Making sense of artificial intelligence and large language models-including ChatGPT-in pediatric hematology/oncology.

ChatGPT and other artificial intelligence (AI) systems have captivated the attention of healthcare providers and researchers for their potential to improve care processes and outcomes. While these technologies hold promise to automate processes, increase efficiency, and reduce cognitive burden, their use also carries risks. In this commentary, we review basic concepts of AI, outline some of the capabilities and limitations of currently available tools, discuss current and future applications in pediatric hematology/oncology, and provide an evaluation and implementation framework that can be used by pediatric hematologist/oncologists considering the use of AI in clinical practice.

Extracting Electronic Health Record Neuroblastoma Treatment Data With High Fidelity Using the REDCap Clinical Data Interoperability Services Module.

PURPOSE: Although the International Neuroblastoma Risk Group Data Commons (INRGdc) has enabled seminal large cohort studies, the research is limited by the lack of real-world, electronic health record (EHR) treatment data. To address this limitation, we evaluated the feasibility of extracting treatment data directly from EHRs using the REDCap Clinical Data Interoperability Services (CDIS) module for future submission to the INRGdc. METHODS: Patients enrolled on the Children's Oncology Group neuroblastoma biology study ANBL00B1 (ClinicalTrials.gov identifier: NCT00904241) who received care at the University of Chicago (UChicago) or the Vanderbilt University Medical Center (VUMC) after the go-live dates for the Fast Healthcare Interoperability Resources (FHIR)-compliant EHRs were identified. Antineoplastic drug orders were extracted using the CDIS module. To validate the CDIS output, antineoplastic agents extracted through FHIR were compared with those queried through EHR relational databases (UChicago's Clinical Research Data Warehouse and VUMC's Epic Clarity database) and manual chart review. RESULTS: The analytic cohort consisted of 41 patients at UChicago and 32 VUMC patients. Antineoplastic drug orders were identified in the extracted EHR records of 39 (95.1%) UChicago patients and 26 (81.3%) VUMC patients. Manual chart review confirmed that patients with missing (n = 8) or discontinued (n = 1) orders in the CDIS output did not receive antineoplastic agents during the timeframe of the study. More than 99% of the antineoplastic drug orders in the EHR relational databases were identified in the corresponding CDIS output. CONCLUSION: Our results demonstrate the feasibility of extracting EHR treatment data with high fidelity using HL7-FHIR via REDCap CDIS for future submission to the INRGdc.

Targeted Enrollment in Pediatric Oncology Trials: A Vision for Just-in-Time Matching.

@PedsDataCommons shares vision for automated clinical trials matching in pediatric oncology.

An open-source platform for pediatric cancer data exploration: a report from Data for the Common Good.

Objective: The Pediatric Cancer Data Commons (PCDC)-a project of Data for the Common Good-houses clinical pediatric oncology data and utilizes the open-source Gen3 platform. To meet the needs of end users, the PCDC development team expanded the out-of-box functionality and developed additional custom features that should be useful to any group developing similar data commons. Materials and Methods: Modifications of the PCDC data portal software were implemented to facilitate desired functionality. Results: Newly developed functionality includes updates to authorization methods, expansion of filtering capabilities, and addition of data analysis functions. Discussion: We describe the process by which custom functionalities were developed. Features are open source and available to be implemented and adapted to suit needs of data portals that utilize the Gen3 platform. Conclusion: Data portals are indispensable tools for facilitating data sharing. Open-source infrastructure facilitates a modular and collaborative approach for meeting needs of end users and stakeholders.

Accelerating pediatric Hodgkin lymphoma research: the Hodgkin Lymphoma Data Collaboration (NODAL).

Data commons have proven to be an indispensable avenue for advancing pediatric cancer research by serving as unified information technology platforms that, when coupled with data standards, facilitate data sharing. The Pediatric Cancer Data Commons, the flagship project of Data for the Common Good (D4CG), collaborates with disease-based consortia to facilitate development of clinical data standards, harmonization and pooling of clinical data from disparate sources, establishment of governance structure, and sharing of clinical data. In the interest of international collaboration, researchers developed the Hodgkin Lymphoma Data Collaboration and forged a relationship with the Pediatric Cancer Data Commons to establish a data commons for pediatric Hodgkin lymphoma. Herein, we describe the progress made in the formation of Hodgkin Lymphoma Data Collaboration and foundational goals to advance pediatric Hodgkin lymphoma research.

Automated Matching of Patients to Clinical Trials: A Patient-Centric Natural Language Processing Approach for Pediatric Leukemia.

PURPOSE: Matching patients to clinical trials is cumbersome and costly. Attempts have been made to automate the matching process; however, most have used a trial-centric approach, which focuses on a single trial. In this study, we developed a patient-centric matching tool that matches patient-specific demographic and clinical information with free-text clinical trial inclusion and exclusion criteria extracted using natural language processing to return a list of relevant clinical trials ordered by the patient's likelihood of eligibility. MATERIALS AND METHODS: Records from pediatric leukemia clinical trials were downloaded from ClinicalTrials.gov. Regular expressions were used to discretize and extract individual trial criteria. A multilabel support vector machine (SVM) was trained to classify sentence embeddings of criteria into relevant clinical categories. Labeled criteria were parsed using regular expressions to extract numbers, comparators, and relationships. In the validation phase, a patient-trial match score was generated for each trial and returned in the form of a ranked list for each patient. RESULTS: In total, 5,251 discretized criteria were extracted from 216 protocols. The most frequent criterion was previous chemotherapy/biologics (17%). The multilabel SVM demonstrated a pooled accuracy of 75%. The text processing pipeline was able to automatically extract 68% of eligibility criteria rules, as compared with 80% in a manual version of the tool. Automated matching was accomplished in approximately 4 seconds, as compared with several hours using manual derivation. CONCLUSION: To our knowledge, this project represents the first open-source attempt to generate a patient-centric clinical trial matching tool. The tool demonstrated acceptable performance when compared with a manual version, and it has potential to save time and money when matching patients to trials.

The Childhood Cancer Data Initiative: Using the Power of Data to Learn From and Improve Outcomes for Every Child and Young Adult With Pediatric Cancer.

Data-driven basic, translational, and clinical research has resulted in improved outcomes for children, adolescents, and young adults (AYAs) with pediatric cancers. However, challenges in sharing data between institutions, particularly in research, prevent addressing substantial unmet needs in children and AYA patients diagnosed with certain pediatric cancers. Systematically collecting and sharing data from every child and AYA can enable greater understanding of pediatric cancers, improve survivorship, and accelerate development of new and more effective therapies. To accomplish this goal, the Childhood Cancer Data Initiative (CCDI) was launched in 2019 at the National Cancer Institute. CCDI is a collaborative community endeavor supported by a 10-year, $50-million (in US dollars) annual federal investment. CCDI aims to learn from every patient diagnosed with a pediatric cancer by designing and building a data ecosystem that facilitates data collection, sharing, and analysis for researchers, clinicians, and patients across the cancer community. For example, CCDI's Molecular Characterization Initiative provides comprehensive clinical molecular characterization for children and AYAs with newly diagnosed cancers. Through these efforts, the CCDI strives to provide clinical benefit to patients and improvements in diagnosis and care through data-focused research support and to build expandable, sustainable data resources and workflows to advance research well past the planned 10 years of the initiative. Importantly, if CCDI demonstrates the success of this model for pediatric cancers, similar approaches can be applied to adults, transforming both clinical research and treatment to improve outcomes for all patients with cancer.

Persistence of Racial and Ethnic Disparities in Risk and Survival for Patients with Neuroblastoma over Two Decades.

BACKGROUND: Racial/ethnic survival disparities in neuroblastoma were first reported more than a decade ago. We sought to investigate if these disparities have persisted with current era therapy. METHODS: Two patient cohorts were identified in the International Neuroblastoma Risk Group Data Commons (INRGdc) (Cohort 1: diagnosed 2001-2009, n=4359; Cohort 2: diagnosed 2010-2019, n=4891). Chi-squared tests were used to assess the relationship between race/ethnicity and clinical and biologic features. Survival was estimated by the Kaplan-Meier method. Cox proportional hazards regression analyses were performed to investigate the association between racial/ethnic groups and prognostic markers. RESULTS: Significantly higher 5-year event-free survival (EFS) and overall survival (OS) were observed for Cohort 2 compared to Cohort 1 (P<0.001 and P<0.001, respectively). Compared to White patients, Black patients in both cohorts had a higher proportion of high-risk disease (Cohort 1: P<0.001; Cohort 2: P<0.001) and worse EFS (Cohort 1: P<0.001; Cohort 2 P<0.001) and OS (Cohort 1: P<0.001; Cohort 2: P<0.001). In Cohort 1, Native Americans also had a higher proportion of high-risk disease (P=0.03) and inferior EFS/OS. No significant survival disparities were observed for low- or intermediate-risk patients in either cohort or high-risk patients in Cohort 1. Hispanic patients with high-risk disease in Cohort 2 had significantly inferior OS (P=0.047). Significantly worse OS, but not EFS, (P=0.006 and P=0.02, respectively) was also observed among Black and Hispanic patients assigned to receive post-Consolidation dinutuximab on clinical trials (n=885). CONCLUSION: Racial/ethnic survival disparities have persisted over time and were observed among high-risk patients assigned to receive post-Consolidation dinutuximab.

Using A Standardized Nomenclature to Semantically Map Oncology-Related Concepts from Common Data Models to a Pediatric Cancer Data Model.

The Pediatric Cancer Data Commons (PCDC) comprises an international community whose ironclad commitment to data sharing is combatting pediatric cancer in an unprecedented way. The byproduct of their data sharing efforts is a gold-standard consensus data model covering many types of pediatric cancer. This article describes an effort to utilize SSSOM, an emerging specification for semantically-rich data mappings, to provide a "hub and spoke" model of mappings from several common data models (CDMs) to the PCDC data model. This provides important contributions to the research community, including: 1) a clear view of the current coverage of these CDMs in the domain of pediatric oncology, and 2) a demonstration of creating standardized mappings. These mappings can allow downstream crosswalk for data transformation and enhance data sharing. This can guide those who currently create and maintain brittle ad hoc data mappings in order to utilize the growing volume of viable research data.

Advancing clinical and translational research in germ cell tumours (GCT): recommendations from the Malignant Germ Cell International Consortium.

Germ cell tumours (GCTs) are a heterogeneous group of rare neoplasms that present in different anatomical sites and across a wide spectrum of patient ages from birth through to adulthood. Once these strata are applied, cohort numbers become modest, hindering inferences regarding management and therapeutic advances. Moreover, patients with GCTs are treated by different medical professionals including paediatric oncologists, neuro-oncologists, medical oncologists, neurosurgeons, gynaecological oncologists, surgeons, and urologists. Silos of care have thus formed, further hampering knowledge dissemination between specialists. Dedicated biobank specimen collection is therefore critical to foster continuous growth in our understanding of similarities and differences by age, gender, and site, particularly for rare cancers such as GCTs. Here, the Malignant Germ Cell International Consortium provides a framework to create a sustainable, global research infrastructure that facilitates acquisition of tissue and liquid biopsies together with matched clinical data sets that reflect the diversity of GCTs. Such an effort would create an invaluable repository of clinical and biological data which can underpin international collaborations that span professional boundaries, translate into clinical practice, and ultimately impact patient outcomes.

Creating a data commons: The INternational Soft Tissue SaRcoma ConsorTium (INSTRuCT).

In this article, we will discuss the genesis, evolution, and progress of the INternational Soft Tissue SaRcoma ConsorTium (INSTRuCT), which aims to foster international research and collaboration focused on pediatric soft tissue sarcoma. We will begin by highlighting the current state of clinical research for pediatric soft tissue sarcomas, including rhabdomyosarcoma and non-rhabdomyosarcoma soft tissue sarcoma. We will then explore challenges and research priorities, describe the development of INSTRuCT, and discuss how the consortium aims to address key research priorities.

Molecular profiling identifies targeted therapy opportunities in pediatric solid cancer.

To evaluate the clinical impact of molecular tumor profiling (MTP) with targeted sequencing panel tests, pediatric patients with extracranial solid tumors were enrolled in a prospective observational cohort study at 12 institutions. In the 345-patient analytical population, median age at diagnosis was 12 years (range 0-27.5); 298 patients (86%) had 1 or more alterations with potential for impact on care. Genomic alterations with diagnostic, prognostic or therapeutic significance were present in 61, 16 and 65% of patients, respectively. After return of the results, impact on care included 17 patients with a clarified diagnostic classification and 240 patients with an MTP result that could be used to select molecularly targeted therapy matched to identified alterations (MTT). Of the 29 patients who received MTT, 24% had an objective response or experienced durable clinical benefit; all but 1 of these patients received targeted therapy matched to a gene fusion. Of the diagnostic variants identified in 209 patients, 77% were gene fusions. MTP with targeted panel tests that includes fusion detection has a substantial clinical impact for young patients with solid tumors.

Mapping Pediatric Oncology Clinical Trial Collaborative Groups on the Global Stage.

PURPOSE: The global pediatric oncology clinical research landscape, particularly in Central and South America, Africa, and Asia, which bear the highest burden of global childhood cancer cases, is less characterized in the literature. Review of how existing pediatric cancer clinical trial groups internationally have been formed and how their research goals have been pursued is critical for building global collaborative research and data-sharing efforts, in line with the WHO Global Initiative for Childhood Cancer. METHODS: A narrative literature review of collaborative groups performing pediatric cancer clinical research in each continent was conducted. An inventory of research groups was assembled and reviewed by current pediatric cancer regional and continental leaders. Each group was narratively described with identification of common structural and research themes among consortia. RESULTS: There is wide variability in the structure, history, and goals of pediatric cancer clinical trial collaborative groups internationally. Several continental regions have longstanding endogenously-formed clinical trial groups that have developed and published numerous adapted treatment regimens to improve outcomes, whereas other regions have consortia focused on developing foundational database registry infrastructure supported by large multinational organizations or twinning relationships. CONCLUSION: There cannot be a one-size-fits-all approach to increasing collaboration between international pediatric cancer clinical trial groups, as this requires a nuanced understanding of local stakeholders and resources necessary to form partnerships. Needs assessments, performed either by local consortia or in conjunction with international partners, have generated productive clinical trial infrastructure. To achieve the goals of the Global Initiative for Childhood Cancer, global partnerships must be sufficiently granular to account for the distinct needs of each collaborating group and should incorporate grassroots approaches, robust twinning relationships, and implementation science.

Applications of Artificial Intelligence in Pediatric Oncology: A Systematic Review.

PURPOSE: There is a need for an improved understanding of clinical and biologic risk factors in pediatric cancer to improve patient outcomes. Machine learning (ML) represents the application of computational inference from advanced statistical methods that can be applied to increasing amount of data available for study in pediatric oncology. The goal of this systematic review was to systematically characterize the state of ML in pediatric oncology and highlight advances and opportunities in the field. METHODS: We conducted a systematic review of the Embase, Scopus, and MEDLINE databases for applications of ML in pediatric oncology. Query results from all three databases were aggregated and duplicate studies were removed. RESULTS: A total of 42 unique articles that examined the applications of ML in pediatric oncology met inclusion criteria for review. We identified 20 studies of CNS tumors, 13 of solid tumors, and nine of leukemia. ML tasks included classification, prediction of treatment response, and dose optimization with a variety of methods being used including neural network, k-nearest neighbor, random forest, naive Bayes, and support vector machines. Strengths of the identified studies included matching or outperforming physician comparators via automated analysis and predicting therapeutic response. Common limitations included significant heterogeneity in reporting standards, clinical applicability, small sample sizes, and missing external validation cohorts. CONCLUSION: We identified areas where ML can enhance clinical care in ways that may not otherwise be achievable. Although ML promises enormous potential in improving diagnostics, decision making, and monitoring for children with cancer, the field remains in early stages and future work will be aided by standards and guidelines to ensure rigorous methodologic design and maximizing clinical utility.

Predicting Response to Chemotherapy in Patients With Newly Diagnosed High-Risk Neuroblastoma: A Report From the International Neuroblastoma Risk Group.

PURPOSE: Metaiodobenzylguanidine (MIBG) scans are a radionucleotide imaging modality that undergo Curie scoring to semiquantitatively assess neuroblastoma burden, which can be used as a marker of therapy response. We hypothesized that a convolutional neural network (CNN) could be developed that uses diagnostic MIBG scans to predict response to induction chemotherapy. METHODS: We analyzed MIBG scans housed in the International Neuroblastoma Risk Group Data Commons from patients enrolled in the Children's Oncology Group high-risk neuroblastoma study ANBL12P1. The primary outcome was response to upfront chemotherapy, defined as a Curie score ≤ 2 after four cycles of induction chemotherapy. We derived and validated a CNN using two-dimensional whole-body MIBG scans from diagnosis and evaluated model performance using area under the receiver operating characteristic curve (AUC). We also developed a clinical classification model to predict response on the basis of age, stage, and MYCN amplification. RESULTS: Among 103 patients with high-risk neuroblastoma included in the final cohort, 67 (65%) were responders. Performance in predicting response to upfront chemotherapy was equivalent using the CNN and the clinical model. Class-activation heatmaps verified that the CNN used areas of disease within the MIBG scans to make predictions. Furthermore, integrating predictions using a geometric mean approach improved detection of responders to upfront chemotherapy (geometric mean AUC 0.73 v CNN AUC 0.63, P < .05; v clinical model AUC 0.65, P < .05). CONCLUSION: We demonstrate feasibility in using machine learning of diagnostic MIBG scans to predict response to induction chemotherapy for patients with high-risk neuroblastoma. We highlight improvements when clinical risk factors are also integrated, laying the foundation for using a multimodal approach to guiding treatment decisions for patients with high-risk neuroblastoma.

Germline Sequencing Improves Tumor-Only Sequencing Interpretation in a Precision Genomic Study of Patients With Pediatric Solid Tumor.

PURPOSE: Molecular tumor profiling is becoming a routine part of clinical cancer care, typically involving tumor-only panel testing without matched germline. We hypothesized that integrated germline sequencing could improve clinical interpretation and enhance the identification of germline variants with significant hereditary risks. MATERIALS AND METHODS: Tumors from pediatric patients with high-risk, extracranial solid malignancies were sequenced with a targeted panel of cancer-associated genes. Later, germline DNA was analyzed for a subset of these genes. We performed a post hoc analysis to identify how an integrated analysis of tumor and germline data would improve clinical interpretation. RESULTS: One hundred sixty participants with both tumor-only and germline sequencing reports were eligible for this analysis. Germline sequencing identified 38 pathogenic or likely pathogenic variants among 35 (22%) patients. Twenty-five (66%) of these were included in the tumor sequencing report. The remaining germline pathogenic or likely pathogenic variants were single-nucleotide variants filtered out of tumor-only analysis because of population frequency or copy-number variation masked by additional copy-number changes in the tumor. In tumor-only sequencing, 308 of 434 (71%) single-nucleotide variants reported were present in the germline, including 31% with suggested clinical utility. Finally, we provide further evidence that the variant allele fraction from tumor-only sequencing is insufficient to differentiate somatic from germline events. CONCLUSION: A paired approach to analyzing tumor and germline sequencing data would be expected to improve the efficiency and accuracy of distinguishing somatic mutations and germline variants, thereby facilitating the process of variant curation and therapeutic interpretation for somatic reports, as well as the identification of variants associated with germline cancer predisposition.

Pediatric Cancer Data Commons: Federating and Democratizing Data for Childhood Cancer Research.

The international pediatric oncology community has a long history of research collaboration. In the United States, the 2019 launch of the Children's Cancer Data Initiative puts the focus on developing a rich and robust data ecosystem for pediatric oncology. In this spirit, we present here our experience in constructing the Pediatric Cancer Data Commons (PCDC) to highlight the significance of this effort in fighting pediatric cancer and improving outcomes and to provide essential information to those creating resources in other disease areas. The University of Chicago's PCDC team has worked with the international research community since 2015 to build data commons for children's cancers. We identified six critical features of successful data commons design and implementation: (1) establish the need for a data commons, (2) develop and deploy the technical infrastructure, (3) establish and implement governance, (4) make the data commons platform easy and intuitive for researchers, (5) socialize the data commons and create working knowledge and expertise in the research community, and (6) plan for longevity and sustainability. Data commons are critical to conducting research on large patient cohorts that will ultimately lead to improved outcomes for children with cancer. There is value in connecting high-quality clinical and phenotype data to external sources of data such as genomic, proteomics, and imaging data. Next steps for the PCDC include creating an informed and invested data-sharing culture, developing sustainable methods of data collection and sharing, standardizing genetic biomarker reporting, incorporating radiologic and molecular analysis data, and building models for electronic patient consent. The methods and processes described here can be extended to any clinical area and provide a blueprint for others wishing to develop similar resources.

Creation of a pharmacogenomics patient portal complementary to an existing institutional provider-facing clinical decision support system.

BACKGROUND: Applied pharmacogenomics presents opportunities for improving patient care through precision medicine, particularly when paired with appropriate clinical decision support (CDS). However, a lack of patient resources for understanding pharmacogenomic test results may hinder shared decision-making and patient confidence in treatment. We sought to create a patient pharmacogenomics education and results delivery platform complementary to a CDS system to facilitate further research on the relevance of patient education to pharmacogenomics. METHODS: We conceptualized a model that extended the data access layer of an existing institutional CDS tool to allow for the pairing of decision supports offered to providers with patient-oriented summaries at the same level of phenotypic specificity. We built a two-part system consisting of a secure portal for patient use and an administrative dashboard for patient summary creation. The system was built in an ASP.NET and AngularJS architecture, and all data was housed in a HIPAA-compliant data center, with PHI secure in transit and at rest. RESULTS: The YourPGx Patient Portal was deployed on the institutional network in June 2019. Fifty-eight unique patient portal summaries have been written so far, which can provide over 4500 results modules to the pilot population of 544 patients. Patient behavior on the portal is being logged for further research. CONCLUSIONS: To our knowledge, this is the first automated system designed and deployed to provide detailed, personalized patient pharmacogenomics education complementary to a clinical decision support system. Future work will expand upon this system to allow for telemedicine and patient notification of new or updated results.

Cancer Informatics for Cancer Centers: Scientific Drivers for Informatics, Data Science, and Care in Pediatric, Adolescent, and Young Adult Cancer.

Cancer Informatics for Cancer Centers (CI4CC) is a grassroots, nonprofit 501c3 organization intended to provide a focused national forum for engagement of senior cancer informatics leaders, primarily aimed at academic cancer centers anywhere in the world but with a special emphasis on the 70 National Cancer Institute-funded cancer centers. This consortium has regularly held topic-focused biannual face-to-face symposiums. These meetings are a place to review cancer informatics and data science priorities and initiatives, providing a forum for discussion of the strategic and pragmatic issues that we faced at our respective institutions and cancer centers. Here, we provide meeting highlights from the latest CI4CC Symposium, which was delayed from its original April 2020 schedule because of the COVID-19 pandemic and held virtually over three days (September 24, October 1, and October 8) in the fall of 2020. In addition to the content presented, we found that holding this event virtually once a week for 6 hours was a great way to keep the kind of deep engagement that a face-to-face meeting engenders. This is the second such publication of CI4CC Symposium highlights, the first covering the meeting that took place in Napa, California, from October 14-16, 2019. We conclude with some thoughts about using data science to learn from every child with cancer, focusing on emerging activities of the National Cancer Institute's Childhood Cancer Data Initiative.