Comparative Analysis of Public Knowledge Bases for Precision Oncology

Evidence blocks for effective presentation of genomic findings at molecular tumor boards: Single institution experience

Genomic profiling, or molecular profiling of the tumor, is becoming a key component of therapeutic decision making in clinical oncology, and is typically carried out via next generation sequencing. However, the interpretation of the results and evaluation of rationale for targeting the uncovered alterations is challenging and requires a deep understanding of cancer biology, genetics, genomics and oncology. Multidisciplinary molecular tumor boards represent a promising strategy in the facilitation of molecularly-informed therapeutic decisions, and usually consist of specialists with various fields of expertise. To effectively communicate the biological and clinical significance of genomic findings, as well as to make molecular tumor board discussions more productive, we developed and implemented evidence blocks into case discussions in our center. We found that this approach facilitated clinicians' understanding of the results of genomic profiling, and resulted in shorter yet more efficient case discussions within the molecular tumor board. Here, we discuss our experience with evidence blocks and how their implementation influenced the molecular tumor board practice.

preon: Fast and accurate entity normalization for drug names and cancer types in precision oncology

MOTIVATION

In precision oncology (PO), clinicians aim to find the best treatment for any patient based on their molecular characterization. A major bottleneck is the manual annotation and evaluation of individual variants, for which usually a range of knowledge bases are screened. To incorporate and integrate the vast information of different databases, fast and accurate methods for harmonizing databases with different types of information are necessary. An essential step for harmonization in PO includes the normalization of tumor entities as well as therapy options for patients.

SUMMARY

preon is a fast and accurate library for the normalization of drug names and cancer types in large-scale data integration.

AVAILABILITY AND IMPLEMENTATION

preon is implemented in Python and freely available via the PyPI repository. Source code and the data underlying this article are available in GitHub at https://github.com/ermshaua/preon/.

The consequences of data dispersion in genomics: a comparative analysis of data sources for precision medicine

BACKGROUND

Genomics-based clinical diagnosis has emerged as a novel medical approach to improve diagnosis and treatment. However, advances in sequencing techniques have increased the generation of genomics data dramatically. This has led to several data management problems, one of which is data dispersion (i.e., genomics data is scattered across hundreds of data repositories). In this context, geneticists try to remediate the above-mentioned problem by limiting the scope of their work to a single data source they know and trust. This work has studied the consequences of focusing on a single data source rather than considering the many different existing genomics data sources.

METHODS

The analysis is based on the data associated with two groups of disorders (i.e., oncology and cardiology) accessible from six well-known genomic data sources (i.e., ClinVar, Ensembl, GWAS Catalog, LOVD, CIViC, and CardioDB). Two dimensions have been considered in this analysis, namely, completeness and concordance. Completeness has been evaluated at two levels. First, by analyzing the information provided by each data source with regard to a conceptual schema data model (i.e., the schema level). Second, by analyzing the DNA variations provided by each data source as related to any of the disorders selected (i.e., the data level). Concordance has been evaluated by comparing the consensus among the data sources regarding the clinical relevance of each variation and disorder.

RESULTS

The data sources with the highest completeness at the schema level are ClinVar, Ensembl, and CIViC. ClinVar has the highest completeness at the data level data source for the oncology and cardiology disorders. However, there are clinically relevant variations that are exclusive to other data sources, and they must be considered in order to provide the best clinical diagnosis. Although the information available in the data sources is predominantly concordant, discordance among the analyzed data exist. This can lead to inaccurate diagnoses.

CONCLUSION

Precision medicine analyses using a single genomics data source leads to incomplete results. Also, there are concordance problems that threaten the correctness of the genomics-based diagnosis results.

Leveraging Large Language Models for Decision Support in Personalized Oncology

Importance

Clinical interpretation of complex biomarkers for precision oncology currently requires manual investigations of previous studies and databases. Conversational large language models (LLMs) might be beneficial as automated tools for assisting clinical decision-making.

Objective

To assess performance and define their role using 4 recent LLMs as support tools for precision oncology.

Design, Setting, and Participants

This diagnostic study examined 10 fictional cases of patients with advanced cancer with genetic alterations. Each case was submitted to 4 different LLMs (ChatGPT, Galactica, Perplexity, and BioMedLM) and 1 expert physician to identify personalized treatment options in 2023. Treatment options were masked and presented to a molecular tumor board (MTB), whose members rated the likelihood of a treatment option coming from an LLM on a scale from 0 to 10 (0, extremely unlikely; 10, extremely likely) and decided whether the treatment option was clinically useful.

Main Outcomes and Measures

Number of treatment options, precision, recall, F1 score of LLMs compared with human experts, recognizability, and usefulness of recommendations.

Results

For 10 fictional cancer patients (4 with lung cancer, 6 with other; median [IQR] 3.5 [3.0-4.8] molecular alterations per patient), a median (IQR) number of 4.0 (4.0-4.0) compared with 3.0 (3.0-5.0), 7.5 (4.3-9.8), 11.5 (7.8-13.0), and 13.0 (11.3-21.5) treatment options each was identified by the human expert and 4 LLMs, respectively. When considering the expert as a criterion standard, LLM-proposed treatment options reached F1 scores of 0.04, 0.17, 0.14, and 0.19 across all patients combined. Combining treatment options from different LLMs allowed a precision of 0.29 and a recall of 0.29 for an F1 score of 0.29. LLM-generated treatment options were recognized as AI-generated with a median (IQR) 7.5 (5.3-9.0) points in contrast to 2.0 (1.0-3.0) points for manually annotated cases. A crucial reason for identifying AI-generated treatment options was insufficient accompanying evidence. For each patient, at least 1 LLM generated a treatment option that was considered helpful by MTB members. Two unique useful treatment options (including 1 unique treatment strategy) were identified only by LLM.

Conclusions and Relevance

In this diagnostic study, treatment options of LLMs in precision oncology did not reach the quality and credibility of human experts; however, they generated helpful ideas that might have complemented established procedures. Considering technological progress, LLMs could play an increasingly important role in assisting with screening and selecting relevant biomedical literature to support evidence-based, personalized treatment decisions.

Adrenergic and mesenchymal signatures are identifiable in cell-free DNA and correlate with metastatic disease burden in children with neuroblastoma

Background

Cell free DNA (cfDNA) profiles of 5-hydroxymethylcytosine (5-hmC), an epigenetic marker of open chromatin and active gene expression, are correlated with metastatic disease burden in patients with neuroblastoma. Neuroblastoma tumors are comprised of adrenergic (ADRN) and mesenchymal (MES) cells, and the relative abundance of each in tumor biopsies has prognostic implications. We hypothesized that ADRN and MES specific signatures could be quantified in cfDNA 5-hmC profiles and would augment the detection of metastatic burden in patients with neuroblastoma.

Methods

We previously performed an integrative analysis to identify ADRN and MES specific genes (n=373 and n=159, respectively). Purified DNA from cell lines was serial diluted with healthy donor cfDNA. Using Gene Set Variation Analysis (GSVA), ADRN and MES signatures were optimized. We then quantified signature scores, and our prior neuroblastoma signature, in cfDNA from 84 samples from 46 high-risk patients including 21 patients with serial samples.

Results

Samples from patients with higher metastatic burden had increased GSVA scores for both ADRN and MES gene signatures (p < 0.001). While ADRN and MES signature scores tracked together in serially collected samples, we identified instances of patients with increases in either MES or ADRN score at relapse.

Conclusions

While it is feasible to identify ADRN and MES signatures using 5-hmC profiles of cfDNA from neuroblastoma patients and correlate these signatures to metastatic burden, additional data are needed to determine the optimal strategies for clinical implementation. Prospective evaluation in larger cohorts is ongoing.

AIMedGraph: a comprehensive multi-relational knowledge graph for precision medicine

The development of high-throughput molecular testing techniques has enabled the large-scale exploration of the underlying molecular causes of diseases and the development of targeted treatment for specific genetic alterations. However, knowledge to interpret the impact of genetic variants on disease or treatment is distributed in different databases, scientific literature studies and clinical guidelines. AIMedGraph was designed to comprehensively collect and interrogate standardized information about genes, genetic alterations and their therapeutic and diagnostic relevance and build a multi-relational, evidence-based knowledge graph. Graph database Neo4j was used to represent precision medicine knowledge as nodes and edges in AIMedGraph. Entities in the current release include 30 340 diseases/phenotypes, 26 140 genes, 187 541 genetic variants, 2821 drugs, 15 125 clinical trials and 797 911 supporting literature studies. Edges in this release cover 621 731 drug interactions, 9279 drug susceptibility impacts, 6330 pharmacogenomics effects, 30 339 variant pathogenicity and 1485 drug adverse reactions. The knowledge graph technique enables hidden knowledge inference and provides insight into potential disease or drug molecular mechanisms. Database URL: http://aimedgraph.tongshugene.net:8201.

Feasibility and outcome of reproducible clinical interpretation of high-dimensional molecular data: a comparison of two molecular tumor boards

BACKGROUND

Structured and harmonized implementation of molecular tumor boards (MTB) for the clinical interpretation of molecular data presents a current challenge for precision oncology. Heterogeneity in the interpretation of molecular data was shown for patients even with a limited number of molecular alterations. Integration of high-dimensional molecular data, including RNA- (RNA-Seq) and whole-exome sequencing (WES), is expected to further complicate clinical application. To analyze challenges for MTB harmonization based on complex molecular datasets, we retrospectively compared clinical interpretation of WES and RNA-Seq data by two independent molecular tumor boards.

METHODS

High-dimensional molecular cancer profiling including WES and RNA-Seq was performed for patients with advanced solid tumors, no available standard therapy, ECOG performance status of 0-1, and available fresh-frozen tissue within the DKTK-MASTER Program from 2016 to 2018. Identical molecular profiling data of 40 patients were independently discussed by two molecular tumor boards (MTB) after prior annotation by specialized physicians, following independent, but similar workflows. Identified biomarkers and resulting treatment options were compared between the MTBs and patients were followed up clinically.

RESULTS

A median of 309 molecular aberrations from WES and RNA-Seq (n = 38) and 82 molecular aberrations from WES only (n = 3) were considered for clinical interpretation for 40 patients (one patient sequenced twice). A median of 3 and 2 targeted treatment options were identified per patient, respectively. Most treatment options were identified for receptor tyrosine kinase, PARP, and mTOR inhibitors, as well as immunotherapy. The mean overlap coefficient between both MTB was 66%. Highest agreement rates were observed with the interpretation of single nucleotide variants, clinical evidence levels 1 and 2, and monotherapy whereas the interpretation of gene expression changes, preclinical evidence levels 3 and 4, and combination therapy yielded lower agreement rates. Patients receiving treatment following concordant MTB recommendations had significantly longer overall survival than patients receiving treatment following discrepant recommendations or physician's choice.

CONCLUSIONS

Reproducible clinical interpretation of high-dimensional molecular data is feasible and agreement rates are encouraging, when compared to previous reports. The interpretation of molecular aberrations beyond single nucleotide variants and preclinically validated biomarkers as well as combination therapies were identified as additional difficulties for ongoing harmonization efforts.

From Information Overload to Actionable Insights: Digital Solutions for Interpreting Cancer Variants from Genomic Testing

Given the increase in genomic testing in routine clinical use, there is a growing need for digital technology solutions to assist pathologists, oncologists, and researchers in translating variant calls into actionable knowledge to personalize patient management plans. In this article, we discuss the challenges facing molecular geneticists and medical oncologists in working with test results from next-generation sequencing for somatic oncology, and propose key considerations for implementing a decision support software to aid the interpretation of clinically important variants. In addition, we review results from an example decision support software, NAVIFY Mutation Profiler. NAVIFY Mutation Profiler is a cloud-based software that provides curation, annotation, interpretation, and reporting of somatic variants identified by next-generation sequencing. The software reports a tiered classification based on consensus recommendations from AMP, ASCO, CAP, and ACMG. Studies with NAVIFY Mutation Profiler demonstrated that the software provided timely updates and accurate curation, as well as interpretation of variant combinations, demonstrating that decision support tools can help advance implementation of precision oncology.

Integrating molecular profiles into clinical frameworks through the Molecular Oncology Almanac to prospectively guide precision oncology

Tumor molecular profiling of single gene-variant ('first-order') genomic alterations informs potential therapeutic approaches. Interactions between such first-order events and global molecular features (for example, mutational signatures) are increasingly associated with clinical outcomes, but these 'second-order' alterations are not yet accounted for in clinical interpretation algorithms and knowledge bases. We introduce the Molecular Oncology Almanac (MOAlmanac), a paired clinical interpretation algorithm and knowledge base to enable integrative interpretation of multimodal genomic data for point-of-care decision making and translational-hypothesis generation. We benchmarked MOAlmanac to a first-order interpretation method across multiple retrospective cohorts and observed an increased number of clinical hypotheses from evaluation of molecular features and profile-to-cell line matchmaking. When applied to a prospective precision oncology trial cohort, MOAlmanac nominated a median of two therapies per patient and identified therapeutic strategies administered in 47% of patients. Overall, we present an open-source computational method for integrative clinical interpretation of individualized molecular profiles.

[Variant interpretation in molecular pathology and oncology : An introduction]

Increasingly extensive genomic diagnostics in cancer precision medicine require uniform evaluation criteria for the classification of variants with regard to their functional and therapeutic implications. In this review we present the most important guidelines and classification systems currently used in daily clinical practice, explain their advantages and disadvantages as well as differences and similarities, and present the step-by-step, systematic process that enables successful variant interpretation.

Targeted Therapies in Advanced Biliary Tract Cancer: An Evolving Paradigm

Biliary tract cancers (BTCs) are a heterogeneous group of adenocarcinomas that originate from the epithelial lining of the biliary tree. BTCs are characterized by presentation with advanced disease precluding curative surgery, rising global incidence, and a poor prognosis. Chemotherapy is the mainstay of the current treatment, which results in a median overall survival of less than one year, underscoring the need for novel therapeutic agents and strategies. Next-generation sequencing-based molecular profiling has shed light on the underpinnings of the complex pathophysiology of BTC and has uncovered numerous actionable targets, leading to the discovery of new therapies tailored to the molecular targets. Therapies targeting fibroblast growth factor receptor (FGFR) fusion, isocitrate dehydrogenase (IDH) mutations, the human epidermal growth factor receptor (HER) family, DNA damage repair (DDR) pathways, and BRAF mutations have produced early encouraging results in selected patients. Current clinical trials evaluating targeted therapies, as monotherapies and in combination with other agents, are paving the way for novel treatment options. Genomic profiling of cell-free circulating tumor DNA that can assist in the identification of an actionable target is another exciting area of development. In this review, we provide a contemporaneous appraisal of the evolving targeted therapies and the ongoing clinical trials that will likely transform the therapeutic paradigm of BTC.

[RAS mutations at the molecular tumor conference]

Members of the rat sarcoma (RAS) gene family belong to the most frequently mutated genes that drive pathogenesis and therapy response. As the discovery of their malignant potential dates back more than three decades, cellular mutated RAS genes and their products belong to the best characterized cancer genes. Despite urgent clinical needs, RAS therapies are still elusive and limited to preclinical studies. However, very recently, novel and promising approaches have become a reality in clinical applications and trials. In the near future, interesting therapeutic options will emerge that are capable of targeting "undruggable" RAS. This will be even more important as the detection of RAS mutations has already been an integral part of routine molecular diagnostics for many years.