Generalization optimizing machine learning to improve CT scan radiomics and assess immune checkpoint inhibitors' response in non-small cell lung cancer: a multicenter cohort study.

Background: Recent developments in artificial intelligence suggest that radiomics may represent a promising non-invasive biomarker to predict response to immune checkpoint inhibitors (ICIs). Nevertheless, validation of radiomics algorithms in independent cohorts remains a challenge due to variations in image acquisition and reconstruction. Using radiomics, we investigated the importance of scan normalization as part of a broader machine learning framework to enable model external generalizability to predict ICI response in non-small cell lung cancer (NSCLC) patients across different centers. Methods: Radiomics features were extracted and compared from 642 advanced NSCLC patients on pre-ICI scans using established open-source PyRadiomics and a proprietary DeepRadiomics deep learning technology. The population was separated into two groups: a discovery cohort of 512 NSCLC patients from three academic centers and a validation cohort that included 130 NSCLC patients from a fourth center. We harmonized images to account for variations in reconstruction kernel, slice thicknesses, and device manufacturers. Multivariable models, evaluated using cross-validation, were used to estimate the predictive value of clinical variables, PD-L1 expression, and PyRadiomics or DeepRadiomics for progression-free survival at 6 months (PFS-6). Results: The best prognostic factor for PFS-6, excluding radiomics features, was obtained with the combination of Clinical + PD-L1 expression (AUC = 0.66 in the discovery and 0.62 in the validation cohort). Without image harmonization, combining Clinical + PyRadiomics or DeepRadiomics delivered an AUC = 0.69 and 0.69, respectively, in the discovery cohort, but dropped to 0.57 and 0.52, in the validation cohort. This lack of generalizability was consistent with observations in principal component analysis clustered by CT scan parameters. Subsequently, image harmonization eliminated these clusters. The combination of Clinical + DeepRadiomics reached an AUC = 0.67 and 0.63 in the discovery and validation cohort, respectively. Conversely, the combination of Clinical + PyRadiomics failed generalizability validations, with AUC = 0.66 and 0.59. Conclusion: We demonstrated that a risk prediction model combining Clinical + DeepRadiomics was generalizable following CT scan harmonization and machine learning generalization methods. These results had similar performances to routine oncology practice using Clinical + PD-L1. This study supports the strong potential of radiomics as a future non-invasive strategy to predict ICI response in advanced NSCLC.

Variants at the APOE /C1/C2/C4 Locus Modulate Cholesterol Efflux Capacity Independently of High-Density Lipoprotein Cholesterol.

Background Macrophage cholesterol efflux to high-density lipoproteins ( HDLs ) is the first step of reverse cholesterol transport. The cholesterol efflux capacity ( CEC ) of HDL particles is a protective risk factor for coronary artery disease independent of HDL cholesterol levels. Using a genome-wide association study approach, we aimed to identify pathways that regulate CEC in humans. Methods and Results We measured CEC in 5293 French Canadians. We tested the genetic association between 4 CEC measures and genotypes at >9 million common autosomal DNA sequence variants. These analyses yielded 10 genome-wide significant signals ( P<6.25×10-9) representing 7 loci. Five of these loci harbor genes with important roles in lipid biology ( CETP , LIPC , LPL , APOA 1/C3/A4/A5, and APOE /C1/C2/C4). Except for the APOE /C1/C2/C4 variant ( rs141622900, P nonadjusted=1.0×10-11; P adjusted=8.8×10-9), the association signals disappear when correcting for HDL cholesterol and triglyceride levels. The additional 2 significant signals were near the PPP 1 CB / PLB 1 and RBFOX 3/ ENPP 7 genes. In secondary analyses, we considered candidate functional variants for 58 genes implicated in HDL biology, as well as 239 variants associated with blood lipid levels and/or coronary artery disease risk by genome-wide association study . These analyses identified 27 significant CEC associations, implicating 5 additional loci ( GCKR , LIPG , PLTP , PPARA , and TRIB 1). Conclusions Our genome-wide association study identified common genetic variation at the APOE /C1/C2/C4 locus as a major determinant of CEC that acts largely independently of HDL cholesterol. We predict that HDL -based therapies aiming at increasing CEC will be modulated by changes in the expression of apolipoproteins in this gene cluster.

Polygenic determinants in extremes of high-density lipoprotein cholesterol.

HDL cholesterol (HDL-C) remains a superior biochemical predictor of CVD risk, but its genetic basis is incompletely defined. In patients with extreme HDL-C concentrations, we concurrently evaluated the contributions of multiple large- and small-effect genetic variants. In a discovery cohort of 255 unrelated lipid clinic patients with extreme HDL-C levels, we used a targeted next-generation sequencing panel to evaluate rare variants in known HDL metabolism genes, simultaneously with common variants bundled into a polygenic trait score. Two additional cohorts were used for validation and included 1,746 individuals from the Montréal Heart Institute Biobank and 1,048 individuals from the University of Pennsylvania. Findings were consistent between cohorts: we found rare heterozygous large-effect variants in 18.7% and 10.9% of low- and high-HDL-C patients, respectively. We also found common variant accumulation, indicated by extreme polygenic trait scores, in an additional 12.8% and 19.3% of overall cases of low- and high-HDL-C extremes, respectively. Thus, the genetic basis of extreme HDL-C concentrations encountered clinically is frequently polygenic, with contributions from both rare large-effect and common small-effect variants. Multiple types of genetic variants should be considered as contributing factors in patients with extreme dyslipidemia.

Polygenic Versus Monogenic Causes of Hypercholesterolemia Ascertained Clinically.

OBJECTIVE: Next-generation sequencing technology is transforming our understanding of heterozygous familial hypercholesterolemia, including revision of prevalence estimates and attribution of polygenic effects. Here, we examined the contributions of monogenic and polygenic factors in patients with severe hypercholesterolemia referred to a specialty clinic. APPROACH AND RESULTS: We applied targeted next-generation sequencing with custom annotation, coupled with evaluation of large-scale copy number variation and polygenic scores for raised low-density lipoprotein cholesterol in a cohort of 313 individuals with severe hypercholesterolemia, defined as low-density lipoprotein cholesterol >5.0 mmol/L (>194 mg/dL). We found that (1) monogenic familial hypercholesterolemia-causing mutations detected by targeted next-generation sequencing were present in 47.3% of individuals; (2) the percentage of individuals with monogenic mutations increased to 53.7% when copy number variations were included; (3) the percentage further increased to 67.1% when individuals with extreme polygenic scores were included; and (4) the percentage of individuals with an identified genetic component increased from 57.0% to 92.0% as low-density lipoprotein cholesterol level increased from 5.0 to >8.0 mmol/L (194 to >310 mg/dL). CONCLUSIONS: In a clinically ascertained sample with severe hypercholesterolemia, we found that most patients had a discrete genetic basis detected using a comprehensive screening approach that includes targeted next-generation sequencing, an assay for copy number variations, and polygenic trait scores.

Whole-genome sequencing in French Canadians from Quebec.

Genome-wide association studies (GWAS) have had a tremendous success in the identification of common DNA sequence variants associated with complex human diseases and traits. However, because of their design, GWAS are largely inappropriate to characterize the role of rare and low-frequency DNA variants on human phenotypic variation. Rarer genetic variation is geographically more restricted, supporting the need for local whole-genome sequencing (WGS) efforts to study these variants in specific populations. Here, we present the first large-scale low-pass WGS of the French-Canadian population. Specifically, we sequenced at ~5.6× coverage the whole genome of 1970 French Canadians recruited by the Montreal Heart Institute Biobank and identified 29 million bi-allelic variants (31 % novel), including 19 million variants with a minor allele frequency (MAF) <0.5 %. Genotypes from the WGS data are highly concordant with genotypes obtained by exome array on the same individuals (99.8 %), even when restricting this analysis to rare variants (MAF <0.5, 99.9 %) or heterozygous sites (98.9 %). To further validate our data set, we showed that we can effectively use it to replicate several genetic associations with myocardial infarction risk and blood lipid levels. Furthermore, we analyze the utility of our WGS data set to generate a French-Canadian-specific imputation reference panel and to infer population structure in the Province of Quebec. Our results illustrate the value of low-pass WGS to study the genetics of human diseases in the founder French-Canadian population.

Testing the role of predicted gene knockouts in human anthropometric trait variation.

Although the role of complete gene inactivation by two loss-of-function mutations inherited in trans is well-established in recessive Mendelian diseases, we have not yet explored how such gene knockouts (KOs) could influence complex human phenotypes. Here, we developed a statistical framework to test the association between gene KOs and quantitative human traits. Our method is flexible, publicly available, and compatible with common genotype format files (e.g. PLINK and vcf). We characterized gene KOs in 4498 participants from the NHLBI Exome Sequence Project (ESP) sequenced at high coverage (>100×), 1976 French Canadians from the Montreal Heart Institute Biobank sequenced at low coverage (5.7×), and >100 000 participants from the Genetic Investigation of ANthropometric Traits (GIANT) Consortium genotyped on an exome array. We tested associations between gene KOs and three anthropometric traits: body mass index (BMI), height and BMI-adjusted waist-to-hip ratio (WHR). Despite our large sample size and multiple datasets available, we could not detect robust associations between specific gene KOs and quantitative anthropometric traits. Our results highlight several limitations and challenges for future gene KO studies in humans, in particular when there is no prior knowledge on the phenotypes that might be affected by the tested gene KOs. They also suggest that gene KOs identified with current DNA sequencing methodologies probably do not strongly influence normal variation in BMI, height, and WHR in the general human population.

Toxicity of metal oxide nanoparticles in Escherichia coli correlates with conduction band and hydration energies.

Metal oxide nanoparticles (MOx NPs) are used for a host of applications, such as electronics, cosmetics, construction, and medicine, and as a result, the safety of these materials to humans and the environment is of considerable interest. A prior study of 24 MOx NPs in mammalian cells revealed that some of these materials show hazard potential. Here, we report the growth inhibitory effects of the same series of MOx NPs in the bacterium Escherichia coli and show that toxicity trends observed in E. coli parallel those seen previously in mammalian cells. Of the 24 materials studied, only ZnO, CuO, CoO, Mn2O3, Co3O4, Ni2O3, and Cr2O3 were found to exert significant growth inhibitory effects; these effects were found to relate to membrane damage and oxidative stress responses in minimal trophic media. A correlation of the toxicological data with physicochemical parameters of MOx NPs revealed that the probability of a MOx NP being toxic increases as the hydration enthalpy becomes less negative and as the conduction band energy approaches those of biological molecules. These observations are consistent with prior results observed in mammalian cells, revealing that mechanisms of toxicity of MOx NPs are consistent across two very different taxa. These results suggest that studying nanotoxicity in E. coli may help to predict toxicity patterns in higher organisms.

Zebrafish high-throughput screening to study the impact of dissolvable metal oxide nanoparticles on the hatching enzyme, ZHE1.

The zebrafish is emerging as a model organism for the safety assessment and hazard ranking of engineered nanomaterials. In this Communication, the implementation of a roboticized high-throughput screening (HTS) platform with automated image analysis is demonstrated to assess the impact of dissolvable oxide nanoparticles on embryo hatching. It is further demonstrated that this hatching interference is mechanistically linked to an effect on the metalloprotease, ZHE 1, which is responsible for degradation of the chorionic membrane. The data indicate that 4 of 24 metal oxide nanoparticles (CuO, ZnO, Cr2 O3 , and NiO) could interfere with embryo hatching by a chelator-sensitive mechanism that involves ligation of critical histidines in the ZHE1 center by the shed metal ions. A recombinant ZHE1 enzymatic assay is established to demonstrate that the dialysates from the same materials responsible for hatching interference also inhibit ZHE1 activity in a dose-dependent fashion. A peptide-based BLAST search identifies several additional aquatic species that express enzymes with homologous histidine-based catalytic centers, suggesting that the ZHE1 mechanistic paradigm could be used to predict the toxicity of a large number of oxide nanoparticles that pose a hazard to aquatic species.

Use of metal oxide nanoparticle band gap to develop a predictive paradigm for oxidative stress and acute pulmonary inflammation.

We demonstrate for 24 metal oxide (MOx) nanoparticles that it is possible to use conduction band energy levels to delineate their toxicological potential at cellular and whole animal levels. Among the materials, the overlap of conduction band energy (E(c)) levels with the cellular redox potential (-4.12 to -4.84 eV) was strongly correlated to the ability of Co(3)O(4), Cr(2)O(3), Ni(2)O(3), Mn(2)O(3), and CoO nanoparticles to induce oxygen radicals, oxidative stress, and inflammation. This outcome is premised on permissible electron transfers from the biological redox couples that maintain the cellular redox equilibrium to the conduction band of the semiconductor particles. Both single-parameter cytotoxic as well as multi-parameter oxidative stress assays in cells showed excellent correlation to the generation of acute neutrophilic inflammation and cytokine responses in the lungs of C57 BL/6 mice. Co(3)O(4), Ni(2)O(3), Mn(2)O(3), and CoO nanoparticles could also oxidize cytochrome c as a representative redox couple involved in redox homeostasis. While CuO and ZnO generated oxidative stress and acute pulmonary inflammation that is not predicted by E(c) levels, the adverse biological effects of these materials could be explained by their solubility, as demonstrated by ICP-MS analysis. These results demonstrate that it is possible to predict the toxicity of a large series of MOx nanoparticles in the lung premised on semiconductor properties and an integrated in vitro/in vivo hazard ranking model premised on oxidative stress. This establishes a robust platform for modeling of MOx structure-activity relationships based on band gap energy levels and particle dissolution. This predictive toxicological paradigm is also of considerable importance for regulatory decision-making about this important class of engineered nanomaterials.