Combining gene expression profiling and machine learning to diagnose B-cell non-Hodgkin lymphoma

Role of artificial intelligence in haematolymphoid diagnostics

The advent of digital pathology and the deployment of high-throughput molecular techniques are generating an unprecedented mass of data. Thanks to advances in computational sciences, artificial intelligence (AI) approaches represent a promising avenue for extracting relevant information from complex data structures. From diagnostic assistance to powerful research tools, the potential fields of application of machine learning techniques in pathology are vast and constitute the subject of considerable research work. The aim of this article is to provide an overview of the potential applications of AI in the field of haematopathology and to define the role that these emerging technologies could play in our laboratories in the short to medium term.

Detection of disease-specific signatures in B cell repertoires of lymphomas using machine learning

The classification of B cell lymphomas-mainly based on light microscopy evaluation by a pathologist-requires many years of training. Since the B cell receptor (BCR) of the lymphoma clonotype and the microenvironmental immune architecture are important features discriminating different lymphoma subsets, we asked whether BCR repertoire next-generation sequencing (NGS) of lymphoma-infiltrated tissues in conjunction with machine learning algorithms could have diagnostic utility in the subclassification of these cancers. We trained a random forest and a linear classifier via logistic regression based on patterns of clonal distribution, VDJ gene usage and physico-chemical properties of the top-n most frequently represented clonotypes in the BCR repertoires of 620 paradigmatic lymphoma samples-nodular lymphocyte predominant B cell lymphoma (NLPBL), diffuse large B cell lymphoma (DLBCL) and chronic lymphocytic leukemia (CLL)-alongside with 291 control samples. With regard to DLBCL and CLL, the models demonstrated optimal performance when utilizing only the most prevalent clonotype for classification, while in NLPBL-that has a dominant background of non-malignant bystander cells-a broader array of clonotypes enhanced model accuracy. Surprisingly, the straightforward logistic regression model performed best in this seemingly complex classification problem, suggesting linear separability in our chosen dimensions. It achieved a weighted F1-score of 0.84 on a test cohort including 125 samples from all three lymphoma entities and 58 samples from healthy individuals. Together, we provide proof-of-concept that at least the 3 studied lymphoma entities can be differentiated from each other using BCR repertoire NGS on lymphoma-infiltrated tissues by a trained machine learning model.

Machine Learning-Supported Diagnosis of Small Blue Round Cell Sarcomas Using Targeted RNA Sequencing

Small blue round cell sarcomas (SBRCSs) are a heterogeneous group of tumors with overlapping morphologic features but markedly varying prognosis. They are characterized by distinct chromosomal alterations, particularly rearrangements leading to gene fusions, whose detection currently represents the most reliable diagnostic marker. Ewing sarcomas are the most common SBRCSs, defined by gene fusions involving EWSR1 and transcription factors of the ETS family, and the most frequent non-EWSR1-rearranged SBRCSs harbor a CIC rearrangement. Unfortunately, currently the identification of CIC::DUX4 translocation events, the most common CIC rearrangement, is challenging. Here, we present a machine-learning approach to support SBRCS diagnosis that relies on gene expression profiles measured via targeted sequencing. The analyses on a curated cohort of 69 soft-tissue tumors showed markedly distinct expression patterns for SBRCS subgroups. A random forest classifier trained on Ewing sarcoma and CIC-rearranged cases predicted probabilities of being CIC-rearranged >0.9 for CIC-rearranged-like sarcomas and <0.6 for other SBRCSs. Testing on a retrospective cohort of 1335 routine diagnostic cases identified 15 candidate CIC-rearranged tumors with a probability >0.75, all of which were supported by expert histopathologic reassessment. Furthermore, the multigene random forest classifier appeared advantageous over using high ETV4 expression alone, previously proposed as a surrogate to identify CIC rearrangement. Taken together, the expression-based classifier can offer valuable support for SBRCS pathologic diagnosis.

Analysis of immunoglobulin/T-cell receptor repertoires by high-throughput RNA sequencing reveals a continuous dynamic of positive clonal selection in follicular lymphoma

Follicular lymphoma (FL) course is highly variable, making its clinical management challenging. In this incurable and recurring pathology, the interval between relapses tends to decrease while aggressiveness increases, sometimes resulting in the transformation to higher-grade lymphoma. These evolutions are particularly difficult to anticipate, resulting from complex clonal evolutions where multiple subclones compete and thrive due to their capacity to proliferate and resist therapies. Here, to apprehend further these processes, we used a high-throughput RNA sequencing approach to address simultaneously the B-cell immunoglobulin repertoires and T-cell immunoglobulin repertoires repertoires of lymphoma cells and their lymphoid microenvironment in a large cohort of 131 FL1/2-3A patients. Our data confirm the existence of a high degree of intra-clonal heterogeneity in this pathology, resulting from ongoing somatic hyper-mutation and class switch recombination. Through the evaluation of the Simpson ecological-diversity index, we show that the contribution of the cancerous cells increases during the course of the disease to the detriment of the reactive compartment, a phenomenon accompanied by a concomitant decrease in the diversity of the tumoral population. Clonal evolution in FL thus contrasts with many tumors, where clonal heterogeneity steadily increases over time and participates in treatment evasion. In this pathology, the selection of lymphoma subclones with proliferative advantages progressively outweighs clonal diversification, ultimately leading in extreme cases to transformation to high-grade lymphoma resulting from the rapid emergence of homogeneous subpopulations.

High PDL1/PDL2 gene expression correlates with worse outcome in primary mediastinal large B-cell lymphoma

Primary mediastinal B-cell lymphoma (PMBL) is an uncommon entity of aggressive B-cell lymphoma with an unusually good prognosis, except for 10-15% of chemotherapy-refractory cases. To identify earlier these higher risk patients, we performed molecular characterization of a retrospective multicenter cohort of patients treated with firstline immunochemotherapy. The traits of the patients with gene-expression profiling data (n = 120) were as follows: median age of 34 years (range, 18-67 years); female sex, 58.3%; elevated lactate dehydrogenase, 82.5%; Eastern Cooperative Oncology Group performance status score of 0 to 1, 85.7%; Ann Arbor stage I/II, 55%; International Prognostic Index score of 1 to 2, 64.4%; and median metabolic tumor volume, 290.4 cm3 (range, 15.7-1147.5 cm3). Among all 137 markers tested for correlation with survival data, only programmed death-ligand (PDL) 1 and PDL2 expression showed a prognostic impact. Overall, both PDL1 and PDL2 genes were highly expressed in 37 patients (30.8%; PDL1high/PDL2high). The baseline clinical characteristics of patients with PDL1high/PDL2high were similar to those of other patients. In univariate analysis, PDL1high/PDL2high status was associated with poor progression-free survival (PFS) (hazard ratio [HR], 4.292) and overall survival (OS; HR, 8.24). In multivariate analysis, PDL1high/PDL2high status was an independent prognostic factor of adverse outcomes (PFS: HR, 5.22; OS: HR, 10.368). We validated these results in an independent cohort of 40 patients and confirmed the significant association between PDL1high/PDL2high status and inferior PFS (HR, 6.11). High PDL1/PDL2 gene expression defines a population with strong immune privilege and poorer outcomes from standard chemotherapy who might benefit from firstline checkpoint inhibitor therapy.

Machine learning models-based on integration of next-generation sequencing testing and tumor cell sizes improve subtype classification of mature B-cell neoplasms

Background

Next-generation sequencing (NGS) panels for mature B-cell neoplasms (MBNs) are widely applied clinically but have yet to be routinely used in a manner that is suitable for subtype differential diagnosis. This study retrospectively investigated newly diagnosed cases of MBNs from our laboratory to investigate mutation landscapes in Chinese patients with MBNs and to combine mutational information and machine learning (ML) into clinical applications for MBNs, especially for subtype classification.

Methods

Samples from the Catalogue Of Somatic Mutations In Cancer (COSMIC) database were collected for ML model construction and cases from our laboratory were used for ML model validation. Five repeats of 10-fold cross-validation Random Forest algorithm was used for ML model construction. Mutation detection was performed by NGS and tumor cell size was confirmed by cell morphology and/or flow cytometry in our laboratory.

Results

Totally 849 newly diagnosed MBN cases from our laboratory were retrospectively identified and included in mutational landscape analyses. Patterns of gene mutations in a variety of MBN subtypes were found, important to investigate tumorigenesis in MBNs. A long list of novel mutations was revealed, valuable to both functional studies and clinical applications. By combining gene mutation information revealed by NGS and ML, we established ML models that provide valuable information for MBN subtype classification. In total, 8895 cases of 8 subtypes of MBNs in the COSMIC database were collected and utilized for ML model construction, and the models were validated on the 849 MBN cases from our laboratory. A series of ML models was constructed in this study, and the most efficient model, with an accuracy of 0.87, was based on integration of NGS testing and tumor cell sizes.

Conclusions

The ML models were of great significance in the differential diagnosis of all cases and different MBN subtypes. Additionally, using NGS results to assist in subtype classification of MBNs by method of ML has positive clinical potential.

Integrative Prognostic Machine Learning Models in Mantle Cell Lymphoma

Patients with mantle cell lymphoma (MCL), an incurable B-cell malignancy, benefit from accurate pretreatment disease stratification. We curated an extensive database of 862 patients diagnosed between 2014 and 2022. A machine learning (ML) gradient-boosted model incorporated baseline features from clinicopathologic, cytogenetic, and genomic data with high predictive power discriminating between patients with indolent or responsive MCL and those with aggressive disease (AUC ROC = 0.83). In addition, we utilized the gradient-boosted framework as a robust feature selection method for multivariate logistic and survival modeling. The best ML models incorporated features from clinical and genomic data types highlighting the need for correlative molecular studies in precision oncology. As proof of concept, we launched our most accurate and practical models using an application interface, which has potential for clinical implementation. We designated the 20-feature ML model-based index the "integrative MIPI" or iMIPI and a similar 10-feature ML index the "integrative simplified MIPI" or iMIPI-s. The top 10 baseline prognostic features represented in the iMIPI-s are: lactase dehydrogenase (LDH), Ki-67%, platelet count, bone marrow involvement percentage, hemoglobin levels, the total number of observed somatic mutations, TP53 mutational status, Eastern Cooperative Oncology Group performance level, beta-2 microglobulin, and morphology. Our findings emphasize that prognostic applications and indices should include molecular features, especially TP53 mutational status. This work demonstrates the clinical utility of complex ML models and provides further evidence for existing prognostic markers in MCL.

Significance

Our model is the first to integrate a dynamic algorithm with multiple clinical and molecular features, allowing for accurate predictions of MCL disease outcomes in a large patient cohort.

Artificial Intelligence-Assisted Diagnostic Cytology and Genomic Testing for Hematologic Disorders

Artificial intelligence (AI) is a rapidly evolving field of computer science that involves the development of computational programs that can mimic human intelligence. In particular, machine learning and deep learning models have enabled the identification and grouping of patterns within data, leading to the development of AI systems that have been applied in various areas of hematology, including digital pathology, alpha thalassemia patient screening, cytogenetics, immunophenotyping, and sequencing. These AI-assisted methods have shown promise in improving diagnostic accuracy and efficiency, identifying novel biomarkers, and predicting treatment outcomes. However, limitations such as limited databases, lack of validation and standardization, systematic errors, and bias prevent AI from completely replacing manual diagnosis in hematology. In addition, the processing of large amounts of patient data and personal information by AI poses potential data privacy issues, necessitating the development of regulations to evaluate AI systems and address ethical concerns in clinical AI systems. Nonetheless, with continued research and development, AI has the potential to revolutionize the field of hematology and improve patient outcomes. To fully realize this potential, however, the challenges facing AI in hematology must be addressed and overcome.

How molecular advances may improve the diagnosis and management of PTCL patients

Peripheral T-cell lymphomas (PTCL) comprised more than 30 rare heterogeneous entities, representing 10 to 15% of adult non-Hodgkin lymphomas. Although their diagnosis is still mainly based on clinical, pathological, and phenotypic features, molecular studies have allowed for a better understanding of the oncogenic mechanisms involved and the refinement of many PTCL entities in the recently updated classifications. The prognosis remains poor for most entities (5-year overall survival < 30%), with current conventional therapies based on anthracyclin-based polychemotherapy regimen, despite many years of clinical trials. The recent use of new targeted therapies appears to be promising for relapsed/refractory patients, such as demethylating agents in T-follicular helper (TFH) PTCL. However further studies are needed to evaluate the proper combination of these drugs in the setting of front-line therapy. In this review, we will summarize the oncogenic events for the main PTCL entities and report the molecular targets that have led to the development of new therapies. We will also discuss the development of innovative high throughput technologies that aid the routine workflow for the histopathological diagnosis and management of PTCL patients.

Disequilibrium between BRCA1 and BRCA2 Circular and Messenger RNAs Plays a Role in Breast Cancer

Breast cancer is a frequent disease for which the discovery of markers that enable early detection or prognostic assessment remains challenging. Circular RNAs (circRNAs) are single-stranded structures in closed loops that are produced by backsplicing. CircRNA and messenger RNA (mRNA) are generated co-transcriptionally, and backsplicing and linear splicing compete against each other. As mRNAs are key players in tumorigenesis, we hypothesize that a disruption of the balance between circRNAs and mRNAs could promote breast cancer. Hence, we developed an assay for a simultaneous study of circRNAs and mRNAs, which we have called splice and expression analyses by exon ligation and high-throughput sequencing (SEALigHTS). Following SEALigHTS validation for BRCA1 and BRCA2, our hypothesis was tested using an independent research set of 95 pairs from tumor and adjacent normal breast tissues. In this research set, ratios of BRCA1 and BRCA2 circRNAs/mRNAs were significantly lower in the tumor breast tissue compared to normal tissue (p = 1.6 × 10-9 and p = 4.4 × 10-5 for BRCA1 and BRCA2, respectively). Overall, we developed an innovative method to study linear splicing and backsplicing, described the repertoire of BRCA1 and BRCA2 circRNAs, including 15 novel ones, and showed for the first time that a disequilibrium between BRCA1 and BRCA2 circRNAs and mRNAs plays a role in breast cancer.

Combined Reverse-Transcriptase Multiplex Ligation-Dependent Probe Amplification and Next-Generation Sequencing Analyses to Assign Unclassified BCL2-/BCL6- Nonrearranged Small B-Cell Lymphoid Neoplasms as Follicular or Nodal Marginal Zone Lymphoma

Distinguishing between follicular lymphoma (FL) and nodal marginal zone lymphoma (NMZL) can be difficult when morphologic and phenotypic features are unusual and characteristic cytogenetic rearrangements are absent. We evaluated the diagnostic contribution of ancillary techniques-including fluorescence in situ hybridization (FISH)-detected 1p36 deletion; reverse-transcriptase, multiplex, ligation-dependent probe amplification (RT-MLPA); and next-generation sequencing (NGS)-for tumors that remain unclassified according to standard criteria. After review, 50 CD5-negative small B-cell lymphoid neoplasms without BCL2 and BCL6 FISH rearrangements were diagnosed as FLs (n = 27), NMZLs (n = 5), or unclassified (n = 18) based on the 2016 World Health Organization Classification of Tumours of Haematopoietic and Lymphoid Tissues. FISH helped identify the 1p36 deletion in 3 FLs and 1 unclassified tumor. Most classified FLs had an RT-MLPA germinal center B-cell (GCB) signature (93%) or were noncontributive (7%). Classified NMZLs had an RT-MLPA activated B-cell signature (20%), had an unassigned signature (40%), or were noncontributive (40%). Among unclassified tumors, the RT-MLPA GCB signature was associated with mutations most commonly found in FLs (CREBBP, EZH2, STAT6, and/or TNFRSF14) (90%). An RT-MLPA-detected GCB signature and/or NGS-detected gene mutations were considered as FL identifiers for 13 tumors. An activated B-cell signature or NOTCH2 mutation supported NMZL diagnosis in 3 tumors. Combining the RT-MLPA and NGS findings successfully discriminated 89% of unclassified tumors in favor of one or the other diagnosis. NGS-detected mutations may be of therapeutic interest. Herein, we detected 3 EZH2 and 8 CREBBP mutations that might be eligible for targeted therapies.

Genomic landscape of mature B-cell non-Hodgkin lymphomas - an appraisal from lymphomagenesis to drug resistance

BACKGROUND

Mature B-cell non-Hodgkin lymphomas are one of the most common hematological malignancies with a divergent clinical presentation, phenotype, and course of disease regulated by underlying genetic mechanism.

MAIN BODY

Genetic and molecular alterations are not only critical for lymphomagenesis but also largely responsible for differing therapeutic response in these neoplasms. In recent years, advanced molecular tools have provided a deeper understanding regarding these oncogenic drives for predicting progression as well as refractory behavior in these diseases. The prognostic models based on gene expression profiling have also been proved effective in various clinical scenarios. However, considerable overlap does exist between the genotypes of individual lymphomas and at the same time where additional molecular lesions may be associated with each entity apart from the key genetic event. Therefore, genomics is one of the cornerstones in the multimodality approach essential for classification and risk stratification of B-cell non-Hodgkin lymphomas.

CONCLUSION

We hereby in this review discuss the wide range of genetic aberrancies associated with tumorigenesis, immune escape, and chemoresistance in major B-cell non-Hodgkin lymphomas.

Distinct Gene Expression Patterns of Calcium Channels and Related Signaling Pathways Discovered in Lymphomas

Cell surface calcium (Ca2+) channels permit Ca2+ ion influx, with Ca2+ taking part in cellular functions such as proliferation, survival, and activation. The expression of voltage-dependent Ca2+ (CaV) channels may modulate the growth of hematologic cancers. Profile analysis of Ca2+ channels, with a focus on the Ca2+ release-activated Ca2+ (CRAC) and L-type CaV channels, was performed on RNA sequencing data from lymphoma cell lines and samples derived from patients with diffuse large B cell lymphoma (DLBCL). CaV1.2 expression was found to be elevated in classical Hodgkin lymphoma (CHL) cell lines when compared to other B cell lymphoma cell lines. In contrast, CHL exhibited reduced expression of ORAI2 and STIM2. In our differential expression analysis comparing activated B cell-like DLBCL (ABC-DLBCL) and germinal centre B cell-like DLBCL (GCB-DLBCL) patient samples, ABC-DLBCL revealed stronger expression of CaV1.3, whereas CaV1.1, CaV1.2, and CaV1.4 showed greater expression levels in GCB-DLBCL. Interestingly, no differences in ORAI/STIM expression were noted in the patient samples. As Ca2+ is known to bind to calmodulin, leading to calcineurin activation and the passage of nuclear factor of activated T cells (NFAT) to the cell nucleus, pathways for calcineurin, calmodulin, NFAT, and Ca2+ signaling were also analyzed by gene set enrichment analysis. The NFAT and Ca2+ signaling pathways were found to be upregulated in the CHL cell lines relative to other B cell lymphoma cell lines. Furthermore, the calmodulin and Ca2+ signaling pathways were shown to be downregulated in the ABC-DLBCL patient samples. The findings of this study suggest that L-type CaV channels and Ca2+-related pathways could serve as differentiating components for biologic therapies in targeted lymphoma treatments.

Integrative diagnosis of primary cutaneous large B-cell lymphomas supports the relevance of cell of origin profiling

Primary cutaneous large B-cell lymphomas (PCLBCL) represent a diagnostic challenge because they are classified as PCLBCL, leg type (PCLBCL, LT) or primary cutaneous follicle centre lymphoma, large cell (PCFCL, LC), which differ by prognosis and therapeutic requirement. Unclassified cases with discordant clinical presentations, morphologies, and immunophenotypes may be classified into the not otherwise specified (PCLBCL, NOS) category based on ancillary molecular analyses. Cell-of-origin profiling as germinal centre (GC) type or non-GC type by immunohistochemistry is not considered reproducible because of variable CD10 expression. In a series of 55 PCLBCL cases with > 80% large cells, we reported 21 PCFCL, LC cases as GC-type and 27 PCLBCL, LT as non-GC-type; 7 cases were considered PCLBCL, NOS. Here, we demonstrate the accuracy of molecular profiling of PCLBCL as GC or non-GC type using a reverse transcriptase multiplex ligation assay (RT-MLPA). RT-MLPA classified the seven PCLBCL, NOS cases in accordance with their mutational profile. An integrative principal component analysis confirmed the main criteria and the relevance of genomic profiling of PCFCL, LC as GC-derived, and PCLBCL, LT as non-GC-derived. Both the cell-of-origin classification of PCLBCL and the integrative analysis identified two clinically relevant subgroups according to overall survival, which may help to standardize PCLBCL diagnosis and patient management.

A novel diagnostic approach for the classification of small B-cell lymphoid neoplasms based on the NanoString platform

Small B-cell lymphoid neoplasms (SBCLNs) are a heterogeneous group of diseases characterized by malignant clonal proliferation of mature B-cells. However, the classification of SBCLNs remains a challenge, especially in cases where histopathological analysis is unavailable or those with atypical laboratory findings or equivocal pathologic data. In this study, gene expression profiling of 1039 samples from 27 gene expression omnibus (GEO) datasets was first investigated to select highly and differentially expressed genes among SBCLNs. Samples from 57 SBCLN cases and 102 nonmalignant control samples were used to train a classifier using the NanoString platform. The classifier was built by employing a cascade binary classification method based on the random forest algorithm with 35 refined gene signatures. Cases were successively classified as chronic lymphocytic leukemia/small lymphocytic lymphoma, conventional mantle cell lymphoma, follicular lymphoma, leukemic non-nodal mantle cell lymphoma, marginal zone lymphoma, lymphoplasmacytic lymphoma/Waldenström's macroglobulinemia, and other undetermined. The classifier algorithm was then validated using an independent cohort of 197 patients with SBCLNs. Under the distribution of our validation cohort, the overall sensitivity and specificity of proposed algorithm model were >95%, respectively, for all the cases with tumor cell content greater than 0.72. Combined with additional genetic aberrations including IGH-BCL2 translocation, MYD88 L265P mutation, and BRAF V600E mutation, the optimal sensitivity and specificity were respectively found at 0.88 and 0.98. In conclusion, the established algorithm demonstrated to be an effective and valuable ancillary diagnostic approach for the sub-classification and pathologic investigation of SBCLN in daily practice.

DNA Methylation-Based Classification of Small B-Cell Lymphomas: A Proof-of-Principle Study

Although most small B-cell lymphomas (SBCLs) can be diagnosed using routine methods, challenges exist. For example, marginal zone lymphomas (MZLs) can be difficult to rule-in, in large part because no widely-used, sensitive, and specific biomarker is available for the marginal zone cell of origin. In this study, it was hypothesized that DNA methylation array profiling can assist with the classification of SBCLs, including MZLs. Extramedullary SBCLs, including challenging cases, were reviewed internally for pathology consensus and profiled. By combining the resulting array data set with data sets from other groups, a set of 26 informative probes was selected and used to train machine learning models to classify 4 common SBCLs: chronic lymphocytic leukemia/small lymphocytic lymphoma, follicular lymphoma, mantle cell lymphoma, and MZL. Prediction probability cutoff was used to separate classifiable from unclassifiable cases, and show that the trained model was able to classify 95% of independent test cases (n = 264/279). The concordance between model predictions and pathology diagnoses was 99.6% (n = 262/263) among classifiable test cases. One validation reference test case was reclassified based on model prediction. The model was also used to predict the diagnoses of two challenging SBCLs. Although the differential examined and data on difficult cases are limited, these results support accurate methylation-based classification of SBCLs. Furthermore, high specificities of predictions suggest that methylation signatures can be used to rule-in MZLs.

Detection of Gene Fusion Transcripts in Peripheral T-Cell Lymphoma Using a Multiplexed Targeted Sequencing Assay

The genetic basis of peripheral T-cell lymphoma (PTCL) is complex and encompasses several recurrent fusion transcripts discovered over the past years by means of massive parallel sequencing. However, there is currently no affordable and rapid technology for their simultaneous detection in clinical samples. Herein, we developed a multiplex ligation-dependent RT-PCR-based assay, followed by high-throughput sequencing, to detect 33 known PTCL-associated fusion transcripts. Anaplastic lymphoma kinase (ALK) fusion transcripts were detected in 15 of 16 ALK-positive anaplastic large-cell lymphomas. The latter case was further characterized by a novel SATB1_ALK fusion transcript. Among 239 other PTCLs, representative of nine entities, non-ALK fusion transcripts were detected in 24 samples, mostly of follicular helper T-cell (TFH) derivation. The most frequent non-ALK fusion transcript was ICOS_CD28 in nine TFH-PTCLs, one PTCL not otherwise specified, and one adult T-cell leukemia/lymphoma, followed by VAV1 rearrangements with multiple partners (STAP2, THAP4, MYO1F, and CD28) in five samples (three PTCL not otherwise specified and two TFH-PTCLs). The other rearrangements were CTLA4_CD28 (one TFH-PTCL), ITK_SYK (two TFH-PTCLs), ITK_FER (one TFH-PTCL), IKZF2_ERBB4 (one TFH-PTCL and one adult T-cell leukemia/lymphoma), and TP63_TBL1XR1 (one ALK-negative anaplastic large-cell lymphoma). All fusions detected by our assay were validated by conventional RT-PCR and Sanger sequencing in 30 samples with adequate material. The simplicity and robustness of this targeted multiplex assay make it an attractive tool for the characterization of these heterogeneous diseases.

Histopathologic and Machine Deep Learning Criteria to Predict Lymphoma Transformation in Bone Marrow Biopsies

CONTEXT.—

Large-cell transformation (LCT) of indolent B-cell lymphomas, such as follicular lymphoma (FL) and chronic lymphocytic leukemia (CLL), signals a worse prognosis, at which point aggressive chemotherapy is initiated. Although LCT is relatively straightforward to diagnose in lymph nodes, a marrow biopsy is often obtained first given its ease of procedure, low cost, and low morbidity. However, consensus criteria for LCT in bone marrow have not been established.

OBJECTIVE.—

To study the accuracy and reproducibility of a trained convolutional neural network in identifying LCT, in light of promising machine learning tools that may introduce greater objectivity to morphologic analysis.

DESIGN.—

We retrospectively identified patients who had a diagnosis of FL or CLL who had undergone bone marrow biopsy for the clinical question of LCT. We scored morphologic criteria and correlated results with clinical disease progression. In addition, whole slide scans were annotated into patches to train convolutional neural networks to discriminate between small and large tumor cells and to predict the patient's probability of transformation.

RESULTS.—

Using morphologic examination, the proportion of large lymphoma cells (≥10% in FL and ≥30% in CLL), chromatin pattern, distinct nucleoli, and proliferation index were significantly correlated with LCT in FL and CLL. Compared to pathologist-derived estimates, machine generated quantification demonstrated better reproducibility and stronger correlation with final outcome data.

CONCLUSIONS.—

These histologic findings may serve as indications of LCT in bone marrow biopsies. The pathologist-augmented with machine system appeared to be the most predictive, arguing for greater efforts to validate and implement these tools to further enhance physician practice.

Translational Applications of Artificial Intelligence and Machine Learning for Diagnostic Pathology in Lymphoid Neoplasms: A Comprehensive and Evolutive Analysis

Genomic analysis and digitalization of medical records have led to a big data scenario within hematopathology. Artificial intelligence and machine learning tools are increasingly used to integrate clinical, histopathological, and genomic data in lymphoid neoplasms. In this study, we identified global trends, cognitive, and social framework of this field from 1990 to 2020. Metadata were obtained from the Clarivate Analytics Web of Science database in January 2021. A total of 525 documents were assessed by document type, research areas, source titles, organizations, and countries. SciMAT and VOSviewer package were used to perform scientific mapping analysis. Geographical distribution showed the USA and People's Republic of China as the most productive countries, reporting up to 190 (36.19%) of all documents. A third-degree polynomic equation predicts that future global production in this area will be three-fold the current number, near 2031. Thematically, current research is focused on the integration of digital image analysis and genomic sequencing in Non-Hodgkin lymphomas, prediction of chemotherapy response and validation of new prognostic models. These findings can serve pathology departments to depict future clinical and research avenues, but also, public institutions and administrations to promote synergies and optimize funding allocation.