A Magnetic Resonance Imaging Radiomics Signature to Distinguish Benign From Malignant Orbital Lesions

OBJECTIVES

Distinguishing benign from malignant orbital lesions remains challenging both clinically and with imaging, leading to risky biopsies. The objective was to differentiate benign from malignant orbital lesions using radiomics on 3 T magnetic resonance imaging (MRI) examinations.

MATERIALS AND METHODS

This institutional review board-approved prospective single-center study enrolled consecutive patients presenting with an orbital lesion undergoing a 3 T MRI prior to surgery from December 2015 to July 2019. Radiomics features were extracted from 6 MRI sequences (T1-weighted images [WIs], DIXON-T2-WI, diffusion-WI, postcontrast DIXON-T1-WI) using the Pyradiomics software. Features were selected based on their intraobserver and interobserver reproducibility, nonredundancy, and with a sequential step forward feature selection method. Selected features were used to train and optimize a Random Forest algorithm on the training set (75%) with 5-fold cross-validation. Performance metrics were computed on a held-out test set (25%) with bootstrap 95% confidence intervals (95% CIs). Five residents, 4 general radiologists, and 3 expert neuroradiologists were evaluated on their ability to visually distinguish benign from malignant lesions on the test set. Performance comparisons between reader groups and the model were performed using McNemar test. The impact of clinical and categorizable imaging data on algorithm performance was also assessed.

RESULTS

A total of 200 patients (116 [58%] women and 84 [42%] men; mean age, 53.0 ± 17.9 years) with 126 of 200 (63%) benign and 74 of 200 (37%) malignant orbital lesions were included in the study. A total of 606 radiomics features were extracted. The best performing model on the training set was composed of 8 features including apparent diffusion coefficient mean value, maximum diameter on T1-WIs, and texture features. Area under the receiver operating characteristic curve, accuracy, sensitivity, and specificity on the test set were respectively 0.869 (95% CI, 0.834-0.898), 0.840 (95% CI, 0.806-0.874), 0.684 (95% CI, 0.615-0.751), and 0.935 (95% CI, 0.905-0.961). The radiomics model outperformed all reader groups, including expert neuroradiologists (P < 0.01). Adding clinical and categorizable imaging data did not significantly impact the algorithm performance (P = 0.49).

CONCLUSIONS

An MRI radiomics signature is helpful in differentiating benign from malignant orbital lesions and may outperform expert radiologists.

Clinical Impact of Lymphadenectomy after Neoadjuvant Chemotherapy in Advanced Epithelial Ovarian Cancer: A Review of Available Data

Recent robust data allow for omitting lymph node dissection for patients with advanced epithelial ovarian cancer (EOC) and without any suspicion of lymph node metastases, without compromising recurrence-free survival (RFS), nor overall survival (OS), in the setting of primary surgical treatment. Evidence supporting the same postulate for patients undergoing complete cytoreductive surgery after neoadjuvant chemotherapy (NACT) is lacking. Throughout a systematic literature review, the aim of our study was to evaluate the impact of lymph node dissection in patients undergoing surgery for advanced-stage EOC after NACT. A total of 1094 patients, included in six retrospective series, underwent either systematic, selective or no lymph node dissection. Only one study reveals a positive effect of lymphadenectomy on OS, and two on RFS. The four remaining series fail to demonstrate any beneficial effect on survival, neither for RFS nor OS. All of them highlight the higher peri- and post-operative complication rate associated with systematic lymph node dissection. Despite heterogeneity in the design of the studies included, there seems to be a trend showing no improvement on OS for systematic lymph node dissection in node negative patients. A well-conducted prospective trial is mandatory to evaluate this matter.

Prevalence and characteristics of pulmonary embolism in 1042 COVID-19 patients with respiratory symptoms: A nested case-control study

INTRODUCTION

Coronavirus disease 2019 (COVID-19) has been associated with cardiovascular complications and coagulation disorders. Previous studies reported pulmonary embolism (PE) in severe COVID-19 patients. Aim of the study was to estimate the prevalence of symptomatic PE in COVID-19 patients and to identify the clinical, radiological or biological characteristics associated with PE.

PATIENTS/METHODS

We conducted a retrospective nested case-control study in 2 French hospitals. Controls were matched in a 1:2 ratio on the basis of age, sex and center. PE patients with COVID-19 were compared to patients in whom PE was ruled out (CTPA controls) and in whom PE has not been investigated (CT controls).

RESULTS

PE was suspected in 269 patients among 1042 COVID-19 patients, and confirmed in 59 patients (5.6%). Half of PE was diagnosed at COVID-19 diagnosis. PE patients did not differ from CT and CTPA controls for thrombosis risk factors. PE patients more often required invasive ventilation compared to CTPA controls (odds ratio (OR) 2.79; 95% confidence interval (CI) 1.33-5.84) and to CT controls (OR 8.07; 95% CI 2.70-23.82). PE patients exhibited more extensive parenchymal lesions (>50%) than CT controls (OR 3.90; 95% CI 1.54-9.94). D-dimer levels were 5.1 (95% CI 1.90-13.76) times higher in PE patients than CTPA controls.

CONCLUSIONS

Our results suggest a PE prevalence in COVID-19 patients close to 5% in the whole population and to 20% of the clinically suspected population. PE seems to be associated with more extensive lung damage and to require more frequently invasive ventilation.

Socio-economic and psychological impact of the COVID-19 outbreak on private practice and public hospital radiologists

•

During the COVID-19 outbreak, many radiologists expressed anxiety, depression and insomnia symptoms.

•

Working in public hospital was a major protective factor for mental issues.

•

Restricted access to education, past medical history and exposition to COVID-19 were common risks factor of anxiety or depression.

Purpose

The COVID-19 pandemic has led to an urgent reorganisation of the healthcare system to prevent hospitals from overflowing and the virus from spreading. Our objective was to evaluate the socioeconomic and psychological impact of the COVID-19 outbreak on radiologists.

Material and methods

French radiologists were invited to answer an online survey during the pandemic through mailing lists. The questionnaire was accessible for nine days. It covered socio-demographic information, exposure to COVID-19 at work and impact on work organisation, and included the Insomnia Severity Index and Hospital Anxiety and Depression Scale. Outcomes were moderate to severe insomnia, definite symptoms of depression or anxiety. Risk and protective factors were identified through multivariate binary logistic regression.

Results

1515 radiologists answered the survey. Overall, 674 (44.5 %) worked in a highCOVID-19 density area, 671 (44.3 %) were women, and 809 (53.4 %) worked in private practice. Among responders, 186 (12.3 %) expressed insomnia, 222 (14.6 %) anxiety, and 189 (12.5 %) depression symptoms. Lack of protective equipment, increased teleradiology activity and negative impact on education were risk factors for insomnia (respectively OR [95 %CI]:1.7[1.1−2.7], 1.5[1.1−2.2], and 2.5[1.8−3.6]). Female gender, respiratory history, working in COVID-19 high density area, increase of COVID-19 related activity, and impacted education were risk factors for anxiety (OR[95 %CI]:1.7[1.2−2.3], 2[1.1−3.4], 1.5[1.1−2], 1.2[1−1.4], and 2.1[1.5−3]). Conversely, working in a public hospital was a protective factor against insomnia, anxiety, and depression (OR[95 %CI]:0.4[0.2−0.7], 0.6[0.4−0.9], and 0.5[0.3−0.8]).

Conclusions

During COVID-19 pandemic, many radiologists expressed depression, anxiety and insomnia symptoms. Working in a public hospital was a protective factor against every psychological symptom. Socio-economic impact was also major especially in private practice.

Pulmonary embolism in patients with COVID-19 pneumonia

Acute respiratory distress syndrome development in patients with coronavirus disease 2019 (COVID-19) pneumonia is associated with a high mortality rate and is the main cause of death in patients with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection [1]. Myocardial injury has also been reported to be significantly associated with fatal outcome, with a 37% mortality rate in patients without prior cardiovascular disease but elevated troponin levels [2]. A D-dimer level of >1 μg·mL⁻¹ has been clearly identified as a risk factor for poor outcome in SARS-Cov-2 infection [3], with recent reports highlighting a high incidence of thrombotic events in intensive care unit (ICU) patients [4]. A normal D-dimer level allows the safe exclusion of pulmonary embolism (PE) in outpatients with a low or intermediate clinical probability of PE, but there is no recommendation to use D-dimer as a positive marker of thrombosis because of lack of specificity.

This study reports an overall 24% (95% CI 17–32%) cumulative incidence of pulmonary embolism in patients with COVID-19 pneumonia, 50% (30–70%) in ICU and 18% (12–27%) in other patientshttps://bit.ly/35s7hjm

Variation in neutrophil to lymphocyte ratio (NLR) as predictor of outcomes in metastatic renal cell carcinoma (mRCC) and non-small cell lung cancer (mNSCLC) patients treated with nivolumab

BACKGROUND

An elevated pre-treatment neutrophil to lymphocytes ratio (NLR) is associated with poor prognosis in various malignancies. Optimal cut-off is highly variable across studies and could not be determined individually for a patient to inform his prognosis. We hypothesize that NLR variations could be more useful than baseline NLR to predict progression-free survival (PFS) and overall survival (OS) in patients (pts) receiving anti-PD1 treatment.

PATIENTS AND METHODS

All pts with metastatic renal cell carcinoma (mRCC) and metastatic non-small cell lung cancer (mNSCLC) who received anti-PD1 nivolumab monotherapy in second-line setting or later were included in this French multicentric retrospective study. NLR values were prospectively collected prior to each nivolumab administration. Clinical characteristics were recorded. Associations between baseline NLR, NLR variations and survival outcomes were determined using Kaplan-Meier's method and multivariable Cox regression models.

RESULTS

161 pts (86 mRCC and 75 mNSCLC) were included with a median follow-up of 18 months. On the whole cohort, any NLR increase at week 6 was significantly associated with worse outcomes compared to NLR decrease, with a median PFS of 11 months vs 3.7 months (p < 0.0001), and a median OS of 28.5 months vs. 18 months (p = 0.013), respectively. In multivariate analysis, NLR increase was significantly associated with worse PFS (HR 2.2; p = 6.10^-5) and OS (HR 2.1; p = 0.005). Consistent results were observed in each cohort when analyzed separately.

CONCLUSION

Any NLR increase at week 6 was associated with worse PFS and OS outcomes. NLR variation is an inexpensive and dynamic marker easily obtained to monitor anti-PD1 efficacy.

Application of Artificial Intelligence to Gastroenterology and Hepatology

Since 2010, substantial progress has been made in artificial intelligence (AI) and its application to medicine. AI is explored in gastroenterology for endoscopic analysis of lesions, in detection of cancer, and to facilitate the analysis of inflammatory lesions or gastrointestinal bleeding during wireless capsule endoscopy. AI is also tested to assess liver fibrosis and to differentiate patients with pancreatic cancer from those with pancreatitis. AI might also be used to establish prognoses of patients or predict their response to treatments, based on multiple factors. We review the ways in which AI may help physicians make a diagnosis or establish a prognosis and discuss its limitations, knowing that further randomized controlled studies will be required before the approval of AI techniques by the health authorities.

[Naftidrofuryl in arterial obstructive disease: A systematic revue of the literature]

INTRODUCTION

Arterial obstructive disease is a disease affecting 11 % of the general population. This prevalence is constantly increasing. Nafronyl is still prescribed despite a decreasing reimbursement rate since 2005. The objective of this study was to summarize data from the scientific literature on the efficacy and safety of nafronyl used for the treatment of peripheral arterial obstructive disease.

METHOD

A systematic review was made on EMBASE, MEDLINE and the Cochrane Library. Randomized controlled trials, systematic reviews and meta-analyses comparing naftidrofuryl with placebo were included. The main outcome was an improvement in the maximum walking distance or pain free walking distance. The quality of the reviews was analysed using a standardised reading grid. Only the best study was retained.

RESULTS

Among 193articles, one meta-analyses were selected. Naftidrofuryl improved the initial pain free walking distance by 60 % at six months, without a demonstrated increase in the risk of adverse reactions.

CONCLUSION

The efficacy of naftidrofuryl over the maximum walking distance in peripheral arterial obstructive disease appears similar to physical exercise or simvastatin.

[Encapsulant peritonitis]

INTRODUCTION

Encapsulating peritonitis is a rare but severe chronic fibrotic condition related to the development of a white fibrous membrane surrounding the digestive tract. Idiopathic forms have been described, however the disease is most often secondary to peritoneal dialysis or more rarely to surgery. Treatment is difficult and not codified.

CASE REPORT

We report here the observation of a 36-year-old patient whose diagnosis of encapsulating peritonitis was made after a long sub-occlusive history, eight years after a gastric ulcer perforation.

DISCUSSION

We discuss the possible etiologies and we present a focus on this rare and little-known entity.

Validated imaging biomarkers as decision-making tools in clinical trials and routine practice: current status and recommendations from the EIBALL* subcommittee of the European Society of Radiology (ESR)

Observer-driven pattern recognition is the standard for interpretation of medical images. To achieve global parity in interpretation, semi-quantitative scoring systems have been developed based on observer assessments; these are widely used in scoring coronary artery disease, the arthritides and neurological conditions and for indicating the likelihood of malignancy. However, in an era of machine learning and artificial intelligence, it is increasingly desirable that we extract quantitative biomarkers from medical images that inform on disease detection, characterisation, monitoring and assessment of response to treatment. Quantitation has the potential to provide objective decision-support tools in the management pathway of patients. Despite this, the quantitative potential of imaging remains under-exploited because of variability of the measurement, lack of harmonised systems for data acquisition and analysis, and crucially, a paucity of evidence on how such quantitation potentially affects clinical decision-making and patient outcome. This article reviews the current evidence for the use of semi-quantitative and quantitative biomarkers in clinical settings at various stages of the disease pathway including diagnosis, staging and prognosis, as well as predicting and detecting treatment response. It critically appraises current practice and sets out recommendations for using imaging objectively to drive patient management decisions.

How to differentiate uterine leiomyosarcoma from leiomyoma with imaging

Uterine leiomyomas, the most frequent benign myomatous tumors of the uterus, often cannot be distinguished from malignant uterine leiomyosarcomas using clinical criteria. Furthermore, imaging differentiation between both entities is frequently challenging due to their potential overlapping features. Because a suspected leiomyoma is often managed conservatively or with minimally invasive treatments, the misdiagnosis of leiomyosarcoma for a benign leiomyoma could potentially result in significant treatment delays, therefore increasing morbidity and mortality. In this review, we provide an overview of the differences between leiomyoma and leiomyosarcoma, mainly focusing on imaging characteristics, but also briefly touching upon their demographic, histopathological and clinical differences. The main indications and limitations of available cross-sectional imaging techniques are discussed, including ultrasound, computed tomography, magnetic resonance imaging (MRI) and positron emission tomography/computed tomography. A particular emphasis is placed on the review of specific MRI features that may allow distinction between leiomyomas and leiomyosarcomas according to the most recent evidence in the literature. The potential contribution of texture analysis is also discussed. In order to help guide-imaging diagnosis, we provide an MRI-based diagnostic algorithm which takes into account morphological and functional features, both individually and in combination, in an attempt to optimize radiologic differentiation of leiomyomas from leiomyosarcomas.

Management of epithelial cancer of the ovary, fallopian tube, and primary peritoneum. Long text of the Joint French Clinical Practice Guidelines issued by FRANCOGYN, CNGOF, SFOG, and GINECO-ARCAGY, and endorsed by INCa. Part 1: Diagnostic exploration and staging, surgery, perioperative care, and pathology

An MRI is recommended for an ovarian mass that is indeterminate on ultrasound. The ROMA score (combining CA125 and HE4) can also be calculated (grade A). In presumed early-stage ovarian or tubal cancers, the following procedures should be performed: an omentectomy (at a minimum, infracolic), an appendectomy, multiple peritoneal biopsies, peritoneal cytology (grade C), and pelvic and para-aortic lymphadenectomies (grade B) for all histologic types, except the expansile mucinous subtypes, for which lymphadenectomies can be omitted (grade C). Minimally invasive surgery is recommended for early-stage ovarian cancer, when there is no risk of tumor rupture (grade B). For FIGO stages III or IV ovarian, tubal, and primary peritoneal cancers, a contrast-enhanced computed tomography (CT) scan of the thorax/abdomen/pelvis is recommended (grade B), as well as laparoscopic exploration to take multiple biopsies (grade A) and a carcinomatosis score (Fagotti score at a minimum) (grade C) to assess the possibility of complete surgery (i.e., leaving no macroscopic tumor residue). Complete surgery by a midline laparotomy is recommended for advanced ovarian, tubal, or primary peritoneal cancer (grade B). For advanced cancers, para-aortic and pelvic lymphadenectomies are recommended when metastatic adenopathy is clinically or radiologically suspected (grade B). When adenopathy is not suspected and when complete peritoneal surgery is performed as the initial surgery for advanced cancer, the lymphadenectomies can be omitted because they do not modify either the medical treatment or overall survival (grade B). Primary surgery (before other treatment) is recommended whenever it appears possible to leave no tumor residue (grade B).

Artificial intelligence to diagnose meniscus tears on MRI

PURPOSE

The purpose of this study was to build and evaluate a high-performance algorithm to detect and characterize the presence of a meniscus tear on magnetic resonance imaging examination (MRI) of the knee.

MATERIAL AND METHODS

An algorithm was trained on a dataset of 1123 MR images of the knee. We separated the main task into three sub-tasks: first to detect the position of both horns, second to detect the presence of a tear, and last to determine the orientation of the tear. An algorithm based on fast-region convolutional neural network (CNN) and faster-region CNN, was developed to classify the tasks. The algorithm was thus used on a test dataset composed of 700 images for external validation. The performance metric was based on area under the curve (AUC) analysis for each task and a final weighted AUC encompassing the three tasks was calculated.

RESULTS

The use of our algorithm yielded an AUC of 0.92 for the detection of the position of the two meniscal horns, of 0.94 for the presence of a meniscal tear and of 083 for determining the orientation of the tear, resulting in a final weighted AUC of 0.90.

CONCLUSION

We demonstrate that our algorithm based on fast-region CNN is able to detect meniscal tears and is a first step towards developing more end-to-end artificial intelligence-powered diagnostic tools.

Epidemiology of Lyme borreliosis through two surveillance systems: the national Sentinelles GP network and the national hospital discharge database, France, 2005 to 2016

BACKGROUND

Lyme borreliosis (LB) is the most frequent vector-borne disease in France. Since 2009, surveillance of LB is conducted by a sentinel network of general practitioners (GPs). This system, in conjunction with the national hospitalisation database was used to estimate the incidence and describe the characteristics of LB in France.

AIM

To describe the estimated incidence and trends in GP consultations and hospital admissions for LB in France and identify risk groups and high-incidence regions.

RESULTS

From 2011 to 2016, the mean yearly incidence rate of LB cases was 53 per 100,000 inhabitants (95% CI: 41-65) ranging from 41 in 2011 to 84 per 100 000 in 2016. A mean of 799 cases per year were hospitalised with LB associated diagnoses 2005-16. The hospitalisation incidence rate (HIR) ranged from 1.1 cases per 100,000 inhabitants in 2005 to 1.5 in 2011 with no statistically significant trend. We observed seasonality with a peak during the summer, important inter-regional variations and a bimodal age distribution in LB incidence and HIR with higher incidence between 5 and 9 year olds and those aged 60 years. Erythema migrans affected 633/667 (95%) of the patients at primary care level. Among hospitalised cases, the most common manifestation was neuroborreliosis 4,906/9,594 (51%).

CONCLUSION

Public health strategies should focus on high-incidence age groups and regions during the months with the highest incidences and should emphasise prevention measures such as regular tick checks after exposure and prompt removal to avoid infection.

Gray-level discretization impacts reproducible MRI radiomics texture features

OBJECTIVES

To assess the influence of gray-level discretization on inter- and intra-observer reproducibility of texture radiomics features on clinical MR images.

MATERIALS AND METHODS

We studied two independent MRI datasets of 74 lacrymal gland tumors and 30 breast lesions from two different centers. Two pairs of readers performed three two-dimensional delineations for each dataset. Texture features were extracted using two radiomics softwares (Pyradiomics and an in-house software). Reproducible features were selected using a combination of intra-class correlation coefficient (ICC) and concordance and coherence coefficient (CCC) with 0.8 and 0.9 as thresholds, respectively. We tested six absolute and eight relative gray-level discretization methods and analyzed the distribution and highest number of reproducible features obtained for each discretization. We also analyzed the number of reproducible features extracted from computer simulated delineations representative of inter-observer variability.

RESULTS

The gray-level discretization method had a direct impact on texture feature reproducibility, independent of observers, software or method of delineation (simulated vs. human). The absolute discretization consistently provided statistically significantly more reproducible features than the relative discretization. Varying the bin number of relative discretization led to statistically significantly more variable results than varying the bin size of absolute discretization.

CONCLUSIONS

When considering inter-observer reproducible results of MRI texture radiomics features, an absolute discretization should be favored to allow the extraction of the highest number of potential candidates for new imaging biomarkers. Whichever the chosen method, it should be systematically documented to allow replicability of results.