OncDRS: An integrative clinical and genomic data platform for enabling translational research and precision medicine

Extraction and Imputation of Eastern Cooperative Oncology Group Performance Status From Unstructured Oncology Notes Using Language Models

PURPOSE

Eastern Cooperative Oncology Group (ECOG) performance status (PS) is a key clinical variable for cancer treatment and research, but it is usually only recorded in unstructured form in the electronic health record. We investigated whether natural language processing (NLP) models can impute ECOG PS using unstructured note text.

MATERIALS AND METHODS

Medical oncology notes were identified from all patients with cancer at our center from 1997 to 2023 and divided at the patient level into training (approximately 80%), tuning/validation (approximately 10%), and test (approximately 10%) sets. Regular expressions were used to extract explicitly documented PS. Extracted PS labels were used to train NLP models to impute ECOG PS (0-1 v 2-4) from the remainder of the notes (with regular expression-extracted PS documentation removed). We assessed associations between imputed PS and overall survival (OS).

RESULTS

ECOG PS was extracted using regular expressions from 495,862 notes, corresponding to 79,698 patients. A Transformer-based Longformer model imputed PS with high discrimination (test set area under the receiver operating characteristic curve 0.95, area under the precision-recall curve 0.73). Imputed poor PS was associated with worse OS, including among notes with no explicit documentation of PS detected (OS hazard ratio, 11.9; 95% CI, 11.1 to 12.8).

CONCLUSION

NLP models can be used to impute performance status from unstructured oncologist notes at scale. This may aid the annotation of oncology data sets for clinical outcomes research and cancer care delivery.

Pilot trial of perampanel on peritumoral hyperexcitability and clinical outcomes in newly diagnosed high-grade glioma

Background

Glutamatergic neuron-glioma synaptogenesis and peritumoral hyperexcitability promote glioma growth in a positive feedback loop. The objective of this study was to evaluate the feasibility and estimated effect sizes of the AMPA-R antagonist, perampanel, on intraoperative electrophysiologic hyperexcitability and clinical outcomes.

Methods

An open-label trial was performed comparing perampanel to standard of care (SOC) in patients undergoing resection of newly-diagnosed radiologic high-grade glioma. Perampanel was administered as a pre-operative loading dose followed by maintenance therapy until progressive disease or up to 12-months. SOC treatment involved levetiracetam for 7-days or as clinically indicated. The primary outcome of hyperexcitability was defined by intra-operative electrocorticography high frequency oscillation (HFO) rates. Seizure-freedom and overall survival (OS) were estimated by the Kaplan-Meier method. Tissue concentrations of perampanel, levetiracetam, and metabolites were measured by mass spectrometry.

Results

HFO rates were similar between perampanel-treated and SOC cohorts. The trial was terminated early after interim analysis for futility, and outcomes assessed in 11 patients (7 perampanel-treated, 4 SOC). Over a median 281 days of post-enrollment follow-up, 27% of patients had seizures, including 14% treated with perampanel and 50% treated with SOC. OS in perampanel-treated patients was similar to a glioblastoma reference cohort (p=0.81). Glutamate concentrations in surface biopsies were positively correlated with HFO rates in adjacent electrode contacts and were not significantly associated with treatment assignment or drug concentrations.

Conclusions

A peri-operative loading regimen of perampanel was safe and well-tolerated, with similar peritumoral hyperexcitability as in levetiracetam-treated patients. Maintenance anti-glutamatergic therapy was not observed to impact survival outcomes.

Glioma genetic profiles associated with electrophysiologic hyperexcitability

BACKGROUND

Distinct genetic alterations determine glioma aggressiveness, however, the diversity of somatic mutations contributing to peritumoral hyperexcitability and seizures over the course of the disease is uncertain. This study aimed to identify tumor somatic mutation profiles associated with clinically significant hyperexcitability.

METHODS

A single center cohort of adults with WHO grades 1-4 glioma and targeted exome sequencing (n = 1716) was analyzed and cross-referenced with a validated EEG database to identify the subset of individuals who underwent continuous EEG monitoring (n = 206). Hyperexcitability was defined by the presence of lateralized periodic discharges and/or electrographic seizures. Cross-validated discriminant analysis models trained exclusively on recurrent somatic mutations were used to identify variants associated with hyperexcitability.

RESULTS

The distribution of WHO grades and tumor mutational burdens were similar between patients with and without hyperexcitability. Discriminant analysis models classified the presence or absence of EEG hyperexcitability with an overall accuracy of 70.9%, regardless of IDH1 R132H inclusion. Predictive variants included nonsense mutations in ATRX and TP53, indel mutations in RBBP8 and CREBBP, and nonsynonymous missense mutations with predicted damaging consequences in EGFR, KRAS, PIK3CA, TP53, and USP28. This profile improved estimates of hyperexcitability in a multivariate analysis controlling for age, sex, tumor location, integrated pathologic diagnosis, recurrence status, and preoperative epilepsy. Predicted somatic mutation variants were over-represented in patients with hyperexcitability compared to individuals without hyperexcitability and those who did not undergo continuous EEG.

CONCLUSION

These findings implicate diverse glioma somatic mutations in cancer genes associated with peritumoral hyperexcitability. Tumor genetic profiling may facilitate glioma-related epilepsy prognostication and management.

Glioma genetic profiles associated with electrophysiologic hyperexcitability

Distinct genetic alterations determine glioma aggressiveness, however the diversity of somatic mutations contributing to peritumoral hyperexcitability and seizures is uncertain. In a large cohort of patients with sequenced gliomas (n=1716), we used discriminant analysis models to identify somatic mutation variants associated with electrographic hyperexcitability in a subset with continuous EEG recording (n=206). Overall tumor mutational burdens were similar between patients with and without hyperexcitability. A cross-validated model trained exclusively on somatic mutations classified the presence or absence of hyperexcitability with an overall accuracy of 70.9%, and improved estimates of hyperexcitability and anti-seizure medication failure in multivariate analysis incorporating traditional demographic factors and tumor molecular classifications. Somatic mutation variants of interest were also over-represented in patients with hyperexcitability compared to internal and external reference cohorts. These findings implicate diverse mutations in cancer genes associated with the development of hyperexcitability and response to treatment.

Histopathologic and Genetic Features of Mismatch Repair-Deficient High-Grade Prostate Cancer

AIMS

Mismatch repair (MMR) deficiency is commonly caused by functional inactivation of MLH1, PMS2, MSH2 or MSH6. The morphologic and molecular correlates of MMR deficiency have been extensively characterized in certain tumor types such as colorectal and endometrial adenocarcinoma. In contrast, the histologic and molecular features of MMR-deficient prostate cancer remain incompletely described. In this study, we evaluated 19 MMR-deficient prostate cancers, including 11 cases without prior systemic treatment.

METHODS AND RESULTS

All treatment-naïve cases (11/11, 100%) were Grade Group 4-5 and had predominant cribriform and/or solid growth patterns. Solid components (any amount) and tumor infiltrating lymphocytes were seen in 7/11 (64%) of these cases each. In 68 MMR-proficient Grade Group 5 prostate cancers, predominant cribriform or solid growth patterns, solid components (any amount) and tumor infiltrating lymphocytes were seen at significantly lower frequencies (31/68, 46%; 9/68, 13% and 6/62, 9%, respectively; p<0.001 for all comparisons). Molecular evaluation of 19 cases demonstrated that MMR-deficiency was secondary to functional loss of MSH2/MSH6 and MLH1/PMS2 in 15 cases (79%) and 4 cases (21%), respectively. Definite or likely germline mutations were present in 4 cases (4/19, 21%). TMPRSS2::ERG rearrangements were identified in 2 cases (2/19, 11%). Recurrent cancer-relevant somatic mutations included (but were not limited to) ATM, TP53, FOXA1, RB1, BRCA2 and PTEN.

CONCLUSIONS

MMR deficiency was most commonly secondary to inactivation of MSH2/MSH6 in this study. Importantly, MMR-deficient high-grade prostatic adenocarcinomas had morphologic features that might be useful to identify selected cases for MMR IHC.

Real-World Analysis of Off-Label Use of Molecularly Targeted Therapy in a Large Academic Medical Center Cohort

PURPOSE

The primary objective of this study is to quantify the use of off-label molecularly targeted therapy and describe the clinical situations in which off-label targeted therapy are used. A key secondary objective is to report the outcomes of patients treated with off-label use of targeted therapy.

PATIENTS AND METHODS

We searched the electronic health record between 2000 and 2020 at our center to characterize the volume, clinical settings, and outcomes associated with off-label use of targeted therapies in different types of solid tumors.

RESULTS

Among 46,712 patients who received targeted therapies, we identified 119 instances of off-label use of targeted therapy. Colon cancer was the most common cancer type to receive off-label targeted therapy in 18 patients (15.1%), followed by 13 with non-small-cell lung cancer (10.9%), eight with cholangiocarcinoma (6.7%), and seven with glioblastoma (5.9%). The most frequent molecular rationale for off-label therapy came from a comprehensive next-generation sequencing test (53.7%). The most frequently mutated gene that provided the rationale for targeted therapy was BRAF (20.1%), with BRAFV600E being the most common molecular alteration overall (15.1%). The median duration of off-label targeted therapy was 3.58 months, and the overall survival of treated patients was 7.59 months. There were 37 patients (31.1%) treated for longer than 6 months, 23 patients (19.3%) who survived ≥ 2 years, and 13 patients who were still on therapy as of June 2020.

CONCLUSION

In this large cohort study of patients with solid tumors, off-label use of targeted therapy was uncommon. With that said, a notable proportion of patients had treatment durations ≥ 6 months and survivals of ≥ 2 years.

Artificial intelligence-aided clinical annotation of a large multi-cancer genomic dataset

To accelerate cancer research that correlates biomarkers with clinical endpoints, methods are needed to ascertain outcomes from electronic health records at scale. Here, we train deep natural language processing (NLP) models to extract outcomes for participants with any of 7 solid tumors in a precision oncology study. Outcomes are extracted from 305,151 imaging reports for 13,130 patients and 233,517 oncologist notes for 13,511 patients, including patients with 6 additional cancer types. NLP models recapitulate outcome annotation from these documents, including the presence of cancer, progression/worsening, response/improvement, and metastases, with excellent discrimination (AUROC > 0.90). Models generalize to cancers excluded from training and yield outcomes correlated with survival. Among patients receiving checkpoint inhibitors, we confirm that high tumor mutation burden is associated with superior progression-free survival ascertained using NLP. Here, we show that deep NLP can accelerate annotation of molecular cancer datasets with clinically meaningful endpoints to facilitate discovery.

Ibrutinib in Steroid-Refractory Chronic Graft-versus-Host Disease, a Single-Center Experience

Chronic graft-versus-host disease (cGVHD) is a leading cause of late morbidity and mortality after allogenic hematopoietic stem cell transplantation. Corticosteroid-based therapies are a mainstay of its initial treatment but there is no consensus in how to treat steroid-refractory cGVHD. Ibrutinib is a Bruton tyrosine kinase and IL-2-inducible kinase inhibitor thought to affect pathways driving cGVHD, and it was approved for the treatment of refractory cGVHD by the Food and Drug Administration (FDA) in August 2017 after a landmark phase 1b/2 study. It was the first medication approved for this indication, but how to best treat refractory cGVHD remains an open question, and there has been limited literature on ibrutinib after the FDA approval. This study sought to characterize the utilization and outcomes associated with ibrutinib use in cGVHD via a retrospective single-center study. Fifty-three patients were identified as having been treated with ibrutinib for cGVHD following FDA approval between September 1, 2017, and December 31, 2020, using an institutional data repository. Their records were reviewed for demographics, cGVHD characteristics, and outcomes. For the entire cohort, two-year overall survival was 76% (95% confidence interval [CI], 60% to 86%), with a median follow-up among survivors of 26 months (range, 1.3 to 39.5 months). However, the 2-year failure-free survival (FFS) after initiation of ibrutinib was 9% (95% CI, 2.6% to 20%), and the median FFS was 4.5 months (95% CI, 2.8 to 7.1 months). Events of FFS included treatment change due to lack of response or toxicity, malignant relapse, or non-treatment related mortality. At the time of this report, 11 patients (21%) remained on ibrutinib. At the time of the FFS event or last follow-up, 6 patients (12%) had a complete or partial response, 34 (64%) had stable disease, and 13 (25%) had progressive disease. Ibrutinib use was associated with no reduction in corticosteroid dose between ibrutinib initiation and FFS event or last follow-up (mean difference, 0.00; P = .98). The most frequently used noncorticosteroid cGVHD therapy after ibrutinib was ruxolitinib (n = 14; 33%). The most common adverse events associated with treatment discontinuation were infection (lung, skin, enterocolitis; n = 6), bleeding and bruising (hematoma, epistaxis, gastrointestinal bleed; n = 5), and muscle aches (n = 2). In a real-world setting, ibrutinib is associated with a modest response rate and FFS and its use in a narrower, more targeted patient population may be indicated.

Clinical Inflection Point Detection on the Basis of EHR Data to Identify Clinical Trial-Ready Patients With Cancer

PURPOSE

To inform precision oncology, methods are needed to use electronic health records (EHRs) to identify patients with cancer who are experiencing clinical inflection points, consistent with worsening prognosis or a high propensity to change treatment, at specific time points. Such patients might benefit from real-time screening for clinical trials.

METHODS

Using serial unstructured imaging reports for patients with solid tumors or lymphoma participating in a single-institution precision medicine study, we trained a deep neural network natural language processing (NLP) model to dynamically predict patients' prognoses and propensity to start new palliative-intent systemic therapy within 30 days. Model performance was evaluated using Harrell's c-index (for prognosis) and the area under the receiver operating characteristic curve (AUC; for new treatment and new clinical trial enrollment). Associations between model outputs and manual annotations of cancer progression were also evaluated using the AUC.

RESULTS

A deep NLP model was trained and evaluated using 302,688 imaging reports for 16,780 patients. In a held-out test set of 34,770 reports for 1,952 additional patients, the model predicted survival with a c-index of 0.76 and initiation of new treatment with an AUC of 0.77. Model-generated prognostic scores were associated with annotation of cancer progression on the basis of manual EHR review (n = 1,488 reports for 110 patients with lung or colorectal cancer) with an AUC of 0.78, and predictions of new treatment were associated with annotation of cancer progression on the basis of manual EHR review with an AUC of 0.84.

CONCLUSION

Training a deep NLP model to identify clinical inflection points among patients with cancer is feasible. This approach could identify patients who may benefit from real-time targeted clinical trial screening interventions at health system scale.

Natural Language Processing to Ascertain Cancer Outcomes From Medical Oncologist Notes

PURPOSE

Cancer research using electronic health records and genomic data sets requires clinical outcomes data, which may be recorded only in unstructured text by treating oncologists. Natural language processing (NLP) could substantially accelerate extraction of this information.

METHODS

Patients with lung cancer who had tumor sequencing as part of a single-institution precision oncology study from 2013 to 2018 were identified. Medical oncologists' progress notes for these patients were reviewed. For each note, curators recorded whether the assessment/plan indicated any cancer, progression/worsening of disease, and/or response to therapy or improving disease. Next, a recurrent neural network was trained using unlabeled notes to extract the assessment/plan from each note. Finally, convolutional neural networks were trained on labeled assessments/plans to predict the probability that each curated outcome was present. Model performance was evaluated using the area under the receiver operating characteristic curve (AUROC) among a held-out test set of 10% of patients. Associations between curated response or progression end points and overall survival were measured using Cox models among patients receiving palliative-intent systemic therapy.

RESULTS

Medical oncologist notes (n = 7,597) were manually curated for 919 patients. In the 10% test set, NLP models replicated human curation with AUROCs of 0.94 for the any-cancer outcome, 0.86 for the progression outcome, and 0.90 for the response outcome. Progression/worsening events identified using NLP models were associated with shortened survival (hazard ratio [HR] for mortality, 2.49; 95% CI, 2.00 to 3.09); response/improvement events were associated with improved survival (HR, 0.45; 95% CI, 0.30 to 0.67).

CONCLUSION

NLP models based on neural networks can extract meaningful outcomes from oncologist notes at scale. Such models may facilitate identification of clinical and genomic features associated with response to cancer treatment.

Enabling Precision Medicine in Cancer Care Through a Molecular Data Warehouse: The Moffitt Experience

PURPOSE

The use of genomics within cancer research and clinical oncology practice has become commonplace. Efforts such as The Cancer Genome Atlas have characterized the cancer genome and suggested a wealth of targets for implementing precision medicine strategies for patients with cancer. The data produced from research studies and clinical care have many potential secondary uses beyond their originally intended purpose. Effective storage, query, retrieval, and visualization of these data are essential to create an infrastructure to enable new discoveries in cancer research.

METHODS

Moffitt Cancer Center implemented a molecular data warehouse to complement the extensive enterprise clinical data warehouse (Health and Research Informatics). Seven different sequencing experiment types were included in the warehouse, with data from institutional research studies and clinical sequencing.

RESULTS

The implementation of the molecular warehouse involved the close collaboration of many teams with different expertise and a use case-focused approach. Cornerstones of project success included project planning, open communication, institutional buy-in, piloting the implementation, implementing custom solutions to address specific problems, data quality improvement, and data governance, unique aspects of which are featured here. We describe our experience in selecting, configuring, and loading molecular data into the molecular data warehouse. Specifically, we developed solutions for heterogeneous genomic sequencing cohorts (many different platforms) and integration with our existing clinical data warehouse.

CONCLUSION

The implementation was ultimately successful despite challenges encountered, many of which can be generalized to other research cancer centers.

Effect of PIK3CA variants on glioma-related epilepsy and response to treatment

Upregulation of the PI3K/AKT/mTOR pathway has been implicated in glioma-related epileptogenesis. In this retrospective analysis, epilepsy characteristics and response to treatment were evaluated in patients with gliomas harboring somatic mutation variants in PIK3CA. A cohort of 134 patients with 150 PIK3CA variants was extracted from previously validated databases. Patients with the hotspot H1047R, R88Q, E542K, and G118D variants comprised a subset (n = 41) for epilepsy phenotyping. In multivariate analysis, the presence of H1047R (n = 15) was associated with worse seizure control (p = 0.026). These results support preclinical findings and suggest that glioma PIK3CA variation may have promise as a biomarker for epilepsy severity and response to treatment.

Identification of a targetable KRAS-mutant epithelial population in non-small cell lung cancer

Lung cancer is the leading cause of cancer deaths. Tumor heterogeneity, which hampers development of targeted therapies, was herein deconvoluted via single cell RNA sequencing in aggressive human adenocarcinomas (carrying Kras-mutations) and comparable murine model. We identified a tumor-specific, mutant-KRAS-associated subpopulation which is conserved in both human and murine lung cancer. We previously reported a key role for the oncogene BMI-1 in adenocarcinomas. We therefore investigated the effects of in vivo PTC596 treatment, which affects BMI-1 activity, in our murine model. Post-treatment, MRI analysis showed decreased tumor size, while single cell transcriptomics concomitantly detected near complete ablation of the mutant-KRAS-associated subpopulation, signifying the presence of a pharmacologically targetable, tumor-associated subpopulation. Our findings therefore hold promise for the development of a targeted therapy for KRAS-mutant adenocarcinomas.

Characterization of genetics in patients with mucosal melanoma treated with immune checkpoint blockade

Mucosal melanoma is a rare form of melanoma which arises from melanocytes in the mucosal membranes and can be effectively treated with immune checkpoint blockade (ICB). However, response rates in mucosal melanoma are lower than those observed for cutaneous melanomas.

Targeted sequencing of up to 447 genes (OncoPanel) was performed on tumors from all mucosal melanoma patients seen at the Dana‐Farber Cancer Institute from 2011 until March 2019.

We identified a total of 46 patients who received ICB with both tumor‐genotype and ICB response data available. Within this cohort of patients, 16 (35%) had durable clinical benefit (DCB) to their first line of ICB. The average mutational burden/megabase was 6.23 and did not correlate with tumor response to ICB. Patients with KIT aberrations had a higher DCB rate compared with patients with wildtype KIT (71 vs. 28%), but this was not found to be statistically significant. For comparison, we analyzed tumor genotypes from an additional 50 mucosal melanoma tumors and 189 cutaneous melanoma tumors. The most frequent mutations in mucosal melanoma were in SF3B1 (27%), KIT (18%), and NF1 (17%), a pattern that is distinct from cutaneous melanomas. In addition, there were genetic differences observed based upon the site of origin of the mucosal melanoma.

Our findings explore clinical features of response in patients with mucosal melanoma treated with ICB and demonstrate a low mutational burden that does not correlate with response. In addition, the lack of significant association between the genetic aberrations tested and response to ICB indicates the need for further exploration in this patient population.

Large-scale analysis of BAP1 expression reveals novel associations with clinical and molecular features of malignant pleural mesothelioma

BRCA1‐associated protein‐1 (BAP1) expression is commonly lost in several tumors including malignant pleural mesothelioma (MPM). Presence or absence of immunohistochemical BAP1 nuclear staining in tumor cells is currently used for differential diagnosis of MPM. In this study, a large cohort of 596 MPM tumors with available clinical data was analyzed to examine associations of BAP1 staining pattern with clinical and molecular features that may reflect the impact of BAP1 mutation on MPM biology. Cases were classified according to the BAP1 staining pattern of tumor cells. Exome and RNA‐sequencing data were available for subsets of cases. Levels of mRNA encoding claudin 15 (CLDN15) and vimentin (VIM) were determined using RT‐qPCR on 483 cases to estimate the relative proportions of epithelial‐like and mesenchymal‐like components in each tumor. Four BAP1 staining patterns were observed: single‐pattern nuclear staining (36%), single‐pattern cytoplasmic staining (25%), single‐pattern absent staining (12%), and combinations of these staining patterns (27%). This study confirmed prior reports that nuclear BAP1 is more frequently associated with wild‐type BAP1 and sarcomatoid histology. However, no associations between BAP1 staining pattern(s) and mutations in specific protein domains and/or mutation type were observed. BAP1 staining patterns were significantly associated (p < 0.001) with BAP1 gene expression, MPM histologic subtypes, molecular clusters, and markers of epithelial‐to‐mesenchymal transition. Frequent observation of combinations of BAP1 staining patterns in MPM tumors indicated intra‐tumoral heterogeneity of BAP1 status. Cytoplasmic BAP1 staining was identified as a putative indicator of favorable prognosis in non‐epithelioid MPM. In conclusion, novel significant associations among different BAP1 staining patterns and subgroups of MPM tumors were observed, suggesting that the role of BAP1 in tumor progression may be more complex than its presumed tumor suppressor function. Cytoplasmic staining was identified as a putative indicator of favorable prognosis in non‐epithelioid MPM, potentially addressing a critical need in clinical decision‐making in this disease. © 2020 The Authors. The Journal of Pathology published by John Wiley & Sons, Ltd. on behalf of The Pathological Society of Great Britain and Ireland.

Precision medicine applications in prostate cancer

Aided by developments in diagnostics and therapeutics, healthcare is increasingly moving toward precision medicine, in which treatment is customized to each individual. We discuss the relevance of precision medicine in prostate cancer, including gene targets, therapeutics and resistance mechanisms. We foresee precision medicine becoming an integral component of prostate cancer management to increase response to therapy and prolong survival.

Disease Ontology: improving and unifying disease annotations across species

Model organisms are vital to uncovering the mechanisms of human disease and developing new therapeutic tools. Researchers collecting and integrating relevant model organism and/or human data often apply disparate terminologies (vocabularies and ontologies), making comparisons and inferences difficult. A unified disease ontology is required that connects data annotated using diverse disease terminologies, and in which the terminology relationships are continuously maintained. The Mouse Genome Database (MGD, http://www.informatics.jax.org), Rat Genome Database (RGD, http://rgd.mcw.edu) and Disease Ontology (DO, http://www.disease-ontology.org) projects are collaborating to augment DO, aligning and incorporating disease terms used by MGD and RGD, and improving DO as a tool for unifying disease annotations across species. Coordinated assessment of MGD's and RGD's disease term annotations identified new terms that enhance DO's representation of human diseases. Expansion of DO term content and cross-references to clinical vocabularies (e.g. OMIM, ORDO, MeSH) has enriched the DO's domain coverage and utility for annotating many types of data generated from experimental and clinical investigations. The extension of anatomy-based DO classification structure of disease improves accessibility of terms and facilitates application of DO for computational research. A consistent representation of disease associations across data types from cellular to whole organism, generated from clinical and model organism studies, will promote the integration, mining and comparative analysis of these data. The coordinated enrichment of the DO and adoption of DO by MGD and RGD demonstrates DO's usability across human data, MGD, RGD and the rest of the model organism database community.

Summary: Analyzing diverse disease data requires a comprehensive, robust disease ontology to integrate annotations and retrieve accurate, interpretable results. MGD, RGD and DO are working in collaboration to achieve this goal.

Somatic mutations in CDH1 and CTNNB1 in primary carcinomas at 13 anatomic sites

Metastases are involved in most cancer deaths. Evidence has suggested that cancer cell detachment from primary tumors might occur largely via the mechanism of epithelial-mesenchymal transition (EMT) activated by epigenetic events, but data addressing other possible triggers of detachment, particularly genetic mutations, have been limited. Using the Profile study of cancer genomics at Dana-Farber Cancer Institute, we examined somatic mutations in the EMT genes CDH1 in 5,106 primary carcinomas and CTNNB1 in 7,578 primary carcinomas across 13 anatomic sites: urinary bladder, breast, colon/rectum, endometrium, esophagus, kidney, lung, ovary, pancreas, prostate, skin (non-melanoma), stomach, and thyroid. For each gene and anatomic site, we calculated the prevalence of primary carcinomas with at least one mutation. Across all anatomic sites, 4% of carcinomas had at least one CDH1 mutation and 4% of carcinomas had at least one CTNNB1 mutation. By anatomic site, the observed prevalence of carcinomas with at least one mutation was less than 5% at 10 sites for CDH1 and 12 sites for CTNNB1. Tumor stage data were available for a subset of breast, colorectal, lung, and prostate tumors. Among patients from this subset who were diagnosed with regional or distant disease, only 4% had a CDH1 mutation and 1% had a CTNNB1 mutation in the primary tumor. The low mutation prevalences, especially among those with diagnoses of regional or distant disease, suggest that somatic mutations in CDH1 and CTNNB1 are unlikely to explain a substantial proportion of cancer cell detachment from primary carcinomas originating at most anatomic sites.

The Use of Translational Research Platforms in Clinical and Biomedical Data Exploration

The rise of precision medicine combined with the variety of biomedical data sources and their heterogeneous nature make the integration and exploration of information that they retain more complicated. In light of these issues, translational research platforms were developed as a promising solution. Research centers have used translational tools for the study of integrated data for hypothesis development and validation, cohort discovery and data-exploration. For this article, we reviewed the literature in order to determine the use of translational research platforms in precision medicine. These tools are used to support scientists in various domains regarding precision medicine research. We identified eight platforms: BRISK, iCOD, iDASH, tranSMART, the recently developed OncDRS, as well as caTRIP, cBio Cancer Portal and G-DOC. The last four platforms explore multidimensional data specifically for cancer research. We focused on tranSMART, for it is the most broadly used platform, since its development in 2012.