Search Results (162)

This paper aims to provide a structured analysis of the performance metrics used in computer-aided detection (CAD) systems, specifically focusing on lung nodule detection in computed tomography (CT) images. By examining key metrics along with their respective strengths and limitations, this study offers guidelines to assist in selecting appropriate metrics. Evaluation methods for CAD systems for lung nodule detection are primarily categorized into per-scan and per-nodule approaches. For per-scan analysis, a key metric is the area under the receiver operating characteristic (ROC) curve (AUROC), which evaluates the ability of the system to distinguish between scans with and without nodules. For per-nodule analysis, the nodule-level sensitivity at fixed false positives per scan is often used, supplemented by the free-response receiver operating characteristic (FROC) curve and the competition performance metric (CPM). However, the CPM does not provide normalized scores because it theoretically ranges from zero to infinity and largely varies depending on the characteristics of the data. To address the advantages and limitations of ROC and FROC curves, an alternative FROC (AFROC) was introduced to combine the strengths of both per-scan and per-nodule analyses. This paper discusses the principles of each metric and their relative strengths, providing insights into their clinical implications and practical utility. Full article

Necrotizing fasciitis, which is categorized as a medical and surgical emergency, is a life-threatening soft tissue infection. Necrotizing fasciitis diagnosis primarily relies on computed tomography (CT), magnetic resonance imaging (MRI), ultrasound scans, surgical biopsy, blood tests, and expert knowledge from doctors or nurses. Necrotizing fasciitis develops rapidly, making early diagnosis crucial. With the rapid progress of information technology and systems, in terms of both hardware and software, deep learning techniques have been employed to address problems in various fields. This study develops an information system using convolutional neural networks (CNNs), Optuna, and digital images (CNNOPTDI) to detect necrotizing fasciitis. The determination of the hyperparameters in convolutional neural networks plays a critical role in influencing classification performance. Therefore, Optuna, an optimization framework for hyperparameter selection, is utilized to optimize the hyperparameters of the CNN models. We collect the images for this study from open data sources such as Open-i and Wikipedia. The numerical results reveal that the developed CNNOPTDI system is feasible and effective in identifying necrotizing fasciitis with very satisfactory classification accuracy. Therefore, a potential future application of the CNNOPTDI system could be in remote medical stations or telemedicine settings to assist with the early detection of necrotizing fasciitis. Full article

Background: In recent years, implant-assisted removable partial dentures (IARPDs) have been used clinically. However, the extent to which additional implants reduce the burden of supporting tissues is unclear. The aim of this study was therefore to investigate the influence of implanted IARPDs on stress sharing among supporting tissues, using finite element (FE) analysis. Methods: FE models were constructed based on the computed tomography (CT) of a patient with a unilateral defect of the mandibular premolars and molars and the surface data of an RPD with cuspids as abutments, designed using computer-aided design software. A titanium implant was placed in the area equivalent to the first premolar, second premolar, or first molar (IARPD4, IARPD5, and IARPD6, respectively). FE analysis was performed for laterally symmetrical models, i.e., bilateral distal free-end IARPDs. A vertical load of 200 N was applied to the central fossa of the artificial premolars or molars (L4, L5, or L6). Results: Equivalent stress in the alveolar mucosa and vertical displacement of the denture was smaller, with IARPDs under L5 and L6 loads, compared to RPDs. However, abutment teeth were displaced upward under an L6 load in the IARPD5 model. Conclusions: Within the limitations of this study, the area corresponding to the first molar was recommended as the location for an implant under the denture base of bilateral distal free-end IARPDs. Implants located in the area corresponding to the second premolar may apply non-physiological extrusion force on abutment teeth under the load on the artificial second molar. Full article

Medical imaging plays a pivotal role in diagnostic medicine with technologies like Magnetic Resonance Imagining (MRI), Computed Tomography (CT), Positron Emission Tomography (PET), and ultrasound scans being widely used to assist radiologists and medical experts in reaching concrete diagnosis. Given the recent massive uplift in the storage and processing capabilities of computers, and the publicly available big data, Artificial Intelligence (AI) has also started contributing to improving diagnostic radiology. Edge computing devices and handheld gadgets can serve as useful tools to process medical data in remote areas with limited network and computational resources. In this research, the capabilities of multiple platforms are evaluated for the real-time deployment of diagnostic tools. MRI classification and segmentation applications developed in previous studies are used for testing the performance using different hardware and software configurations. Cost–benefit analysis is carried out using a workstation with a NVIDIA Graphics Processing Unit (GPU), Jetson Xavier NX, Raspberry Pi 4B, and Android phone, using MATLAB, Python, and Android Studio. The mean computational times for the classification app on the PC, Jetson Xavier NX, and Raspberry Pi are 1.2074, 3.7627, and 3.4747 s, respectively. On the low-cost Android phone, this time is observed to be 0.1068 s using the Dynamic Range Quantized TFLite version of the baseline model, with slight degradation in accuracy. For the segmentation app, the times are 1.8241, 5.2641, 6.2162, and 3.2023 s, respectively, when using JPEG inputs. The Jetson Xavier NX and Android phone stand out as the best platforms due to their compact size, fast inference times, and affordability. Full article

Background/Objectives: Endometriosis is characterized by the presence of ectopic endometrial-like glands and stroma outside the endometrial cavity, which mainly occurs in the pelvic cavity. Pulmonary endometriosis, or thoracic endometriosis syndrome (TES), describes the rare presence of endometrial-like cells in the thoracic cavity and includes catamenial pneumothorax, catamenial hemothorax, hemoptysis, and lung nodules. Our aim is to summarize the results of all reported cases of TES. Methods: Extensive research was conducted through MEDLINE/PUBMED using the keywords “thoracic endometriosis”, “thoracic endometriosis syndrome”, “catamenial pneumothorax”, “catamenial hemoptysis”, and “TES”. Following PRISMA guidelines, all published cases of TES between January 1950 and March 2024 were included. A systematic review of 202 studies in English, including 592 patients, was performed. Results: The median age of women with TES is 33.8 years old. The most common clinical presentation is catamenial pneumothorax (68.4%), while lesions are mainly found in the right lung unilaterally (79.9%). Chest computed tomography (CT) was used alone or after an X-ray to determine the pathological findings. Ground-glass opacity nodules and cystic lesions represent the most common finding in CT, while pneumothorax is the most common finding in X-rays. Video-assisted thoracoscopic surgery (VATS) is the main therapeutic approach, usually in combination with hormonal therapy, including GnRH analogues, progestins, androgens, or combined oral contraceptives. Hormonal therapy was also administered as monotherapy. Symptom recurrence was reported in 10.1% of all cases after the treatment. Conclusions: High clinical awareness and a multidisciplinary approach are necessary for the best clinical outcome for TES patients. More studies are required to extract safer conclusions. Full article

Objectives: To explore the factors affecting the lymph node metastasis (LNM) detection performance of prostate-specific membrane antigen positron emission tomography/computed tomography (PSMA PET/CT) and to evaluate its prognostic value for biochemical recurrence after radical prostatectomy (RP). Methods: Patients who had intermediate- or high-risk prostate cancer and underwent robot-assisted (RA)RP between 2017 and 2021 were included. Initial lymph node staging was carried out using PSMA PET/CT. Sensitivity, specificity, and positive (PPV) and negative (NPV) predictive values were calculated. A cut-off value for LNM tumor deposit size that maximizes specificity was investigated and a post hoc specificity analysis was carried out. In survival analysis for biochemical progression-free survival (bPFS) after RP, Kaplan–Meier curves of molecular imaging (mi)N0 and miN1 patients were compared using the log-rank test and separate Cox regression models were developed to reveal the significance of PSMA PET/CT staging in pre- and post-surgery settings. Results: In 583 patients with a prevalence of pathology-proven LNM of 27.4%, overall sensitivity, specificity, PPV, and NPV of PSMA PET/CT per patient were 26.3% [95%CI 18.9–35.5], 93.9% [95%CI 84.9–100], 61.8% [95%CI 44.5–83.5], and 77.1% [95%CI 69.7–85.1], respectively. PSMA PET/CT showed a better sensitivity as LNM tumor deposit size increased (p = 0.003 OR 2.4 [95%CI 1.3–4.4]) and a better specificity in pT3-4 tumors (96.1%) versus pT2 (91.1%, p = 0.024 OR 2.7 [95%CI 1.1–6.3]). After adjustment according to 5.5 mm LNM tumor deposit size, which showed the best discriminative performance (AUC: 0.905 [95%CI 0.804–1.000, p < 0.001]), overall sensitivity tripled (90.2%, p < 0.001). The 1-year bPFS was 56.0% and 83.3% for miN1 and miN0 patients, respectively (p < 0.001). Whereas miN0pN1 was not, miN1pN1 disease was independently associated with decreased bPFS (HR:2.1 95%CI 1.3–3.4, p < 0.001). Conclusions: PSMA PET/CT has a lymph node tumor burden-dependent and cohort-driven diagnostic ability but consequently a strong independent prognostic value for predicting biochemical recurrence after RARP. Full article

The aim of this study is to examine the prognostic role of initial chest computed tomography severity score index (CTSS) and its association with demographic, socio-epidemiological, and clinical parameters in COVID-19 hospitalized patients. A retrospective study included patients who were hospitalized in the COVID Hospital of the Clinical Hospital Center Kosovska Mitrovica from July 2020 to March 2022. We compared patient characteristics and outcome of their hospital stay with values of CT severity score (mild, moderate, and severe form of the disease). Patients with severe disease were statistically significantly older, they treated more days, and they presented statistically significant highest mortality rate compared to mild and moderate forms. Smokers and obese were significantly more frequent among patients with higher CT, while vaccinated patients were more common among those with a mild form. Biochemical parameters at admission also showed statistical significance between the examined groups. We can conclude that by employing the initial CT severity score as the strongest predictor of mortality, it is possible to predict the outcome in hospitalized patients. A comprehensive examination of the patient upon admission, including determining the extent of inflammatory changes in the lungs using computed tomography, the levels of oxygen saturation, and other laboratory parameters, can assist doctors in making an adequate clinical evaluation and apply appropriate therapeutic protocols in the treatment of COVID-19. Full article

The integration of deep learning into radiology has the potential to enhance diagnostic processes, yet its acceptance in clinical practice remains limited due to various challenges. This study aimed to develop and evaluate a fine-tuned large language model (LLM), based on Llama 3-8B, to automate the generation of accurate and concise conclusions in magnetic resonance imaging (MRI) and computed tomography (CT) radiology reports, thereby assisting radiologists and improving reporting efficiency. A dataset comprising 15,000 radiology reports was collected from the University of Medicine and Pharmacy of Craiova’s Imaging Center, covering a diverse range of MRI and CT examinations made by four experienced radiologists. The Llama 3-8B model was fine-tuned using transfer-learning techniques, incorporating parameter quantization to 4-bit precision and low-rank adaptation (LoRA) with a rank of 16 to optimize computational efficiency on consumer-grade GPUs. The model was trained over five epochs using an NVIDIA RTX 3090 GPU, with intermediary checkpoints saved for monitoring. Performance was evaluated quantitatively using Bidirectional Encoder Representations from Transformers Score (BERTScore), Recall-Oriented Understudy for Gisting Evaluation (ROUGE), Bilingual Evaluation Understudy (BLEU), and Metric for Evaluation of Translation with Explicit Ordering (METEOR) metrics on a held-out test set. Additionally, a qualitative assessment was conducted, involving 13 independent radiologists who participated in a Turing-like test and provided ratings for the AI-generated conclusions. The fine-tuned model demonstrated strong quantitative performance, achieving a BERTScore F1 of 0.8054, a ROUGE-1 F1 of 0.4998, a ROUGE-L F1 of 0.4628, and a METEOR score of 0.4282. In the human evaluation, the artificial intelligence (AI)-generated conclusions were preferred over human-written ones in approximately 21.8% of cases, indicating that the model’s outputs were competitive with those of experienced radiologists. The average rating of the AI-generated conclusions was 3.65 out of 5, reflecting a generally favorable assessment. Notably, the model maintained its consistency across various types of reports and demonstrated the ability to generalize to unseen data. The fine-tuned Llama 3-8B model effectively generates accurate and coherent conclusions for MRI and CT radiology reports. By automating the conclusion-writing process, this approach can assist radiologists in reducing their workload and enhancing report consistency, potentially addressing some barriers to the adoption of deep learning in clinical practice. The positive evaluations from independent radiologists underscore the model’s potential utility. While the model demonstrated strong performance, limitations such as dataset bias, limited sample diversity, a lack of clinical judgment, and the need for large computational resources require further refinement and real-world validation. Future work should explore the integration of such models into clinical workflows, address ethical and legal considerations, and extend this approach to generate complete radiology reports. Full article

In clinical settings, computed tomography (CT), magnetic resonance imaging (MRI), or positron emission tomography (PET) are commonly employed in brain imaging to assist clinicians in determining the type of stroke in patients. However, these modalities are associated with potential hazards or limitations. In contrast, microwave imaging emerges as a promising technique, offering advantages such as non-ionizing radiation, low cost, lightweight, and portability. The primary challenges faced by microwave tomography include the severe ill-posedness of the electromagnetic inverse scattering problem and the time-consuming nature and unsatisfactory resolution of iterative quantitative algorithms. This paper proposes a learning electric field enhancement imaging method (LEFEIM) to achieve quantitative brain imaging based on a microwave tomography system. LEFEIM comprises two cascaded networks. The first, based on a convolutional neural network, utilizes the electric field from the receiving antenna to predict the electric field distribution within the imaging domain. The second network employs the electric field distribution as input to learn the dielectric constant distribution, thereby realizing quantitative brain imaging. Compared to the Born Iterative Method (BIM), LEFEIM significantly improves imaging time, while enhancing imaging quality and goodness-of-fit to a certain extent. Simultaneously, LEFEIM exhibits anti-noise capabilities. Full article

Ischemic stroke (IS) is a severe complication and leading cause of mortality in patients under extracorporeal membrane oxygenation (ECMO). The aim of our narrative review is to summarize the existing evidence and provide a deep examination of the diagnosis and treatment of acute ischemic stroke patients undergoing ECMO support. The incidence rate of ISs is estimated to be between 1 and 8%, while the mortality rate ranges from 44 to 76%, depending on several factors, including ECMO type, duration of support and patient characteristics. Several mechanisms leading to ISs during ECMO have been identified, with thromboembolic events and cerebral hypoperfusion being the most common causes. However, considering that most of the ECMO patients are severely ill or under sedation, stroke symptoms are often underdiagnosed. Multimodal monitoring and daily clinical assessment could be useful preventive techniques. Early recognition of neurological deficits is of paramount importance for prompt therapeutic interventions. All ECMO patients with suspected strokes should immediately receive brain computed tomography (CT) and CT angiography (CTA) for the identification of large vessel occlusion (LVO) and assessment of collateral blood flow. CT perfusion (CTP) can further assist in the detection of viable tissue (penumbra), especially in cases of strokes of unknown onset. Catheter angiography is required to confirm LVO detected on CTA. Intravenous thrombolytic therapy is usually contraindicated in ECMO as most patients are on active anticoagulation treatment. Therefore, mechanical thrombectomy is the preferred treatment option in cases where there is evidence of LVO. The choice of the arterial vascular access used to perform mechanical thrombectomy should be discussed between interventional radiologists and an ECMO team. Anticoagulation management during the acute phase of IS should be individualized after the thromboembolic risk has been carefully balanced against hemorrhagic risk. A multidisciplinary approach is essential for the optimal management of ISs in patients treated with ECMO. Full article

The widespread implementation of lung cancer screening and thin-slice computed tomography (CT) has led to the more frequent detection of small nodules, which are commonly referred to thoracic surgeons. Surgical resection is the final diagnostic and treatment option for such nodules; however, surgeons must perform preoperative or intraoperative markings for the identification of such nodules and their precise resection. Historically, hook-wire marking has been performed more frequently worldwide; however, lethal complications, such as air embolism, have been reported. Therefore, several surgeons have recently attempted to develop novel preoperative and intraoperative markers. For example, transbronchial markings, such as virtual-assisted lung mapping and intraoperative markings using cone-beam computed tomography, have been developed. This review explores various marking methods that have been practically applied for a better understanding of preoperative and intraoperative markings in thoracic surgery. Recently, several attempts have been made to perform intraoperative molecular imaging and dynamic virtual three-dimensional computed tomography for the localization, diagnosis, and margin assessment of small nodules. In this narrative review, the current status and future perspectives of preoperative and intraoperative markings in thoracic surgery are examined for a better understanding of these techniques. Full article

Background: Computed tomography to body divergence (CTBD) is one of the main barriers to bronchoscopic techniques for the diagnosis of peripherally located lung nodules. Cone-beam CT (CBCT) guidance is being rapidly adopted to correct for this phenomenon and to potentially increase diagnostic outcomes. In this trial, we hypothesized that the addition of mobile CBCT (m-CBCT) could improve the rate of tool in lesion (TIL) and the diagnostic yield of shape-sensing robotic-assisted bronchoscopy (SS-RAB). Methods: This was a prospective, single-arm study, which enrolled patients with peripheral lung nodules of 1–3 cm and compared the rate of TIL and the diagnostic yield of SS-RAB alone and combined with mCBCT. Results: A total of 67 subjects were enrolled, the median nodule size was 1.7 cm (range, 0.9–3 cm). TIL was achieved in 23 patients (34.3%) with SS-RAB alone, and 66 patients (98.6%) with the addition of mCBCT (p < 0.0001). The diagnostic yield of SS-RAB alone was 29.9% (95% CI, 29.3–42.3%) and it was 86.6% (95% CI, 76–93.7%) with the addition of mCBCT (p < 0.0001). There were no pneumothoraxes or any bronchoscopy-related complications, and the median total dose–area product (DAP) was 50.5 Gy-cm². Conclusions: The addition of mCBCT guidance to SS-RAB allows bronchoscopists to compensate for CTBD, leading to an increase in TIL and diagnostic yield, with acceptable radiation exposure. Full article

Background: Endoscopic endonasal skull base surgery has become a viable alternative to open procedures for the surgical treatment of benign and malignant lesions in the sinonasal and skull base regions. As in sinus surgery, skull base surgery may cause crusting and posterior rhinorrhea, particularly when a nasoseptal flap is required for skull base reconstruction. Post-operative radiological sinonasal findings have been reported previously with no clear correlation with intraoperative decision-making. As in open surgery, endoscopic surgery is not standardized and there is variability in the intervention to assist with exposure and skull base repair. These modifications, including middle turbinate resection, nasoseptal flap, fat graft, and maxillary antrostomy have the potential for nasal morbidity. The aim of this study was to evaluate whether specific interventions during surgery or specific patient and tumor characteristics harbor a more significant risk of causing nasal morbidity post-operatively, as demonstrated by post-operative imaging. Methods: A retrospective analysis of all patients who underwent endoscopic endonasal skull base surgery for pituitary lesions at two major referral centers was performed. Data on demographic, clinical, and pathological features were collected, and pre- and post-operative imaging studies (computed tomography (CT) and magnetic resonance imaging (MRI)) were reviewed and scored according to the Lund–Mackay (LM) scoring system. Results: The study included 183 patients. Radiographic evidence of sinusitis was observed in 30 patients (LM score > 4) in post-operative imaging studies. Patients who underwent middle turbinectomy or nasoseptal flap were found to have significantly higher LM scores on follow-up imaging. A nasoseptal flap was found to be associated with an average increase in LM score of 1.67 points and middle turbinectomy with an average increase of 2.21 points. There was no correlation between tumor size and findings that were compatible with sinusitis on post-operative imaging. Conclusions: The findings of the present study suggest that endoscopic endonasal skull base surgery is associated with radiological evidence of sinusitis. Nasoseptal flap reconstruction and middle turbinectomy were strongly associated with radiographic sinusitis and should be judiciously performed during surgery. A clinical correlation is needed for further recommendations. Full article

Background/Objectives: Recent advances in artificial intelligence, particularly in cardiac imaging, can potentially enhance patients’ diagnosis and prognosis and identify novel imaging markers. We propose an automated, computer-aided algorithm utilizing native cardiac computed tomography (CT) imaging to identify myocardial fibrosis. This study aims to evaluate its performance compared to CMR markers of fibrosis in a cohort of patients diagnosed with breast cancer. Methods: The study included patients diagnosed with early HER2+ breast cancer, who presented LV dysfunction (LVEF < 50%) and myocardial fibrosis detected on CMR at the time of diagnosis. The patients were also evaluated by cardiac CT, and the extracted images were processed for the implementation of the automatic, computer-assisted algorithm, which marked as fibrosis every pixel that fell within the range of 60–90 HU. The percentage of pixels with fibrosis was subsequently compared with CMR parameters. Results: A total of eight patients (n = 8) were included in the study. High positive correlations between the algorithm’s result and the ECV fraction (r = 0.59, p = 0.126) and native T1 (r = 0.6, p = 0.112) were observed, and a very high positive correlation with LGE of the LV(g) and the LV-LGE/LV mass percentage (r = 0.77, p = 0.025; r = 0.81, p = 0.015). A very high negative correlation was found with GLS (r = −0.77, p = 0.026). The algorithm presented an intraclass correlation coefficient of 1 (95% CI 0.99–1), p < 0.001. Conclusions: The present pilot study proposes a novel promising imaging marker for myocardial fibrosis, generated by an automatic algorithm based on native cardiac CT images. Full article

Kidney cancer is a malignant tumor with a high mortality rate. The accurate segmentation of tumors from computed tomography (CT) scans can assist physicians in clinical diagnosis. We introduced a new segmentation network called DWR-SegFormer to address the challenge of accurately segmenting kidney tumors in CT images. The method involved binarizing the label maps of clear cell renal cell carcinoma and papillary renal cell carcinoma CT images for identification, and the cancer lesion area was obtained by the label so that the model could accurately identify the area and enhance the feature extraction ability. Secondly, an optimized segmentation model combining a DWR attention mechanism and SegFormer network was constructed. MiT-B0 was used as the encoder of the model to establish long-distance feature dependencies and effectively extract feature information at different resolutions. The decoder with a multi-branch DWR module was implemented to utilize multi-scale feature information effectively and enhance segmentation accuracy. Comparing the experimental results with other existing models shows that the model significantly outperformed the comparison methods in all evaluation metrics on the CT image dataset of clear cell renal cancer. Furthermore, the experimental findings highlight the robustness of the proposed model across other datasets. Full article