastic T-cell lymphoma and NPM1-mutated acute myeloid leukaemia, while Lewis and colleagues identified CH-associated mutations shared by neoplastic T-cell and myeloid compartments in 15 of 22 patients [10,11]. Subsequent analyses involving RHOA and TET2 support multistep evolution in which early haematopoietic alterations precede later lymphoid-associated events [12,13].

The paradox is therefore two-sided: indiscriminate subtraction of blood-shared variants may discard genuine lymphoma ancestry, whereas indiscriminate retention of tissue-detected variants may misattribute infiltrating CH to malignant cells [9,10,11]. Neither error can be remedied by sequencing depth alone, because the uncertainty concerns biological provenance rather than analytical sensitivity [9,10,11].

Table 1 distinguishes these competing explanations and the evidence required to resolve them [9,10,11,12,13,14].

4. B-Cell Lymphoma: Emerging Evidence, Uncertain Diagnostic Consequence

Wiegand and colleagues investigated 43 patients with B-cell non-Hodgkin lymphoma using complementary molecular approaches, identifying peripheral-blood CH-associated variants at a variant allele frequency of at least 1% in 55% of patients and identical CH clones in 41% of corresponding lymphoma specimens [14]. Their analyses distinguished non-enriched infiltrating clones from populations preferentially expanded within lymphoma tissue and identified lymphoma-associated alterations in haematopoietic progenitors from selected indolent B-cell neoplasms [14].

Although the investigators designated certain preferentially expanded populations as tumour-promoting CH, clonal enrichment does not independently establish mutation-specific causation or demonstrate direct promotion of malignant growth [14]. Equally, an alteration shared by haematopoietic progenitors and malignant B cells may document ancestral involvement without identifying the event responsible for transformation [14].

Chen and colleagues subsequently evaluated 304 treatment-naïve patients with diffuse large B-cell lymphoma (DLBCL), detecting tumour-infiltrating CH in 13.5% and demonstrating an independent association with adverse clinical outcomes [15]. Their investigation also identified inflammatory myeloid transcriptional enrichment and a 24-gene prognostic signature associated with survival across four cohorts, retaining multivariable significance in three [15].

These findings establish a compelling rationale for biological and prognostic investigation, but do not demonstrate that routine compartment-resolved sequencing improves lymphoma classification, treatment selection or survival [14,15]. Furthermore, differences in ascertainment, cohort composition and molecular methodology preclude direct equivalence between the reported frequencies [14,15].

5. Clonal Proliferation Does Not Establish Malignant Transformation

The distinction between molecular clonality and malignancy becomes particularly consequential when atypical lymphoid proliferations lack sufficient morphological or immunophenotypic evidence for an established lymphoma diagnosis [1,2]. Brune and colleagues described seven unusual lymphoid hyperplasias harbouring lymphoma-associated abnormalities and proposed possible tissue equivalents of lymphoid CH, advancing a precursor-lesion hypothesis rather than validating a new diagnostic entity [16].

One patient subsequently developed follicular lymphoma carrying the same non-productive immunoglobulin rearrangement identified in the earlier lesion, supporting a possible evolutionary relationship without establishing a generalisable risk of transformation [16]. Attygalle and colleagues independently documented parallel evolution of distinct lymphoid proliferations from a shared mutated haematopoietic precursor, demonstrating that clonal relationships need not imply strictly linear progression [17].

In a separate cohort of 30 patients with T-follicular helper-cell lymphoma, Wei and colleagues detected peripheral-blood CH-associated mutations in 40% and reported adverse clinical associations, although the limited sample size and observational design constrain broader inference [18]. Collectively, these findings establish neither a universal precursor state nor a validated indication for surveillance based solely upon CH-associated mutations [16,17,18].

Immunoglobulin and T-cell receptor clonality identify rearrangement-defined populations rather than malignancy itself; their diagnostic significance depends upon concordance with morphology, immunophenotype and clinical findings [19]. Molecular evidence must therefore inform pathological classification without allowing clonal expansion to become a surrogate for malignant transformation [1,2,16,19].

Table 2 delineates the principal disease-specific attribution questions and limitations of the available evidence [9,10,11,12,13,14,15,16,17,18,19].

6. The Analytical Boundary: Detection is Not Attribution

Bulk sequencing establishes the presence of a variant within the analysed specimen but ordinarily cannot determine whether it resides in malignant lymphocytes or accompanying haematopoietic populations [6,9,14]. Variant allele frequency is additionally influenced by cellular composition, copy-number abnormalities, allelic imbalance and assay performance; without appropriate assumptions, it cannot be equated with the proportion of malignant cells carrying the alteration [14].

Paired tumour–normal sequencing improves molecular interpretation in haematological malignancies, as demonstrated by Ptashkin and colleagues, but the choice of comparator is consequential [20]. Peripheral blood may harbour CH-associated or shared ancestral mutations, while saliva and buccal specimens may contain haematopoietic DNA; appropriately validated non-haematopoietic material may therefore be necessary for selected constitutional investigations [20,22,23].

Established AMP/ASCO/CAP guidance addresses variant interpretation and reporting, while ACMG recommendations address the identification and interpretation of suspected constitutional and acquired alterations [21,22,23]. These safeguards are indispensable, but they do not independently reconstruct the evolutionary relationships between malignant and non-malignant populations carrying identical somatic variants [9,10,11,20,21,22,23].

Flow-sorted sequencing and laser-capture microdissection can strengthen cellular localisation, although contamination, limited material and incomplete population sampling remain important constraints [9,14]. Single-cell genotyping can associate mutations with defined cellular phenotypes, but allelic dropout, insufficient coverage and unsampled subclones complicate interpretation, particularly of negative findings [14]. A negative result in an enriched population cannot establish biological absence without adequate cellular purity, locus-specific coverage, analytical sensitivity and appropriate technical controls; conversely, detecting a mutation within that population establishes association, not pathogenic necessity [9,14,21].

The appropriate question is therefore not which technology offers the greatest resolution, but which unresolved biological inference the selected method can legitimately address [9,14,20,21,22,23].

Table 3 distinguishes what available approaches establish from conclusions requiring additional evidence [9,14,19,20,21,22,23].

7. Biological Association is Not Causal Involvement

Tumour-infiltrating CH may influence cancer-associated immunity, as investigations in solid malignancies have linked genetically altered haematopoietic populations to changes in the tumour microenvironment [24]. Those mechanisms, however, cannot be transferred uncritically to lymphoma, where malignant and non-malignant haematopoietic compartments may share an ancestral genotype [9,10,11,12,13,14,15,24].

Current lymphoma studies provide substantive evidence of cellular relationships and prognostic associations but do not consistently establish mutation-specific mechanisms of tumour promotion [14,15]. Demonstrating causation requires functional perturbation capable of distinguishing clonal abundance from clonal activity, while clinical utility requires reproducible incremental benefit beyond established diagnostic or prognostic methods [14,15,21,24].

Cellular ownership, ancestral participation, pathogenic causation and clinical actionability must consequently remain distinct interpretations, even when each concerns the same sequence variant [9,10,11,12,13,14,15,21].

8. Evidentiary Sufficiency and Proportionate Diagnostic Investigation

The principal conceptual contribution of this review is a claim-specific framework of evidentiary sufficiency, not another system of variant classification [7,19,20,21,22,23]. Analytical detection requires a technically validated sequence finding; cellular attribution requires adequately characterised populations; shared ancestry requires concordant evolutionary evidence; causal involvement requires functional corroboration; and clinical utility requires validated improvement in a defined decision [9,10,11,12,13,14,15,19,20,21,22,23,24].

These evidentiary requirements are neither interchangeable nor automatically sequential: analytical validity does not establish cellular ownership, cellular ownership does not establish shared ancestry, shared ancestry does not establish oncogenic causation, and oncogenic involvement does not independently establish clinical utility [9,10,11,12,13,14,15,19,20,21].

Nor do these distinctions imply that every variant warrants progressively more elaborate investigation [14,20,21]. Additional testing is justified only where attribution uncertainty could materially alter an integrated pathological conclusion and an available method can discriminate between the competing explanations [1,2,14,20,21].

An incidental alteration that neither challenges an established diagnosis nor addresses a defined clinical question does not, by itself, justify compartment-resolved sequencing [14,21]. Conversely, a potentially consequential molecular–pathological discordance warrants reassessment of specimen composition, analytical validity, comparator suitability and the feasibility of targeted cellular enrichment [9,14,20,21,22,23].

Where attribution remains unresolved, reporting should identify the variant as present in the submitted specimen rather than unequivocally assigning it to the malignant clone [9,14,21]. Table 4 sets out the proposed evidentiary thresholds and their relationship to diagnostic consequence; it constitutes a critical interpretative synthesis, not a validated clinical algorithm [14,15,20,21].

9. The Decisive Validation Requirement

Future investigations should prospectively compare bulk-tissue interpretations with prespecified compartment-resolved reference assessments, incorporating suitable constitutional controls, orthogonal confirmation and disease-specific stratification [9,14,20,21,22,23]. Analytical endpoints should prioritise attribution concordance, resolution of uncertainty and reproducibility across sampling and enrichment strategies, rather than simply increasing variant detection [14,20,21].

Clinical evaluation should independently examine changes in integrated diagnosis, interobserver agreement, incremental prognostic discrimination and demonstrable effects upon justified management decisions [1,2,14,15,21]. Failure to translate improved cellular attribution into a clinically meaningful difference would constitute an equally informative result, demonstrating that greater biological resolution does not necessarily warrant routine diagnostic implementation [14,15,21].

10. Conclusion

A mutation shared by malignant and non-malignant haematopoietic cells may represent neither an irrelevant contaminant nor a lymphoma-defining driver, but an ancestral alteration whose significance cannot be established through bulk sequencing alone [9,10,11,12,13,14,15]. The objective is therefore not universal compartmental analysis, but attribution proportionate to the scientific claim and its potential diagnostic consequence [14,20,21].

Analytical sensitivity determines what can be detected; pathological certainty depends upon what the evidence can legitimately establish [9,14,21].

Compliance with Ethical Standards