Clinical application of convolutional neural network lung nodule detection software: An Australian quaternary hospital experience.

INTRODUCTION: Early-stage lung cancer diagnosis through detection of nodules on computed tomography (CT) remains integral to patient survivorship, promoting national screening programmes and diagnostic tools using artificial intelligence (AI) convolutional neural networks (CNN); the software of AI-Rad Companion™ (AIRC), capable of self-optimising feature recognition. This study aims to demonstrate the practical value of AI-based lung nodule detection in a clinical setting; a limited body of research. METHODS: One hundred and eighty-three non-contrast CT chest studies from a single centre were assessed for AIRC software analysis. Prospectively collected data from AIRC detection and characterisation of lung nodules (size: ≥3 mm) were assessed against the reference standard; reported findings of a blinded consultant radiologist. RESULTS: One hundred and sixty-seven CT chest studies were included; 52% indicated for nodule or lung cancer surveillance. Of 289 lung nodules, 219 (75.8%) nodules (mean size: 10.1 mm) were detected by both modalities, 28 (9.7%) were detected by AIRC alone and 42 (14.5%) by radiologist alone. Solid nodules missed by AIRC were larger than those missed by radiologist (11.5 mm vs 4.7 mm, P < 0.001). AIRC software sensitivity was 87.3%, with significant false positive and negative rates demonstrating 12.5% specificity (PPV 0.6, NPV 0.4). CONCLUSION: In a population of high nodule prevalence, AIRC lung nodule detection software demonstrates sensitivity comparable to that of consultant radiologist. The clinical significance of larger sized nodules missed by AIRC software presents a barrier to current integration in practice. We consider this research highly relevant in providing focus for ongoing software development, potentiating the future success of AI-based tools within diagnostic radiology.

Impact of Different Artificial Intelligence User Interfaces on Lung Nodule and Mass Detection on Chest Radiographs.

Purpose: To explore the impact of different user interfaces (UIs) for artificial intelligence (AI) outputs on radiologist performance and user preference in detecting lung nodules and masses on chest radiographs. Materials and Methods: A retrospective paired-reader study with a 4-week washout period was used to evaluate three different AI UIs compared with no AI output. Ten radiologists (eight radiology attending physicians and two trainees) evaluated 140 chest radiographs (81 with histologically confirmed nodules and 59 confirmed as normal with CT), with either no AI or one of three UI outputs: (a) text-only, (b) combined AI confidence score and text, or (c) combined text, AI confidence score, and image overlay. Areas under the receiver operating characteristic curve were calculated to compare radiologist diagnostic performance with each UI with their diagnostic performance without AI. Radiologists reported their UI preference. Results: The area under the receiver operating characteristic curve improved when radiologists used the text-only output compared with no AI (0.87 vs 0.82; P < .001). There was no difference in performance for the combined text and AI confidence score output compared with no AI (0.77 vs 0.82; P = .46) and for the combined text, AI confidence score, and image overlay output compared with no AI (0.80 vs 0.82; P = .66). Eight of the 10 radiologists (80%) preferred the combined text, AI confidence score, and image overlay output over the other two interfaces. Conclusion: Text-only UI output significantly improved radiologist performance compared with no AI in the detection of lung nodules and masses on chest radiographs, but user preference did not correspond with user performance.Keywords: Artificial Intelligence, Chest Radiograph, Conventional Radiography, Lung Nodule, Mass Detection© RSNA, 2023.

Strategies for the Management of Solitary Pulmonary Nodules: A Survey of Patient Preferences.

BACKGROUND: Guideline-based strategies for evaluation of solitary pulmonary nodules are tailored to the likelihood of malignancy. Surveillance, biopsy, and resection are all reasonable approaches in fit individuals when the likelihood of malignancy is intermediate. Given the paucity of data demonstrating superior outcomes and important trade-offs among strategies, guidelines emphasize the importance of eliciting patient preferences and engaging in shared decision making; however, there is little literature on what patient preferences actually are. METHODS: This study conducted a cross-sectional, interview-administered questionnaire survey in 100 adults recruited from a metropolitan teaching hospital (The Royal Melbourne Hospital, Parkville, Victoria, Australia). The questionnaire used a hypothetical scenario designed to elicit patient preferences for different management strategies of solitary pulmonary nodules with a probability of malignancy between 10% and 70%. RESULTS: The mean age of participants was 62 years (range, 45 to 80 years), 56% were male, and 94% were current smokers or ex-smokers. Ninety-four percent completed all questions. At 10% probability of malignancy, 36.3% preferred surveillance, 42.4% preferred needle biopsy, and 21.2% preferred surgical resection. Preference for surgical resection increased to 53.5% and 86.2% when the probability of malignancy was 30% and 70%, respectively. Changes in the diagnostic yield of computed tomography biopsy significantly altered preferences when the probability of malignancy was 10% or 30%. CONCLUSIONS: The majority of participants surveyed expressed a preference for some type of biopsy over observation at all levels of solitary pulmonary nodule probability of malignancy evaluated. In an era of increasing solitary pulmonary nodule detection and patient-centered care, if confirmed in broader studies, these findings have considerable implications for processes of care and resource allocation.

Chest imaging findings in COVID-19-positive patients in an Australian tertiary hospital.

INTRODUCTION: Coronavirus disease 2019 (COVID-19) has infected over 215 million individuals worldwide. Chest radiographs (CXR) and computed tomography (CT) have assisted with diagnosis and assessment of COVID-19. Previous reports have described peripheral and lower zone predominant opacities on chest radiographs. Whilst the most common patterns on CT are bilateral, peripheral basal predominant ground glass opacities (Wong et al., Radiology, 296, 2020, E72; Karimian and Azami, Pol J Radiol, 86, 2021, e31). This study describes the imaging findings in an Australian tertiary hospital population. METHODS: COVID-PCR-positive patients who had chest imaging (CXR, CT and ventilation perfusion (V/Q) scans) from January 2020 to August 2020 were included. Distribution, location and pattern of involvement was recorded. Evaluation of the assessors was performed using Fleiss Kappa calculations for review of radiographic findings and qualitative analysis of CT findings. RESULTS: A total of 681 studies (616 CXRs, 59 CTs, 6 V/Q) from 181 patients were reviewed. The most common chest radiograph finding was bilateral lower lobe predominant diffuse opacification and most common CT pattern being ground glass opacities. Of the CT imaging, 33 were CT Pulmonary Angiograms of which five demonstrated acute pulmonary emboli. There was good inter-rater agreement between radiologists in assessment of imaging appearances on CXR (kappa 0.29-0.73) and CT studies. CONCLUSION: A review of imaging in an Australian tertiary hospital demonstrates similar patterns of COVID-19 infection on chest X-ray and CT imaging when compared to the international population.

Assessing signal intensity change on well-registered images: comparing subtraction, color-encoded subtraction, and parallel display formats.

PURPOSE: To assess the sensitivity, specificity, accuracy, and interobserver reliability of subtraction, color-encoded subtraction, and parallel display formats in assessing signal intensity (SI) differences between well-registered images. MATERIALS AND METHODS: Institutional ethics approval for the study and a waiver of individual patient consent were obtained. Five radiologists graded the severity of fatty liver by using a seven-point scale for four imaging sets created from 179 pairs of dual-echo in- and opposed-phase magnetic resonance images from 179 patients. The four sets contained images displayed in parallel, subtraction images, color-encoded subtraction images, and images from the three previous formats presented together. The order of the images and sets was randomized. Sensitivity and specificity were assessed with the McNemar test. Accuracy was assessed by using three-way analysis of variance, with Tukey post hoc methods used to assess differences between the four formats. Interobserver reliability was assessed by using the Fleiss κ value. RESULTS: Subtraction (P = .016 at a 5% SI difference threshold) and color-encoded subtraction (P = .031 at a 4% SI difference threshold) formats had higher sensitivity than did the parallel format. The accuracy of the subtraction format was superior to that of the parallel format (P < .0001). Interobserver reliability of the subtraction (κ = 0.53) and color-encoded subtraction (κ = 0.39) formats was superior to that of the parallel format (κ = 0.33) (P < .0001 and P = .0085, respectively). CONCLUSION: When images are well registered, subtraction and color-encoded subtraction techniques offer advantages over the traditional parallel presentation format for the assessment of SI differences.

Correlation of positron emission tomography ventilation-perfusion matching with CT densitometry in severe emphysema.

BACKGROUND: Emphysema severity is frequently measured on CT via densitometry. Correlation with scintigraphic and spirometric functional measures of ventilation or perfusion varies widely, and no prior study has evaluated correlation between densitometry and lobar ventilation/perfusion in patients with severe emphysema. The aim of this study was to evaluate the utility and findings of gallium-68 (68Ga) ventilation/perfusion positron emission tomography-CT (68Ga-VQ/PET-CT) in severe emphysema assessment. METHODS: Fourteen consecutive patients undergoing evaluation for bronchoscopic lung volume reduction between March 2015 and March 2018 underwent 68Ga-VQ/PET-CT assessment for lobar functional lung mapping, in addition to CT densitometry. Correlations between CT densitometry and 68Ga-VQ/PET-CT parameters for individual lobar lung function were sought. RESULTS: CT densitometry assessment of emphysema correlated only weakly (R2 = 0.13) with lobar perfusion and was not correlated with ventilation (R2 = 0.04). Densitometry was moderately (R2 = 0.67) correlated with V/Q units in upper lobes, though poorly reflected physiological function in lower lobes (R2 = 0.19). Emphysema severity, as measured by CT densitometry, was moderately correlated with proportion of normal V/Q units and matched V/Q defects in individual lobes. CONCLUSIONS: Assessment of lobar pulmonary function by 68Ga-VQ/PET-CT provides physiologic information not evident on CT densitometry such as ventilation and perfusion specifics and matched defects. Further research is needed to see if the discordant findings on 68Ga-VQ/PET-CT provide prognostic information or can be used to modify patient management and improve outcomes.

Ultralow dose CT for follow-up of solid pulmonary nodules: A pilot single-center study using Bland-Altman analysis.

Solid pulmonary nodules are a common finding requiring serial computed tomography (CT) imaging. We sought to explore the detection and measurement accuracy of an ultralow-dose CT (ULDCT) protocol compared with our standard low-dose CT (LDCT) nodule follow-up protocol.In this pragmatic single-center pilot prospective cohort study, patients scheduled for clinically indicated CT surveillance of 1 or more known solid pulmonary nodules >2 mm underwent ULDCT immediately after routine LDCT. The Bland-Altman 95% limits of agreement for diameter and volumetry were calculated.In all, 57 patients underwent 60 imaging episodes, with 170 evaluable nodules. ULDCT detected all known solid pulmonary nodules >2 mm. Bland-Altman analyses demonstrated clinically agreement for both nodule diameter and volume, both of which fell within prespecified limits.This single-center pilot study suggests that ULDCT may be of use in surveillance of known solid pulmonary nodules >2 mm.

Percutaneous gastrostomy and gastrojejunostomy.

Gastrostomy allows enteral nutrition to continue in patients who are unable to meet their caloric requirements orally. Though the indications for gastrostomy placement are varied, dysphagia secondary to a neurological condition is the most common. These catheters were initially placed surgically, but percutaneous endoscopic placement is now the routine in most centers. Interventional radiologists have been performing this procedure under fluoroscopic guidance for several years with encouraging results. Percutaneous radiological gastrostomy is reported to have a success rate comparable to that of the endoscopic method, with lower morbidity and mortality rates. A further benefit is that it may be performed in patients for whom the endoscopic method would be difficult or dangerous, such as those with head and neck malignancies. One of the main factors currently limiting the use of this procedure is the shortage of interventional radiology facilities and specialists.This article describes a technique for routine percutaneous radiological gastrostomy catheter placement and procedural variations for difficult cases. Indications and contraindications will be discussed, as will complication rates and how these compare with the traditional methods of gastrostomy tube placement.