Laboratoire d'imagerie translationnelle en oncologie

PURPOSE: To propose a new quality scoring tool, METhodological RadiomICs Score (METRICS), to assess and improve research quality of radiomics studies. METHODS: We conducted an online modified Delphi study with a group of international experts. It was performed in three consecutive stages: Stage#1, item preparation; Stage#2, panel discussion among EuSoMII Auditing Group members to identify the items to be voted; and Stage#3, four rounds of the modified Delphi exercise by panelists to determine the items eligible for the METRICS and their weights. The consensus threshold was 75%. Based on the median ranks derived from expert panel opinion and their rank-sum based conversion to importance scores, the category and item weights were calculated. RESULT: In total, 59 panelists from 19 countries participated in selection and ranking of the items and categories. Final METRICS tool included 30 items within 9 categories. According to their weights, the categories were in descending order of importance: study design, imaging data, image processing and feature extraction, metrics and comparison, testing, feature processing, preparation for modeling, segmentation, and open science. A web application and a repository were developed to streamline the calculation of the METRICS score and to collect feedback from the radiomics community. CONCLUSION: In this work, we developed a scoring tool for assessing the methodological quality of the radiomics research, with a large international panel and a modified Delphi protocol. With its conditional format to cover methodological variations, it provides a well-constructed framework for the key methodological concepts to assess the quality of radiomic research papers. CRITICAL RELEVANCE STATEMENT: A quality assessment tool, METhodological RadiomICs Score (METRICS), is made available by a large group of international domain experts, with transparent methodology, aiming at evaluating and improving research quality in radiomics and machine learning. KEY POINTS: • A methodological scoring tool, METRICS, was developed for assessing the quality of radiomics research, with a large international expert panel and a modified Delphi protocol. • The proposed scoring tool presents expert opinion-based importance weights of categories and items with a transparent methodology for the first time. • METRICS accounts for varying use cases, from handcrafted radiomics to entirely deep learning-based pipelines. • A web application has been developed to help with the calculation of the METRICS score ( https://metricsscore.github.io/metrics/METRICS.html ) and a repository created to collect feedback from the radiomics community ( https://github.com/metricsscore/metrics ).

Abstract Background Manual quantification of the metabolic tumor volume (MTV) from whole-body 18 F-FDG PET/CT is time consuming and therefore usually not applied in clinical routine. It has been shown that neural networks might assist nuclear medicine physicians in such quantification tasks. However, little is known if such neural networks have to be designed for a specific type of cancer or whether they can be applied to various cancers. Therefore, the aim of this study was to evaluate the accuracy of a neural network in a cancer that was not used for its training. Methods Fifty consecutive breast cancer patients that underwent 18 F-FDG PET/CT were included in this retrospective analysis. The PET-Assisted Reporting System (PARS) prototype that uses a neural network trained on lymphoma and lung cancer 18 F-FDG PET/CT data had to detect pathological foci and determine their anatomical location. Consensus reads of two nuclear medicine physicians together with follow-up data served as diagnostic reference standard; 1072 18 F-FDG avid foci were manually segmented. The accuracy of the neural network was evaluated with regard to lesion detection, anatomical position determination, and total tumor volume quantification. Results If PERCIST measurable foci were regarded, the neural network displayed high per patient sensitivity and specificity in detecting suspicious 18 F-FDG foci (92%; CI = 79–97% and 98%; CI = 94–99%). If all FDG-avid foci were regarded, the sensitivity degraded (39%; CI = 30–50%). The localization accuracy was high for body part (98%; CI = 95–99%), region (88%; CI = 84–90%), and subregion (79%; CI = 74–84%). There was a high correlation of AI derived and manually segmented MTV ( R 2 = 0.91; p < 0.001). AI-derived whole-body MTV (HR = 1.275; CI = 1.208–1.713; p < 0.001) was a significant prognosticator for overall survival. AI-derived lymph node MTV (HR = 1.190; CI = 1.022–1.384; p = 0.025) and liver MTV (HR = 1.149; CI = 1.001–1.318; p = 0.048) were predictive for overall survival in a multivariate analysis. Conclusion Although trained on lymphoma and lung cancer, PARS showed good accuracy in the detection of PERCIST measurable lesions. Therefore, the neural network seems not prone to the clever Hans effect. However, the network has poor accuracy if all manually segmented lesions were used as reference standard. Both the whole body and organ-wise MTV were significant prognosticators of overall survival in advanced breast cancer.

State-of-the-art patient management frequently mandates the investigation of both anatomy and physiology of the patients. Hybrid imaging modalities such as the PET/MRI, PET/CT and SPECT/CT have the ability to provide both structural and functional information of the investigated tissues in a single examination. With the introduction of such advanced hardware fusion, new problems arise such as the exceedingly large amount of multi-modality data that requires novel approaches of how to extract a maximum of clinical information from large sets of multi-dimensional imaging data. Artificial intelligence (AI) has emerged as one of the leading technologies that has shown promise in facilitating highly integrative analysis of multi-parametric data. Specifically, the usefulness of AI algorithms in the medical imaging field has been heavily investigated in the realms of (1) image acquisition and reconstruction, (2) post-processing and (3) data mining and modelling. Here, we aim to provide an overview of the challenges encountered in hybrid imaging and discuss how AI algorithms can facilitate potential solutions. In addition, we highlight the pitfalls and challenges in using advanced AI algorithms in the context of hybrid imaging and provide suggestions for building robust AI solutions that enable reproducible and transparent research.

The International Society of Paediatric Oncology Renal Tumour Study Group (SIOP-RTSG) advocate treating children with Wilms tumour (WT) with preoperative chemotherapy, whereas the Renal Tumor Committee of the Children's Oncology Group (COG) advocates primary nephrectomy (without biopsy) when feasible. Successive SIOP-RTSG trial protocols recommended pretreatment biopsy of children with unilateral tumours only where there were features to suggest an increased probability of a non-WT requiring a change in management. The UK experience in the SIOP WT 2001 trial showed that an alternate approach of performing biopsies on all children with renal tumour masses to determine histology at diagnosis rarely changes management, and can result in misdiagnosis (particularly patients in the age range typical for WT). Although a more selective approach to biopsy has been routine practice in all other countries participating in SIOP-RTSG trials, there was variation between national groups. To address this variation and provide evidence-based recommendations for the indications and recommended approach to renal tumour biopsy within the SIOP paradigm, an international, multidisciplinary working group of SIOP-RTSG members was convened. We describe the resulting recommendations of this group, which are to be incorporated in the ongoing SIOP-RTSG UMBRELLA study.

PURPOSE OF REVIEW: The aim of this study was to highlight the diagnostic and management challenges of primary vitreoretinal lymphoma (PVRL) through a review of the literature and a European survey on real-life practices for PVRL. RECENT FINDINGS: The care of PVRL patients is heterogeneous between specialists and countries. Upfront systemic treatment based on high-dose methotrexate chemotherapy, with or without local treatment, might reduce or delay the risk of brain relapse.Ibrutinib, lenalidomide with or without rituximab, and temozolomide are effective for patients with relapsed/refractory PVRL and should be tested as first-line treatments. SUMMARY: The prognosis of PVRL remains dismal. No firm conclusion regarding optimal treatment can yet be drawn. The risk of brain relapse remains high. Diagnostic procedures and assessment of therapeutic responses need to be homogenized. Collaboration between specialists involved in PVRL and multicentric prospective therapeutic studies are strongly needed. The recommendations of the French group for primary oculocerebral lymphoma (LOC network) are provided, as a basis for further European collaborative work.

Purpose: To design and validate a preprocessing procedure dedicated to T2-weighted MR images of lung cancers so as to improve the ability of radiomic features to distinguish between adenocarcinoma and other histological types. Materials and Methods: A discovery set of 52 patients with advanced lung cancer who underwent T2-weighted MR imaging at 3 Tesla in a single center study from August 2017 to May 2019 was used. Findings were then validated using a validation set of 19 additional patients included from May to October 2019. Tumor type was obtained from the pathology report after trans-thoracic needle biopsy, metastatic lymph node or metastasis samples, or surgical excisions. MR images were preprocessed using N4ITK bias field correction and by normalizing voxel intensities with fat as a reference region. Segmentation and extraction of radiomic features were performed with LIFEx software on the raw images, on the N4ITK-corrected images and on the fully preprocessed images. Two analyses were conducted where radiomic features were extracted: from the whole tumor volume (3D analysis); 2) from all slices encompassing the tumor (2D analysis). Receiver operating characteristic (ROC) analysis was used to identify features that could distinguish between adenocarcinoma and other histological types. Sham experiments were also designed to control the number of false positive findings. Results: There were 31 (12) adenocarcinomas and 21 (7) other histological types in the discovery (validation) set. In 2D, preprocessing increased the number of discriminant radiomic features from 8 without preprocessing to 22 with preprocessing. 2D analysis yielded more features able to identify adenocarcinoma than 3D analysis (12 discriminant radiomic features after preprocessing in 3D). Preprocessing did not increase false positive findings as no discriminant features were identified in any of the sham experiments. The greatest sensitivity of the 2D analysis applied to preprocessed data was confirmed in the validation set. Conclusion: Correction for magnetic field inhomogeneities and normalization of voxel values are essential to reveal the full potential of radiomic features to identify the tumor histological type from MR T2-weighted images, with classification performance similar to those reported in PET/CT and in multiphase CT in lung cancers.

Dozens of articles describing artificial intelligence (AI) developments are submitted to medical imaging journals every month, including in the nuclear medicine field. Our mission, as a nuclear medicine community, is to contribute to a better understanding of normal and pathologic processes by

PURPOSE: FDOPA PET shows good performance for the diagnosis of striatal dopaminergic denervation, making it a valuable tool for the differential diagnosis of Parkinsonism. Textural features are image biomarkers that could potentially improve the early diagnosis and monitoring of neurodegenerative parkinsonian syndromes. We explored the performances of textural features for binary classification of FDOPA scans. METHODS: We used two FDOPA PET datasets: 443 scans for feature selection, and 100 scans from a different PET/CT system for model testing. Scans were labelled according to expert interpretation (dopaminergic denervation versus no dopaminergic denervation). We built LASSO logistic regression models using 43 biomarkers including 32 textural features. Clinical data were also collected using a shortened UPDRS scale. RESULTS: The model built from the clinical data alone had a mean area under the receiver operating characteristics (AUROC) of 63.91. Conventional imaging features reached a maximum score of 93.47 but the addition of textural features significantly improved the AUROC to 95.73 (p < 0.001), and 96.10 (p < 0.001) when limiting the model to the top three features: GLCM_Correlation, Skewness and Compacity. Testing the model on the external dataset yielded an AUROC of 96.00, with 95% sensitivity and 97% specificity. GLCM_Correlation was one of the most independent features on correlation analysis, and systematically had the heaviest weight in the classification model. CONCLUSION: A simple model with three radiomic features can identify pathologic FDOPA PET scans with excellent sensitivity and specificity. Textural features show promise for the diagnosis of parkinsonian syndromes.

Purpose Predicting H3.1, TP53, and ACVR1 mutations in DIPG could aid in the selection of therapeutic options. The contribution of clinical data and multi-modal MRI were studied for these three predictive tasks. To keep the maximum number of subjects, which is essential for a rare disease, missing data were considered. A multi-modal model was proposed, collecting all available data for each patient, without performing any imputation. Methods A retrospective cohort of 80 patients with confirmed DIPG and at least one of the four MR modalities (T1w, T1c, T2w, and FLAIR), acquired with two different MR scanners was built. A pipeline including standardization of MR data and extraction of radiomic features within the tumor was applied. The values of radiomic features between the two MR scanners were realigned using the ComBat method. For each prediction task, the most robust features were selected based on a recursive feature elimination with cross-validation. Five different models, one based on clinical data and one per MR modality, were developed using logistic regression classifiers. The prediction of the multi-modal model was defined as the average of all possible prediction results among five for each patient. The performances of the models were compared using a leave-one-out approach. Results The percentage of missing modalities ranged from 6 to 11% across modalities and tasks. The performance of each individual model was dependent on each specific task, with an AUC of the ROC curve ranging from 0.63 to 0.80. The multi-modal model outperformed the clinical model for each prediction tasks, thus demonstrating the added value of MRI. Furthermore, regardless of performance criteria, the multi-modal model came in the first place or second place (very close to first). In the leave-one-out approach, the prediction of H3.1 (resp. ACVR1 and TP53) mutations achieved a balanced accuracy of 87.8% (resp. 82.1 and 78.3%). Conclusion Compared with a single modality approach, the multi-modal model combining multiple MRI modalities and clinical features was the most powerful to predict H3.1, ACVR1, and TP53 mutations and provided prediction, even in the case of missing modality. It could be proposed in the absence of a conclusive biopsy.

PurposeAccurate dosimetry is paramount to study the FLASH biological effect since dose and dose rate are critical dosimetric parameters governing its underlying mechanisms. With the goal of assessing the suitability of standard clinical dosimeters in a very-high dose rate (VHDR) experimental setup, we evaluated the ion collection efficiency of several commercially available air-vented ionization chambers (IC) in conventional and VHDR proton irradiation conditions.MethodsA cyclotron at the Orsay Proton Therapy Center was used to deliver VHDR pencil beam scanning irradiation. Ion recombination correction factors (ks) were determined for several detectors (Advanced Markus, PPC05, Nano Razor, CC01) at the entrance of the plateau and at the Bragg peak, using the Niatel model, the Two-voltage method and Boag’s analytical formula for continuous beams.ResultsMean dose rates ranged from 4 Gy/s to 385 Gy/s, and instantaneous dose rates up to 1000 Gy/s were obtained with the experimental set-up. Recombination correction factors below 2 % were obtained for all chambers, except for the Nano Razor, at VHDRs with variations among detectors, while ks values were significantly smaller (0.8 %) for conventional dose rates.ConclusionsWhile the collection efficiency of the probed ICs in scanned VHDR proton therapy is comparable to those in the conventional regime with recombination coefficiens smaller than 1 % for mean dose rates up to 177 Gy/s, the reduction in collection efficiency for higher dose rates cannot be ignored when measuring the absorbed dose in pre-clinical proton scanned FLASH experiments and clinical trials.

PURPOSES: The aim of the study was to assess the efficacy of a treatment protocol that combines photodynamic therapy (PDT) and nitroglycerin (NG) on human retinoblastoma tumors xenografted on mice. We aimed to increase the PDT efficiency (in our least treatment-responsive retinoblastoma line) with better PS delivery to the tumor generated by NG, which is known to dilate vessels and enhance the permeability and retention of macromolecules in solid tumors. METHODS: In vivo follow-up of the therapeutic effects was performed by sodium MRI, which directly monitors variations in sodium concentrations non-invasively and can be used to track the tumor response to therapy. NG ointment was applied one hour before PDT. The PDT protocol involves double-tumor targeting, i.e., cellular and vascular. The first PS dose was injected followed by a second one, separated by a 3 h interval. The timelapse allowed the PS molecules to penetrate tumor cells. Ten minutes after the second dose, the PS was red-light-activated. RESULTS: In this study, we observed that the PDT effect was enhanced by applying nitroglycerin ointment to the tumor-bearing animal's skin. PDT initiates the bystander effect on retinoblastomas, and NG increases this effect by increasing the intratumoral concentration of PS, which induces a higher production of ROS in the illuminated region and thus increases the propagation of the cell death signal deeper into the tumor (bystander effect).

The real-world implementation of federated learning is complex and requires research and development actions at the crossroad between different domains ranging from data science, to software programming, networking, and security. While today several FL libraries are proposed to data scientists and users, most of these frameworks are not designed to find seamless application in medical use-cases, due to the specific challenges and requirements of working with medical data and hospital infrastructures. Moreover, governance, design principles, and security assumptions of these frameworks are generally not clearly illustrated, thus preventing the adoption in sensitive applications. Motivated by the current technological landscape of FL in healthcare, in this document we present Fed-BioMed: a research and development initiative aiming at translating federated learning (FL) into real-world medical research applications. We describe our design space, targeted users, domain constraints, and how these factors affect our current and future software architecture.

Proton therapy allows to avoid excess radiation dose on normal tissues. However, there are some limitations. Indeed, passive delivery of proton beams results in an increase in the lateral dose upstream of the tumor and active scanning leads to strong differences in dose delivery. This study aims to assess possible differences in the transcriptomic response of skin in C57BL/6 mice after TBI irradiation by active or passive proton beams at the dose of 6 Gy compared to unirradiated mice. In that purpose, total RNA was extracted from skin samples 3 months after irradiation and RNA-Seq was performed. Results showed that active and passive delivery lead to completely different transcription profiles. Indeed, 140 and 167 genes were differentially expressed after active and passive scanning compared to unirradiated, respectively, with only one common gene corresponding to RIKEN cDNA 9930021J03. Moreover, protein-protein interactions performed by STRING analysis showed that 31 and 25 genes are functionally related after active and passive delivery, respectively, with no common gene between both types of proton delivery. Analysis showed that active scanning led to the regulation of genes involved in skin development which was not the case with passive delivery. Moreover, 14 ncRNA were differentially regulated after active scanning against none for passive delivery. Active scanning led to 49 potential mRNA-ncRNA pairs with one ncRNA mainly involved, Gm44383 which is a miRNA. The 43 genes potentially regulated by the miRNA Gm44393 confirmed an important role of active scanning on skin keratin pathway. Our results demonstrated that there are differences in skin gene expression still 3 months after proton irradiation versus unirradiated mouse skin. And strong differences do exist in late skin gene expression between scattered or scanned proton beams. Further investigations are strongly needed to understand this discrepancy and to improve treatments by proton therapy.

learning.Representation learning involves training a neural network to learn a parsimonious yet comprehensive representation of the useful information present in the images.This information can then be used as latent deep radiomic features to build models corresponding to different classification or prediction tasks.Unlike handcrafted features, not only the values, but the definitions of the features involved in deep radiomics themselves depend on the training data and model architecture used to train a classification or prediction model.While deep learning is gaining ground thanks to the increasing availability of open datasets and shared deep learning models (https://nmmitools.org,https://monai.io),one might ask whether researchers should focus efforts exclusively on deep learning rather than radiomics.This short paper examines the respective positions of handcrafted and end-to-end deep radiomics in relation to key factors to be considered when developing and implementing classification or outcome prediction models.It has been inspired by the content of a moderated debate held during the SNMMI 2023 meeting involving two advocates of handcrafted radiomics and two advocates of deep radiomics.

International audience

Gallium-68-labeled FAPI-46 has recently been proposed as a novel positron emission tomography imaging probe to diagnose and monitor a wide variety of cancers. Promising results from several ongoing clinical trials have led to a soaring demand for this radiotracer. Typical [ 68 Ga]Ga-FAPI-46 labeling protocols do not cope with multiple generator elutions, leaving radiopharmacies unable to scale-up the production and meet the demand. Here, we propose a robust and efficient automated radiosynthesis of [ 68 Ga]Ga-FAPI-46 on the Trasis miniAllinOne synthesizer, featuring a prepurification step which allows multiple generator elutions and ensures compatibility with a wide range of gallium-68 generators. Our approach was to optimize the prepurification step by first testing five different cationic cartridge chemistries. Only the strong cationic exchange (SCX) cartridges tested had sufficient affinities for quantitative trapping of &amp;gt;99.9%, while the weak cationics did not exceed 50%. Packaging, rinsing, or flowing of the selected SCX cartridges was not noticeable, but improvements in fluidics managed to save time. Based on our previous development experience of [ 68 Ga]Ga-FAPI-46, radiolabeling optimization was also carried out at different temperatures during 10 min. At temperatures above 100°C, radiochemical yield (RCY) &amp;gt; 80% was achieved without significantly increasing the chemical impurities (&amp;lt;5.5 μg mL -1 ). The optimized sequence was reproducibly conducted with three different brands of widely used generators (RCY &amp;gt;88%). A comparison with radiosyntheses carried out without prepurification steps was also conclusive in terms of RCY, radiochemical yield, and chemical purity. Finally, high-activity tests using elutions from three generators were also successful for these parameters. [ 68 Ga]Ga-FAPI-46 was consistently obtained in good radiochemical yields (&amp;gt;89%, n = 3 ), and the final product quality was compliant with internal specifications based on European Pharmacopoeia. This process is suitable for GMP production and allows scaling-up of routine productions, higher throughput, and, ultimately, better patient care.

Purpose: Effective dose to circulating immune cells (EDIC) is associated with survival in lung and esophageal cancer patients. This study aimed to evaluate the benefit of intensity-modulated proton therapy (IMPT) for EDIC reduction compared with volumetric modulated arc therapy (VMAT) in mediastinal Hodgkin lymphoma (mHL) patients. Materials and Methods: Ten consecutive mHL patients treated with involved-site IMPT after frontline chemotherapy were included. The mean dose to the heart, lung, and liver and the integral dose to the body were obtained, and we calculated EDIC based on these variables. The effective dose to circulating immune cells was compared between IMPT and VMAT schedules. Results: < .01). Integral dose reduction was the main driver of EDIC reduction with IMPT, followed by lung sparing. Conclusion: Intensity-modulated proton therapy significantly reduced EDIC in mHL patients undergoing consolidation involved-site radiation therapy. Integral dose reduction combined with improved lung sparing was the main driver of EDIC reduction with IMPT.

Ziel/Aim Automatic quantification of the metabolic tumor volume (MTV) from PET data by neural networks can simplify the clinical implementation of the parameter, whose manual estimation is a laborious process. It remains unclear how specific a neural network has to be designed for the investigated disease. The aim of this study was to evaluate the accuracy of a neural network that was trained on lymphoma and lung cancer FDG PET/CT data in the evaluation of breast cancer patients.

Laser-driven proton sources have long been developed with an eye on their potential for medical application to radiation therapy. These sources are compact, versatile, and show peculiar characteristics such as extreme instantaneous dose rates, short duration and broad energy spectrum. Typical temporal modality of laser-driven irradiation, the so-called fast-fractionation, results from the composition of multiple, temporally separated, ultra-short dose fractions. In this paper we present the use of a high-energy laser system for delivering the target dose in a single nanosecond pulse, for ultra-fast irradiation of biological samples. A transport line composed by two permanent-magnet quadrupoles and a scattering system is used to improve the dose profile and to control the delivered dose-per-pulse. A single-shot dosimetry protocol for the broad-spectrum proton source using Monte Carlo simulations was developed. Doses as high as 20 Gy could be delivered in a single shot, lasting less than 10 ns over a 1 cm diameter biological sample, at a dose-rate exceeding [Formula: see text]. Exploratory application of extreme laser-driven irradiation conditions, falling within the FLASH irradiation protocol, are presented for irradiation in vitro and in vivo. A reduction of radiation-induced oxidative stress in vitro and radiation-induced developmental damage compatible with the onset of FLASH effect were observed in vivo, whereas anti-tumoral efficacy was confirmed by cell survival assay.

This paper proposes the estimation of a mutual shape from a set of different segmentation results using both active contours and information theory. The mutual shape is here defined as a consensus shape estimated from a set of different segmentations of the same object. In an original manner, such a shape is defined as the minimum of a criterion that benefits from both the mutual information and the joint entropy of the input segmentations. This energy criterion is justified using similarities between information theory quantities and area measures, and presented in a continuous variational framework. In order to solve this shape optimization problem, shape derivatives are computed for each term of the criterion and interpreted as an evolution equation of an active contour. A mutual shape is then estimated together with the sensitivity and specificity of each segmentation. Some synthetic examples allow us to cast the light on the difference between the mutual shape and an average shape. The applicability of our framework has also been tested for segmentation evaluation and fusion of different types of real images (natural color images, old manuscripts, medical images).