Development of a Structured Query Language and Natural Language Processing Algorithm to Identify Lung Nodules in a Cancer Centre

Importance: The stratification of indeterminate lung nodules is a growing problem, but the burden of lung nodules on healthcare services is not well-described. Manual service evaluation and research cohort curation can be time-consuming and potentially improved by automation.

Objective: To automate lung nodule identification in a tertiary cancer centre.

Methods: This retrospective cohort study used Electronic Healthcare Records to identify CT reports generated between 31st October 2011 and 24th July 2020. A structured query language/natural language processing tool was developed to classify reports according to lung nodule status. Performance was externally validated. Sentences were used to train machine-learning classifiers to predict concerning nodule features in 2,000 patients.

Results: 14,586 patients with lung nodules were identified. The cancer types most commonly associated with lung nodules were lung (39%), neuro-endocrine (38%), skin (35%), colorectal (33%) and sarcoma (33%). Lung nodule patients had a greater proportion of metastatic diagnoses (45 vs. 23%, p < 0.001), a higher mean post-baseline scan number (6.56 vs. 1.93, p < 0.001), and a shorter mean scan interval (4.1 vs. 5.9 months, p < 0.001) than those without nodules. Inter-observer agreement for sentence classification was 0.94 internally and 0.98 externally. Sensitivity and specificity for nodule identification were 93 and 99% internally, and 100 and 100% at external validation, respectively. A linear-support vector machine model predicted concerning sentence features with 94% accuracy.

Conclusion: We have developed and validated an accurate tool for automated lung nodule identification that is valuable for service evaluation and research data acquisition.

Personalized cancer medicine: molecular diagnostics, predictive biomarkers, and drug resistance

The progressive elucidation of the molecular pathogenesis of cancer has fueled the rational development of targeted drugs for patient populations stratified by genetic characteristics. Here we discuss general challenges relating to molecular diagnostics and describe predictive biomarkers for personalized cancer medicine. We also highlight resistance mechanisms for epidermal growth factor receptor (EGFR) kinase inhibitors in lung cancer. We envisage a future requiring the use of longitudinal genome sequencing and other omics technologies alongside combinatorial treatment to overcome cellular and molecular heterogeneity and prevent resistance caused by clonal evolution.

JARID2 is a direct target of the PAX3-FOXO1 fusion protein and inhibits myogenic differentiation of rhabdomyosarcoma cells

Rhabdomyosarcomas (RMS) are the most frequent soft-tissue sarcoma in children and characteristically show features of developing skeletal muscle. The alveolar subtype is frequently associated with a PAX3-FOXO1 fusion protein that is known to contribute to the undifferentiated myogenic phenotype of RMS cells. Histone methylation of lysine residues controls developmental processes in both normal and malignant cell contexts. Here we show that JARID2, which encodes a protein known to recruit various complexes with histone-methylating activity to their target genes, is significantly overexpressed in RMS with PAX3-FOXO1 compared with the fusion gene-negative RMS (t-test; P < 0.0001). Multivariate analyses showed that higher JARID2 levels are also associated with metastases at diagnosis, independent of fusion gene status and RMS subtype (n = 120; P = 0.039). JARID2 levels were altered by silencing or overexpressing PAX3-FOXO1 in RMS cell lines with and without the fusion gene, respectively. Consistent with this, we demonstrated that JARID2 is a direct transcriptional target of the PAX3-FOXO1 fusion protein. Silencing JARID2 resulted in reduced cell proliferation coupled with myogenic differentiation, including increased expression of Myogenin (MYOG) and Myosin Light Chain (MYL1) in RMS cell lines representative of both the alveolar and embryonal subtypes. Induced myogenic differentiation was associated with a decrease in JARID2 levels and this phenotype could be rescued by overexpressing JARID2. Furthermore, we that showed JARID2 binds to and alters the methylation status of histone H3 lysine 27 in the promoter regions of MYOG and MYL1 and that the interaction of JARID2 at these promoters is dependent on EED, a core component of the polycomb repressive complex 2 (PRC2). Therefore, JARID2 is a downstream effector of PAX3-FOXO1 that maintains an undifferentiated myogenic phenotype that is characteristic of RMS. JARID2 and other components of PRC2 may represent novel therapeutic targets for treating RMS patients.

Combining multiple structure and sequence alignments to improve sequence detection and alignment: application to the SH2 domains of Janus kinases

In this paper, an approach is described that combines multiple structure alignments and multiple sequence alignments to generate sequence profiles for protein families. First, multiple sequence alignments are generated from sequences that are closely related to each sequence of known three-dimensional structure. These alignments then are merged through a multiple structure alignment of family members of known structure. The merged alignment is used to generate a Hidden Markov Model for the family in question. The Hidden Markov Model can be used to search for new family members or to improve alignments for distantly related family members that already have been identified. Application of a profile generated for SH2 domains indicates that the Janus family of nonreceptor protein tyrosine kinases contains SH2 domains. This conclusion is strongly supported by the results of secondary structure-prediction programs, threading calculations, and the analysis of comparative models generated for these domains. One of the Janus kinases, human TYK2, has an SH2 domain that contains a histidine instead of the conserved arginine at the key phosphotyrosine-binding position, betaB5. Calculations of the pK(a) values of the betaB5 arginines in a number of SH2 domains and of the betaB5 histidine in a homology model of TYK2 suggest that this histidine is likely to be neutral around pH 7, thus indicating that it may have lost the ability to bind phosphotyrosine. If this indeed is the case, TYK2 may contain a domain with an SH2 fold that has a modified binding specificity.

Protein structure prediction

The prediction of protein structure, based primarily on sequence and structure homology, has become an increasingly important activity. Homology models have become more accurate and their range of applicability has increased. Progress has come, in part, from the flood of sequence and structure information that has appeared over the past few years, and also from improvements in analysis tools. These include profile methods for sequence searches, the use of three-dimensional structure information in sequence alignment and new homology modeling tools, specifically in the prediction of loop and side-chain conformations. There have also been important advances in understanding the physical chemical basis of protein stability and the corresponding use of physical chemical potential functions to identify correctly folded from incorrectly folded protein conformations.

Canonical structures for the hypervariable regions of T cell alphabeta receptors

T cell alphabeta receptors have binding sites for peptide-MHC complexes formed by six hypervariable regions. Analysis of the six atomic structures known for Valpha and for Vbeta domains shows that their first and second hypervariable regions have one of three or four different main-chain conformations (canonical structures). Six of these canonical structures have the same conformation in complexes with peptide-MHC complexes, the free receptor and/or in an isolated V domain. Thus, for at least the first and second hypervariable regions in the currently known structures, the conformation of the canonical structures is well defined in the free state and is conserved on formation of complexes with peptide-MHC. We identified the key residues that are mainly responsible for the conformation of each canonical structure. The first and second hypervariable regions of Valpha and Vbeta domains are encoded by the germline V segments. Humans have 37 functional Valpha segments and 47 Vbeta segments, and mice have 20 Vbeta segments. Inspection of the size of their hypervariable regions, and of sites that contain key residues, indicates that close to 70 % of Valpha segments and 90 % of Vbeta segments have hypervariable regions with a conformation of one of the known canonical structures. The alpha and beta V gene segments in both humans and mice have only a few combinations of different canonical structure in their first and second hypervariable regions. In human Vbeta domains, the number of different sequences with these canonical structure combinations is larger than in mice, whilst for Valpha domains it is probably smaller.

Standard conformations for the canonical structures of immunoglobulins

A comparative analysis of the main-chain conformation of the L1, L2, L3, H1 and H2 hypervariable regions in 17 immunoglobulin structures that have been accurately determined at high resolution is described. This involves 79 hypervariable regions in all. We also analysed a part of the H3 region in 12 of the 15 VH domains considered here. On the basis of the residues at key sites the 79 hypervariable regions can be assigned to one of 18 different canonical structures. We show that 71 of these hypervariable regions have a conformation that is very close to what can be defined as a "standard" conformation of each canonical structure. These standard conformations are described in detail. The other eight hypervariable regions have small deviations from the standard conformations that, in six cases, involve only the rotation of a single peptide group. Most H3 hypervariable regions have the same conformation in the part that is close to the framework and the details of this conformation are also described here.