Search

PURPOSE The ability of next-generation sequencing (NGS) assays to interrogate thousands of genomic loci has revolutionized genetic testing. However, translation to the clinic is impeded by false-negative results that pose a risk to patients. In response, regulatory bodies are calling for reliability measures to be reported alongside NGS results. Existing methods to estimate reliability do not account for sample- and position-specific variability, which can be significant. Here, we report an approach that computes reliability metrics for every genomic position and sample interrogated by an NGS assay. METHODS Our approach predicts the limit of detection (LOD), the lowest reliably detectable variant fraction, by taking technical factors into account. We initially explored how LOD is affected by input material amount, library conversion rate, sequencing coverage, and sequencing error rate. This revealed that LOD depends heavily on genomic context and sample properties. Using these insights, we developed a computational approach to predict LOD on the basis of a biophysical model of the NGS workflow. We focused on targeted assays for cell-free DNA, but, in principle, this approach applies to any NGS assay. RESULTS We validated our approach by showing that it accurately predicts LOD and distinguishes reliable from unreliable results when screening 580 lung cancer samples for actionable mutations. Compared with a standard variant calling workflow, our approach avoided most false negatives and improved interassay concordance from 94% to 99%. CONCLUSION Our approach, which we name LAVA (LOD-aware variant analysis), reports the LOD for every position and sample interrogated by an NGS assay. This enables reliable results to be identified and improves the transparency and safety of genetic tests.

8542 Background: The NADIM trial (NCT03081689), led by the Spanish Lung Cancer Group, assessed the antitumor activity and safety of neoadjuvant chemoimmunotherapy for resectable stage IIIA NSCLC. Patients received neoadjuvant nivolumab and paclitaxel-carboplatin for three cycles before surgical resection, followed by one year of adjuvant nivolumab. At 24 months, progression-free survival (PFS) was 77%, suggesting that neoadjuvant chemoimmunotherapy represents a promising option in this setting. Pathological complete response (pCR) could potentially be used as an important surrogate endpoint for survival. We present here a re-analysis of the NADIM cohort aiming to develop a machine learning algorithm to predict the pCR status based on multimodal baseline data. Methods: We combined baseline clinical data (e.g., age, smoking status), biological data (e.g., tumor histology, mutations), radiology reports and radiomics analysis of the baseline CT scan in a multimodal analysis. While 46 patients were enrolled in the NADIM trial, only 28 had a complete set of data available for this retrospective study. For each patient, tumors were segmented on the baseline CT-scan in 3D by a Deep Learning algorithm. Radiomics features were extracted following the IBSI standards and combined with the other data modalities. A filter-based variable selection method was applied before training several machine learning algorithms. The optimization criterion was the Area Under the ROC Curve (AUC). Due to the small size of the cohort, a leave-one-out cross-validation approach was used to properly estimate the model performance. For a sub-cohort of 20 patients for which data have been collected longitudinally during the neoadjuvant treatment, an additional Delta-radiomics model was used to predict the pCR status. Results: An XGBoost algorithm with a linear base learner displayed an AUC of 0.69, a precision of 75%, a sensitivity of 83% and a specificity of 50%. Features with highest weight in the algorithm were a mix of radiological, radiomics, biological and clinical features (including the neutrophils to lymphocytes ratio, mutations and histology) highlighting the importance of a truly multimodal analysis. Indeed, withdrawing a specific data modality (e.g., radiomics or biological features), led to a decrease of ̃15% of the AUC. Inclusion of the Delta-radiomics analysis on the data collected longitudinally prior to surgery led to an improved AUC of 0.76 in that patient sub-cohort. Conclusions: This study is, to our knowledge, the first to offer a multimodal analysis of the response to neoadjuvant treatment for surgically resectable stage IIIA NSCLC and is a proof of concept that a machine learning algorithm can be used to predict the pCR in this context. These preliminary results are being confirmed in the ongoing NADIM II trial. Clinical trial information: NCT03838159.

601 Background: Triple negative breast cancer (TNBC) is a biologically and clinically heterogenous disease, associated with poorer outcomes when compared with other subtypes of breast cancer. In early-stage TNBC, surgery with curative intent remains the mainstay of therapy. Neoadjuvant chemotherapy is often given prior to surgery and achieving pathological complete response (pCR) has been associated with improved long-term outcomes in terms of progression-free survival (PFS) and overall survival (OS). There is thus high clinical interest in the ability to accurately predict pCR status using baseline data. Methods: A retrospective cohort of 57 patients with early-stage TNBC treated with neoadjuvant chemotherapy was analyzed to develop a machine learning-based algorithm predictive of pCR likelihood at the individual patient level. Multimodal baseline data were collected including clinical, biological (e.g., histology, genomic profile including CDC2, CDC20, KPNA2, MYBL2, complete blood count, Ki67), imaging data (baseline PET/CT scan), the radiology report and clinical outcomes data (pCR, PFS, OS). For each patient, tumors were segmented in 3D by an experimented nuclear physician using the SOPHiA Radiomics platform. Radiomics features were then extracted following the IBSI standards and then combined with the other data modalities. A filter-based variable selection method was applied before training several machine learning algorithms. The optimization criterion was the Area Under the ROC Curve (AUC). Due to the small size of the cohort, a nested leave-pair-out cross-validation approach was used to properly estimate the model performance. Results: The best result was obtained with the SVM algorithm with a linear kernel, reaching an AUC of 0.82, a precision of 71%, a sensitivity of 71% and a specificity of 70%. The features with highest weight in the algorithm were a mix of radiological, radiomics, biological and clinical features, highlighting the importance of a truly multimodal analysis. Indeed, withdrawing a specific data modality (e.g., radiomics features or biological features), led to a decrease of ̃10% of the AUC. Patients were then stratified into two groups based upon their predicted pCR status. These two groups displayed a statistically significant difference in PFS (p

4547 Background: Surgery is the standard of care for localized kidney cancer. Diagnostic imaging plays a critical role in disease staging and informs the extent of surgical resection (partial or radical nephrectomy, extended resection). In clinical routine, up to 15% of the tumors initially assessed as T1-T2 on imaging is upgraded to pT3a status post-surgery, implying a higher risk of relapse. The ability to correctly predict pT3a status pre-surgery would inform the surgical approach. An individualized prediction of the risk of clinical T1 or T2 tumors to be upstaged to pT3a is thus of high surgical interest. Methods: UroCCR is a French national network of 37 multidisciplinary teams for kidney cancer management that collects longitudinal data on the routine clinical care of its patients. A retrospective cohort of 4,395 cases of clinical T1-T2 kidney tumors was analyzed to develop a machine learning-based algorithm predictive of post-surgical pT3a upstaging risk at the individual patient level. For each patient, pre-surgical data were collected, including gender, age, symptoms, tumor size, tumor location, RENAL score, ECOG performance status, ASA score, and post-surgical pathological status. Sites were randomly assigned to the training or testing cohort, and their respective patient cases split between cohorts in a 60/40 ratio. Missing values were addressed through imputation performed with a k-nearest neighbor algorithm. Algorithms were trained on a data set of 2,636 patients and hyperparameters were optimized using a Bayes cross-validation (10-fold) approach. The area under the precision-recall curve (prAUC) was used as optimization metric. The performance of the algorithms for pT3a status prediction was then evaluated on the test dataset of 1,759 patients using precision-recall curves. Results: A logistic regression algorithm reached an AUC of 0.77 and a prAUC of 0.41. Higher values of the tumor size or age at surgery, the hilar location and the presence of symptoms at diagnosis were all associated with an increase of the predicted probability of pT3a upstaging. For each patient, Shapley values graphs were generated to display the pT3a upstaging probability and the relative contribution of each feature to the prediction. Three risk groups were defined based on the relative computed probability of pT3a upstaging, which displayed a statistically significant difference in Disease-Free Survival (DFS) (p < 0.0001), suggesting that pre-surgical multimodal data analysis could help predict long-term outcomes. Conclusions: This study suggests that machine learning applied to pre-surgical multimodal data can predict the risk of pT3a upstaging of a localized kidney tumor and inform long-term outcomes at the individual patient level. The results have been validated on an external cohort of 1,759 patients with data from the clinical routine.

Objectives Genotyping of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has been instrumental in monitoring viral evolution and transmission during the pandemic. The quality of the sequence data obtained from these genotyping efforts depends on several factors, including the quantity/integrity of the input material, the technology, and laboratory-specific implementation. The current lack of guidelines for SARS-CoV-2 genotyping leads to inclusion of error-containing genome sequences in genomic epidemiology studies. We aimed to establish clear and broadly applicable recommendations for reliable virus genotyping. Methods We established and used a sequencing data analysis workflow that reliably identifies and removes technical artefacts; such artefacts can result in miscalls when using alternative pipelines to process clinical samples and synthetic viral genomes with an amplicon-based genotyping approach. We evaluated the impact of experimental factors, including viral load and sequencing depth, on correct sequence determination. Results We found that at least 1000 viral genomes are necessary to confidently detect variants in the SARS-CoV-2 genome at frequencies of ≥10%. The broad applicability of our recommendations was validated in over 200 clinical samples from six independent laboratories. The genotypes we determined for clinical isolates with sufficient quality cluster by sampling location and period. Our analysis also supports the rise in frequencies of 20A.EU1 and 20A.EU2, two recently reported European strains whose dissemination was facilitated by travel during the summer of 2020. Conclusions We present much-needed recommendations for the reliable determination of SARS-CoV-2 genome sequences and demonstrate their broad applicability in a large cohort of clinical samples.

BackgroundSARS-CoV-2 genotyping has been instrumental to monitor virus evolution and transmission during the pandemic. The reliability of the information extracted from the genotyping efforts depends on a number of aspects, including the quality of the input material, applied technology and potential laboratory-specific biases. These variables must be monitored to ensure genotype reliability. The current lack of guidelines for SARS-CoV-2 genotyping leads to inclusion of error-containing genome sequences in studies of viral spread and evolution.ResultsWe used clinical samples and synthetic viral genomes to evaluate the impact of experimental factors, including viral load and sequencing depth, on correct sequence determination using an amplicon-based approach. We found that at least 1000 viral genomes are necessary to confidently detect variants in the genome at frequencies of 10% or higher. The broad applicability of our recommendations was validated in >200 clinical samples from six independent laboratories. The genotypes of clinical isolates with viral load above the recommended threshold cluster by sampling location and period. Our analysis also supports the rise in frequency of 20A.EU1 and 20A.EU2, two recently reported European strains whose dissemination was favoured by travelling during the summer 2020.ConclusionsWe present much-needed recommendations for reliable determination of SARS-CoV-2 genome sequence and demonstrate their broad applicability in a large cohort of clinical samples.

An inflammatory response initiated by the NLRP3 inflammasome is triggered by a variety of situations of host 'danger', including infection and metabolic dysregulation. Previous studies suggested that NLRP3 inflammasome activity is negatively regulated by autophagy and positively regulated by reactive oxygen species (ROS) derived from an uncharacterized organelle. Here we show that mitophagy/autophagy blockade leads to the accumulation of damaged, ROS-generating mitochondria, and this in turn activates the NLRP3 inflammasome. Resting NLRP3 localizes to endoplasmic reticulum structures, whereas on inflammasome activation both NLRP3 and its adaptor ASC redistribute to the perinuclear space where they co-localize with endoplasmic reticulum and mitochondria organelle clusters. Notably, both ROS generation and inflammasome activation are suppressed when mitochondrial activity is dysregulated by inhibition of the voltage-dependent anion channel. This indicates that NLRP3 inflammasome senses mitochondrial dysfunction and may explain the frequent association of mitochondrial damage with inflammatory diseases.

Uncontrolled endoplasmic reticulum (ER) stress responses are proposed to contribute to the pathology of chronic inflammatory diseases such as type 2 diabetes or atherosclerosis. However, the connection between ER stress and inflammation remains largely unexplored. Here, we show that ER stress causes activation of the NLRP3 inflammasome, with subsequent release of the pro-inflammatory cytokine interleukin-1β. This ER-triggered proinflammatory signal shares the same requirement for reactive oxygen species production and potassium efflux compared with other known NLRP3 inflammasome activators, but is independent of the classical unfolded protein response (UPR). We thus propose that the NLRP3 inflammasome senses and responds to ER stress downstream of a previously uncharacterized ER stress response signaling pathway distinct from the UPR, thus providing mechanistic insight to the link between ER stress and chronic inflammatory diseases.

Although the importance of the NOD-like receptor family, pyrin domain containing 3 (NLRP3) inflammasome in health and disease is well appreciated, a precise characterization of NLRP3 expression is yet undetermined. To this purpose, we generated a knock-in mouse in which the Nlrp3 coding sequence was substituted for the GFP (enhanced GFP [egfp]) gene. In this way, the expression of eGFP is driven by the endogenous regulatory elements of the Nlrp3 gene. In this study, we show that eGFP expression indeed mirrors that of NLRP3. Interestingly, splenic neutrophils, macrophages, and, in particular, monocytes and conventional dendritic cells showed robust eGFP fluorescence, whereas lymphoid subsets, eosinophils, and plasmacytoid dendritic cells showed negligible eGFP levels. NLRP3 expression was highly inducible in macrophages, both by MyD88- and Trif-dependent pathways. In vivo, when mice were challenged with diverse inflammatory stimuli, differences in both the number of eGFP-expressing cells and fluorescence intensity were observed in the draining lymph node. Thus, NLRP3 levels at the site of adaptive response initiation are controlled by recruitment of NLRP3-expressing cells and by NLRP3 induction.

Nature 469, 221–225 (2011) On page 2 of the print and PDF version of this Letter, there are two citations to Fig. 2f. The first citation to Fig. 2e, f, which refers to active IL-1β in the culture supernatant after knockdown of beclin 1 and ATG5, should instead be to Fig. 2d, e; the second citation to Fig.

The interleukin-1 (IL-1) family of cytokines has been implicated in the pathogenesis of atherosclerosis in previous studies. The NLRP3 inflammasome has recently emerged as a pivotal regulator of IL-1β maturation and secretion by macrophages. Little is currently known about a possible role for the NLRP3 inflammasome in atherosclerosis progression in vivo. We generated ApoE-/- Nlrp3-/-, ApoE-/- Asc-/- and ApoE-/- caspase-1-/- double-deficient mice, fed them a high-fat diet for 11 weeks and subsequently assessed atherosclerosis progression and plaque phenotype. No differences in atherosclerosis progression, infiltration of plaques by macrophages, nor plaque stability and phenotype across the genotypes studied were found. Our results demonstrate that the NLRP3 inflammasome is not critically implicated in atherosclerosis progression in the ApoE mouse model.

BACKGROUND: Characteristic symptoms of malaria include recurrent fever attacks and neurodegeneration, signs that are also found in patients with a hyperactive Nalp3 inflammasome. Plasmodium species produce a crystal called hemozoin that is generated by detoxification of heme after hemoglobin degradation in infected red blood cells. Thus, we hypothesized that hemozoin could activate the Nalp3 inflammasome, due to its particulate nature reminiscent of other inflammasome-activating agents. METHODOLOGY/PRINCIPAL FINDINGS: We found that hemozoin acts as a proinflammatory danger signal that activates the Nalp3 inflammasome, causing the release of IL-1beta. Similar to other Nalp3-activating particles, hemozoin activity is blocked by inhibiting phagocytosis, K(+) efflux and NADPH oxidase. In vivo, intraperitoneal injection of hemozoin results in acute peritonitis, which is impaired in Nalp3-, caspase-1- and IL-1R-deficient mice. Likewise, the pathogenesis of cerebral malaria is dampened in Nalp3-deficient mice infected with Plasmodium berghei sporozoites, while parasitemia remains unchanged. SIGNIFICANCE/CONCLUSIONS: The potent pro-inflammatory effect of hemozoin through inflammasome activation may possibly be implicated in plasmodium-associated pathologies such as cerebral malaria.