Search

Abstract Bone segmentation is an important step to perform biomechanical failure load simulations on in-vivo CT data of patients with bone metastasis, as it is a mandatory operation to obtain meshes needed for numerical simulations. Segmentation can be a tedious and time consuming task when done manually, and expert segmentations are subject to intra- and inter-operator variability. Deep learning methods are increasingly employed to automatically carry out image segmentation tasks. These networks usually need to be trained on a large image dataset along with the manual segmentations to maximize generalization to new images, but it is not always possible to have access to a multitude of CT-scans with the associated ground truth. It then becomes necessary to use training techniques to make the best use of the limited available data. In this paper, we propose a dedicated pipeline of preprocessing, deep learning based segmentation method and post-processing for in-vivo human femurs and vertebrae segmentation from CT-scans volumes. We experimented with three U-Net architectures and showed that out-of-the-box models enable automatic and high-quality volume segmentation if carefully trained. We compared the failure load simulation results obtained on femurs and vertebrae using either automatic or manual segmentations and studied the sensitivity of the simulations on small variations of the automatic segmentation. The failure loads obtained using automatic segmentations were comparable to those obtained using manual expert segmentations for all the femurs and vertebrae tested, demonstrating the effectiveness of the automated segmentation approach for failure load simulations.

Objectives: We evaluated the contribution of lung lesion quantification on chest CT using a clinical Artificial Intelligence (AI) software in predicting death and intensive care units (ICU) admission for COVID-19 patients. Methods: For 349 patients with positive COVID-19-PCR test that underwent a chest CT scan at admittance or during hospitalization, we applied the AI for lung and lung lesion segmentation to obtain lesion volume (LV), and LV/Total Lung Volume (TLV) ratio. ROC analysis was used to extract the best CT criterion in predicting death and ICU admission. Two prognostic models using multivariate logistic regressions were constructed to predict each outcome and were compared using AUC values. The first model (“Clinical”) was based on patients’ characteristics and clinical symptoms only. The second model (“Clinical+LV/TLV”) included also the best CT criterion. Results: LV/TLV ratio demonstrated best performance for both outcomes; AUC of 67.8% (95% CI: 59.5 - 76.1) and 81.1% (95% CI: 75.7 - 86.5) respectively. Regarding death prediction, AUC values were 76.2% (95% CI: 69.9 - 82.6) and 79.9% (95%IC: 74.4 - 85.5) for the “Clinical” and the “Clinical+LV/TLV” models respectively, showing significant performance increase (+ 3.7%; p-value

Introduction With the spread of COVID-19 epidemic, health plans must be adapted continuously. There is an urgent need to define the best care courses of patients with COVID-19, especially in intensive care units (ICUs), according to their individualised benefit/risk ratio. Since older age is associated with poorer short-term and long-term outcomes, prediction models are needed, that may assist clinicians in their ICU admission decision. Senior-COVID-Rea was designed to evaluate, in patients over 60 years old admitted in ICU for severe COVID-19 disease, the impact of age and geriatric and paraclinical parameters on their mortality 30 days after ICU admission.Methods and analysis This is a multicentre survey protocol to be conducted in seven hospitals of the Auvergne-Rhône-Alpes region, France. All patients over 60 years old admitted in ICU for severe COVID-19 infection (or their legally acceptable representative) will be proposed to enter the study and to fill in a questionnaire regarding their functional and nutritional parameters 1 month before COVID-19 infection. Paraclinical parameters at ICU admission will be collected: lymphocytes and neutrophils counts, high-fluorescent lymphoid cells and immature granulocytes percentages (Sysmex data), D-dimers, C-reactive protein, lactate dehydrogenase (LDH), creatinine, CT scan for lung extension rate as well as clinical resuscitation scores, and the delay between the first signs of infection and ICU admission. The primary outcome will be the overall survival at day 30 post-ICU admission. The analysis of factors predicting mortality at day 30 will be carried out using univariate and multivariate logistic regressions. Multivariate logistic regression will consider up to 15 factors.The ambition of this trial, which takes into account the different approaches of geriatric vulnerability, is to define the respective abilities of different operational criteria of frailty to predict patients’ outcomes.Ethics and dissemination The study protocol was ethically approved. The results of the primary and secondary objectives will be published in peer-reviewed journals.Trial registration number NCT04422340.

Detection of molecular alterations in lung cancer bone metastasis (LCBM) is particularly difficult when decalcification procedure is needed. The Idylla™ real-time (RT)-PCR is compared to the routine method used in our laboratory, which combines next generation and Sanger sequencing, for the detection of EGFR mutations in LCBM.LCBM subjected to EDTA or formic acid decalcification were analysed for EGFR mutational status using two methods: first, the Ion Torrent Ampliseq next generation sequencing (NGS) assay +/- Sanger sequencing was used prospectively; then, the fully-automated, RT-PCR based molecular testing system Idylla™ EGFR Mutation Test was applied retrospectively.Out of the 34 LCBM assayed, 14 (41.2%) were unsuitable for NGS analysis and five remained unsuitable after additional Sanger EGFR sequencing (5/34, 14.7%). Using Idylla™, valid results were observed for 33/34 samples (97.1%). The concordance between the NGS +/- Sanger sequencing method and the RT-PCR method was 89.7% (26/29), one false positive EGFR S768I mutation and two false negative results were observed using Idylla™; one of these false negative cases was diagnosed by Sanger sequencing with a rare exon 19 EGFR mutation not covered by the Idylla™ EGFR Mutation Test design.Detection of EGFR mutations in decalcified LCBM is challenging using NGS, more than half of samples showing invalid results. Alternative methods should thus be preferred to spare clinical samples and decrease delay. The Idylla™ EGFR Mutation Test shows a good performance on decalcified bone samples and could be used as a first step. In case of negative results, a sequencing approach is mandatory to check the presence of rare EGFR mutations sensitive to EGFR tyrosine kinase inhibitors. Keywords: Lung carcinoma, EGFR mutation, Bone metastases, Decalcification, Tyrosine kinase inhibitor

As for molecular alterations of lung adenocarcinoma, it is critical that pathologists are able to give PD-L1 expression status before first-line of treatment. The present study compared PD-L1 expression (clone 22-C3) in decalcified using EDTA or formic acid and non-decalcified lung cancer metastases bone samples. Amongst the 84 bone samples analysed for PD-L1 expression, and independently of decalcification, TPS ≥ 1% was 25.0% and ≥ 50% was 11.4%. There was no significant difference between decalcified samples (n = 45) and non-decalcified samples (n = 39) for both TPS ≥ 1% (p = 0.32) and TPS ≥ 50% (p = 1). To conclude, we confirm decalcified bone metastasis specimens may be used for PD-L1 IHC in routine practice. These results also highlight potentially interesting specificities of the bone microenvironment that should be further studied Keywords: Lung carcinoma, PD-L1, Bone metastases, Decalcification, Immunotherapy