Artificial intelligence solution to classify pulmonary nodules on CT

PURPOSE

The purpose of this study was to create an algorithm to detect and classify pulmonary nodules in two categories based on their volume greater than 100 mm³ or not, using machine learning and deep learning techniques.

MATERIALS AND METHOD

The dataset used to train the model was provided by the organization team of the SFR (French Radiological Society) Data Challenge 2019. An asynchronous and parallel 3-stages pipeline was developed to process all the data (a data "pre-processing" stage; a "nodule detection" stage; a "classifier" stage). Lung segmentation was achieved using 3D U-NET algorithm; nodule detection was done using 3D Retina-UNET and classifier stage with a support vector machine algorithm on selected features. Performances were assessed using area under receiver operating characteristics curve (AUROC).

RESULTS

The pipeline showed good performance for pathological nodule detection and patient diagnosis. With the preparation dataset, an AUROC of 0.9058 (95% confidence interval [CI]: 0.8746-0.9362) was obtained, 87% yielding accuracy (95% CI: 84.83%-91.03%) for the "nodule detection" stage, corresponding to 86% specificity (95% CI: 82%-92%) and 89% sensitivity (95% CI: 84.83%-91.03%).

CONCLUSION

A fully functional pipeline using 3D U-NET, 3D Retina-UNET and classifier stage with a support vector machine algorithm was developed, resulting in high capabilities for pulmonary nodule classification.

Radiological assessment after stereotactic body radiation of lung tumours

The increasing use of stereotactic body radiation therapy for lung tumours comes along with new post-therapeutic imaging findings that should be known by physicians involved in patient follow-up. Radiation-induced lung injury is much more frequent than after conventional radiation therapy, it can also be delayed and has a different radiological presentation. Radiation-induced lung injury after stereotactic body radiation therapy involves the lung parenchyma surrounding the target tumour and appears as a dynamic process continuing for years after completion of the treatment. Thus, the radiological pattern and the severity of radiation-induced lung injury are prone to changes during follow-up, which can make it difficult to differentiate from local recurrence. Contrary to radiation-induced lung injury, local recurrence after stereotactic body radiation therapy is rare. Other complications mainly depend on tumour location and include airway complications, rib fractures and organizing pneumonia. The aim of this article is to provide a wide overview of radiological changes occurring after SBRT for lung tumours. Awareness of changes following stereotactic body radiation therapy should help avoiding unnecessary interventions for pseudo tumoral presentations.

Incidental diagnosis of Covid-19 pneumonia on chest computed tomography

PURPOSE

The purpose of this study was to determine the prevalence and imaging characteristics of incidentally diagnosed COVID-19 pneumonia on computed tomography (CT).

MATERIALS AND METHODS

This retrospective study was conducted between March 20th and March 31st, 2020 at Cochin hospital, Paris France. Thoracic CT examinations of all patients referred for another reason than a suspicion of SARS-CoV-2 infection were reviewed. CT images were analyzed by a chest radiologist to confirm the presence of findings consistent with COVID-19 pneumonia and quantify disease extent. Clinical and biological data (C-reactive protein serum level [CRP] and white blood cell count) of patients with CT findings suggestive for COVID-19 pneumonia were retrieved from the electronic medical chart.

RESULTS

During the study period, among 205 diagnostic CT examinations, six examinations (6/205, 3%) in 6 different patients (4 men, 2 women; median age, 57 years) revealed images highly suggestive of COVID-19 pneumonia. The final diagnosis was confirmed by RT-PCR. Three inpatients were suspected of extra thoracic infection whereas three outpatients were either fully asymptomatic or presented with fatigue only. All had increased CRP serum level and lymphopenia. Disease extent on CT was mild to moderate in 5/6 patients (83%) and severe in 1/6 patient (17%).

CONCLUSION

Cumulative incidence of fortuitous diagnosis if COVID-19 pneumonia did not exceed 3% during the highest pandemic phase and was predominantly associated with limited lung involvement.

COVID-19 pneumonia: A review of typical CT findings and differential diagnosis

The standard of reference for confirming COVID-19 relies on microbiological tests such as real-time polymerase chain reaction (RT-PCR) or sequencing. However, these tests might not be available in an emergency setting. Computed tomography (CT) can be used as an important complement for the diagnosis of COVID-19 pneumonia in the current epidemic context. In this review, we present the typical CT features of COVID-19 pneumonia and discuss the main differential diagnosis.

Optimization of acquisition parameters for reduced-dose thoracic CT: A phantom study

PURPOSE

The purpose of this study was to analyze the impact of different options for reduced-dose computed tomography (CT) on image noise and visibility of pulmonary structures in order to define the best choice of parameters when performing ultra-low dose acquisitions of the chest in clinical routine.

MATERIALS AND METHODS

Using an anthropomorphic chest phantom, CT images were acquired at four defined low dose levels (computed tomography dose index [CTDI_vol]=0.15, 0.20, 0.30 and 0.40mGy), by changing tube voltage, pitch factor, or rotation time and adapting tube current to reach the predefined CTDI_vol-values. Images were reconstructed using two different levels of iteration (adaptive statistical iterative reconstruction [ASIR®]-v70% and ASIR®-v100%). Signal-to-noise ratio (SNR) as well as contrast-to-noise ratio (CNR) was calculated. Visibility of pulmonary structures (bronchi/vessels) were assessed by two readers on a 5-point-Likert scale.

RESULTS

Best visual image assessments and CNR/SNR were obtained with high tube voltage, while lowest scores were reached with lower pitch factor followed by high tube current. Protocols favoring lower pitch factor resulted in decreased visibility of bronchi/vessels, especially in the periphery. Decreasing radiation dose from 0.40 to 0.30mGy was not associated with a significant decrease in visual scores (P<0.05), however decreasing radiation dose from 0.30mGy to 0.15mGy was associated with a lower visibility of most of the evaluated structures (P<0.001). While image noise could be significantly reduced when ASIR®-v100% instead of ASIR®-v70% was used, the visibility-scores of pulmonary structures did not change significantly.

CONCLUSION

Favoring high tube voltage is the best option for reduced-dose protocols. A decrease of SNR and CNR does not necessarily go along with reduced visibility of pulmonary structures.

Nodular reverse halo sign in active pulmonary tuberculosis: A rare CT feature?

PURPOSE

The purpose of this study was to investigate the prevalence of the nodular reverse halo sign (NRHS) in chest computed tomography (CT) in patients with active pulmonary tuberculosis.

MATERIALS AND METHODS

From March 2018 to March 2019, 29 consecutive patients with a culture-confirmed active pulmonary tuberculosis and who underwent chest CT examination during hospital-admission were retrospectively included in the study. There were 24 men and 5 women with a mean age of 40.9±16.7 (SD) years (range: 18-80years). Chest CT examinations of included patients were evaluated for the presence of NRHS and other tuberculosis-related CT signs.

RESULTS

CT revealed the NRHS in 5 patients (5/29; 17%). The other CT signs of tuberculosis included consolidations in 18 patients (18/29; 62%), tree-in-bud pattern in 14 patients (14/29; 48%), cavitation in 12 patients (12/29; 41%), sparse nodules in 10 patients (10/29; 34%), and pleural effusion in 8 patients (8/29; 28%).

CONCLUSION

CT shows NRHS in 17% of patients with active pulmonary tuberculosis, indicating that the sign is not as rare as previously thought in patients with this condition.

Combined pulmonary fibrosis and emphysema in systemic sclerosis: A syndrome associated with heavy morbidity and mortality

BACKGROUND

The syndrome of combined pulmonary fibrosis and emphysema (CPFE) primarily due to tobacco smoking has been reported in connective tissue disease, but little is known about its characteristics in systemic sclerosis (SSc).

METHODS

In this retrospective multi-center case-control study, we identified 36 SSc patients with CPFE, and compared them with 72 SSc controls with interstitial lung disease (ILD) without emphysema.

RESULTS

Rate of CPFE in SSc patients with CT scan was 3.6%, and 7.6% among SSc patients with ILD. CPFE-SSc patients were more likely to be male (75 % vs 18%, p < 0.0001), smokers (83 % vs 33%, p < 0.0001), and to have limited cutaneous SSc (53 % vs 24% p < 0.01) than ILD-SSc controls. No specific autoantibody was significantly associated with CPFE. At diagnosis, CPFE-SSc patients had a greater decrease in carbon monoxide diffusing capacity (DLCO 39 ± 13 % vs 51 ± 12% of predicted value, p < 0.0001) when compared to SSc-ILD controls, whereas lung volumes (total lung capacity and forced vital capacity) were similar. During follow-up, CPFE-SSc patients more frequently developed precapillary pulmonary hypertension (PH) (44 % vs 11%, p < 10^-4), experienced more frequent unscheduled hospitalizations (50 % vs 25%, p < 0.01), and had decreased survival (p < 0.02 by Kaplan-Meier survival analysis) as compared to ILD-SSc controls.

CONCLUSIONS

The CPFE syndrome is a distinct pulmonary manifestation in SSc, with higher morbidity and mortality. Early diagnosis of CPFE by chest CT in SSc patients (especially smokers) may result in earlier smoking cessation, screening for PH, and appropriate management.

[Bronchiectasis imaging]

Bronchiectasis are defined as an irreversible focal or diffuse dilatation of the bronchi and can be associated with significant morbidity. The prevalence is currently increasing, probably due to an increased use of thoracic computed tomography (CT). Indeed, the diagnosis relies on imaging and chest CT is the gold standard technique. The main diagnosis criterion is an increased bronchial diameter as compared to that of the companion artery. However, false positives are possible when the artery diameter is decreased, which is called pseudo-bronchiectasis. Other features such as the lack of bronchial tapering, and visibility of bronchi within 1cm of the pleural surface are also diagnostic criteria, and other CT features of bronchial disease are commonly seen. Thoracic imaging also allows severity assessment and long-term monitoring of structural abnormalities. The distribution pattern and the presence of associated findings on chest CT help identifying specific causes of bronchiectasis. Lung MRI and ultra-low dose CT and are promising imaging modalities that may play a role in the future. The objectives of this review are to describe imaging features for the diagnosis and severity assessment of bronchiectasis, to review findings suggesting the cause of bronchiectasis, and to present the new developments in bronchiectasis imaging.

[Bronchiectasis in adults: Etiological diagnosis]

Bronchiectasis is defined by a permanent and irreversible increase in airway caliber. Computed tomography (CT) scan is required for diagnosis and allows differentiating localized (affecting only one lobe) from diffuse (affecting two or more lobes) bronchiectasis. Localized bronchiectasis is usually related to a local cause (e.g., tumor, compression by lymph nodes, foreign body) whereas there are multiple causes of diffuse bronchiectasis. Main causes of diffuse bronchiectasis can be grouped into 5 categories: (1) immune deficiencies; (2) genetic disorders responsible for abnormal mucociliairy clearance (e.g., cystic fibrosis and primary ciliary dyskinesia); (3) autoinflammatory and connective tissue diseases (e.g., rheumatoid arthritis, Sjogren syndrome); (4) chronic airway inflammatory disorders (e.g., allergic broncho-pulmonary aspergillosis) or chronic infections (e.g., by nontuberculous mycobacteria) and; (5) sequelae of pulmonary insults (e.g., infection, radiotherapy). In the absence of a specific cause, bronchiectasis are called idiopathic. Bronchoscopy is the cornerstone for etiological diagnosis in localized bronchiectasis. Investigation of diffuse bronchiectasis includes standardized biological tests, eventually completed by specialized explorations (e.g., analysis of epithelial ion transport and/or ciliary beating). After this systematic approach, an etiological diagnosis is obtained in approximately 60% of cases, often resulting in therapeutic consequences and in some cases in genetic counselling.

[Ageing with cystic fibrosis: Classical and emerging comorbidities in adults with cystic fibrosis]

Cystic fibrosis (CF) is a genetic disease with pulmonary involvement being predominant and often leading to respiratory failure and premature death. Non-pulmonary complications related to CF transmembrane conductance regulator (CFTR) defect are numerous and account for significant morbidity (exocrine pancreatic insufficiency, diabetes, CF-related liver disease, chronic sinusitis, osteoporosis). Improvement in patients' care led to a continuous increase in life expectancy, with a subsequent increase in the number of adult CF patients worldwide. Increased life expectancy comes with increased prevalence of CF-related comorbidities, but also with new emerging complications directly related to ageing (chronic kidney disease, cardiovascular risk factors, cancers). CFTR modulators might also contribute to modify the face of CF epidemiology and prognosis. Ageing with CF has become a challenge for CF patients and caregivers. This review summarizes classic and emerging comorbidities in the context of current growth and ageing of the CF population. It also addresses potential roles of CFTR modulators.

[Acute pulmonary and non-pulmonary complications in adults with cystic fibrosis]

INTRODUCTION

Cystic fibrosis (CF) is a genetic disease primarily affecting the lungs, which could lead to chronic respiratory failure and premature death. CF patients are usually followed in specialized centers, but may present outside of these centers when they seek care for acute pulmonary and/or non-pulmonary complications. The aim of this paper is to provide appropriate knowledge necessary for managing respiratory and non-respiratory emergencies in CF adults.

METHODS

The review is based on international guidelines, extensive search of the available literature using Pubmed, and experience of the CF reference center at Cochin hospital (Paris, France). Complications occurring after solid organ transplantation (e.g., lung and/or liver) are excluded from this review.

RESULTS

Main acute respiratory complications are pulmonary exacerbations, hemoptysis, pneumothorax and allergic bronchopulmonary aspergillosis. Acute non-respiratory complications include hyponatremic dehydration, acute pancreatitis, acute complications of gallstones, distal intestinal obstruction syndrome, symptomatic nephrolithiasis, acute kidney injury, drug intolerances and catheter-related acute complications.

CONCLUSION

This review summarizes acute pulmonary and non-pulmonary complications occurring in adults with CF, focusing on diagnosis and principles of treatment, with the aim of providing a reference that can be used in clinical practice.

[New TNM classification of non-small cell lung cancer]

Initial staging is a key part of the initial evaluation of non-small cell lung cancer. It relies on the 7th edition of the TNM classification. Proposals have been recently developed for the 8th edition of the classification, which is due to be enacted in early 2017. Among these proposals, the weight of tumor size has been increased and new N descriptors have been introduced to further describe N category depending on the number station involved. Regarding M descriptors, oligometastatic disease is distinguished from multiple distant extrathoracic metastases.

Lung cancer screening: Current status and perspective

Lung cancer screening by low-dose computed tomography (CT) has been proven to reduce lung cancer-related mortality by 20%, in the National Lung cancer Screening Trial (NLST). Low-dose CT acquisition protocols should result in effective dose not exceeding 1.5mSv, and should be read by radiologists who have an expertise in the field. A volumetry-based approach similar to that used in the NELSON study allows reducing the rate of positive screens around 2%. The major harm of screening is overdiagnosis, which correspond to the detection of indolent tumours that would not cause the subject death. Ongoing research for the development of useful biomarkers to be combined to imaging could improve the accuracy of lung cancer screening.

DIPNECH: when to suggest this diagnosis on CT

Diffuse idiopathic pulmonary neuroendocrine cell hyperplasia (DIPNECH) is an under-recognized disease characterized by proliferation of neuroendocrine cells in the bronchial wall. It is considered a pre-invasive lesion for lung carcinoid tumours and is found in 5.4% of patients undergoing surgical resection for lung carcinoid tumours. Other manifestations of DIPNECH include bronchial obstruction and formation of tumorlets. DIPNECH preferentially affects middle-aged women. Patients are either asymptomatic or present with long-standing dyspnoea due to obstructive syndrome that can be mistaken for asthma. At CT, mosaic attenuation with multiple small nodules is very suggestive of DIPNECH. The aim of this review is to describe DIPNECH-related CT features and correlate them with histology, in order to help radiologists suggest this diagnosis and distinguish DIPNECH from other causes of mosaic perfusion.