Radiological Imaging Research Articles

Breast Implants: Low Rate of Annual Check-Ups Results in Delayed Presentation of Ruptured Implants.

Background: Breast-implant-based reconstruction is one of the most performed procedures in plastic surgery. Despite the high durability of breast implants, various complications are accompanied with prolonged inlay duration, particularly implant rupture. Many aftereffects can be associated with implant rupture, especially siliconoma and BIA-ALCL. Without regular implant check-ups, implant-related issues may remain underrecognized. Here, we analyzed the number of breast implant carriers needing revisions and if patients adhered to annual implant follow-up recommendations. Methods: We reviewed 1128 breast procedures at the department of plastic surgery at the University Clinic of St. Poelten between August 1st 2018 and December 31st 2023. Patients were analyzed to see whether regular check-ups of their breast implants were performed. Additionally, implant-related complications were investigated, as well as if they were noticed by implant carriers. Results: Only 15.46% of breasts implants were regularly checked at least once a year in our cohort. The remaining 84.54% of patients consulted our department due to pain or aesthetic discomfort without periodical follow-ups. Most implant ruptures (73.8%) were diagnosed in patients consulting acutely due to pain or capsular contraction after an average of 17.36 ± 10.57 years. Routine examination uncovered 26.2% of silent implant ruptures without patients yet complaining of clinical symptoms as early as 15.44 ± 11.17 years. Conclusions: Most implant ruptures develop clinical symptoms as an indicator that removal is warranted. However, only regular follow-ups can identify implant complications several years earlier and possibly reduce severe sequalae such as BIA-ALCL. This highlights the significance and necessity of annual breast implant controls by surgeons and radiologic imaging to prevent devastating implant-associated aftereffects.

Artificial Intelligence in Prostate Cancer Diagnosis

Prostate cancer (PCa) is a cancer with a broad spectrum of biological behavior and it is a heterogeneous nature. In order to prevent overdiagnosis and overtreatment, and to detect clinically significant PCa, standardized scoring and grading systems are used in imaging and pathological examinations. However, reproducibility and agreement between readers in these diagnostic stages, which require experience, are low. Promising results have been achieved by integrating artificial intelligence (AI)-based applications into the diagnosis and management of PCa. In radiological and pathological imaging, computer-aided diagnostic tools have increased clinical efficiency and achieved diagnostic accuracy comparable to that of experienced healthcare professionals. This review provides an overview of AI applications used in radiological imaging, prostate biopsy, and histopathological examination in the diagnosis of PCa.

Choanal Atresia about Two Cases at the Hospital and University Center Gabriel Toure from Bamako (Mali)

Choanal atresia is a congenital defect in the permeabilization of the posterior end of the nasal cavities. Its bilateral form is responsible for respiratory distress that can be life-threatening. The unilateral form was suspected in view of the permanent unilateral right nasal obstruction. We report the observation of two clinical cases admitted to the Radiology and Medical Imaging department, CHU GT of Bamako (Mali). The aim of our study was to highlight the importance of screening for this condition at birth. Surgical treatment allowed us to obtain an overall success rate of 100%. Divulsion was used to pass the emergency stage after short-term palliative medical treatment in the case of bilateral choanal atresia. The operative techniques used were divulsion and nasal endoscopic route.

Radiomics in radiology: What the radiologist needs to know about technical aspects and clinical impact.

Radiomics represents the science of extracting and analyzing a multitude of quantitative features from medical imaging, revealing the quantitative potential of radiologic images. This scientific review aims to provide radiologists with a comprehensive understanding of radiomics, emphasizing its principles, applications, challenges, limits, and prospects. The limitations of standardization in current scientific production are analyzed, along with possible solutions proposed by some of the referenced papers. As the continuous evolution of medical imaging is ongoing, radiologists must be aware of new perspectives to play a central role in patient management.

Evaluation of the relationship between thrombophilia gene mutations and demographic data, imaging findings and stroke subtypes in young patients followed up with a diagnosis of stroke

Aims: A stroke is defined as a focal or generalized neurological deficit that develops suddenly, lasts for more than 24 hours, or ends in death within this period, with a cause other than a vascular cause. Stroke is among the leading causes of morbidity and mortality in the general population. In our study, we examined the prevalence of all prothrombotic gene mutations observed in patients with and without predisposing risk factors for young patients with stroke and discussed their roles in stroke formation in the context of existing literature. Methods: In the present study, we conducted a retrospective review of the medical records of 300 patients aged 18 to 49 years who were hospitalized in the neurology clinic of our hospital between June 2023 and June 2024 and diagnosed with acute arterial ischemic stroke based on anamnesis, neurological examination, and radiological imaging (CT and MRI) results. The study cohort comprised 47 patients (23 males, 24 females) aged between 18 and 49 years with a diagnosis of ischemic stroke. In our study, we examined ischemic stroke-related gene polymorphisms in patients with early-onset ischemic stroke (before the age of 50 years). This study aims to investigate the prevalence of Factor V-Leiden, prothrombin-G20210A, methylenetetrahydrofolate reductase (MTHFR) C677T, MTHFR A1298C, SERPIN 1, and Factor II polymorphisms in young patients with ischemic stroke. Results: A total of 47 patients, comprising 23 males and 24 females, were included in the study. The mean age of the patients was 40.34 ± 6.37 years. According to the TOAST classification system, 17 patients (36.2%) were diagnosed with large artery atherothrombosis, one patient (2.1%) had a cardioembolic infarction, and 29 patients (61.7%) had a small artery occlusion. In our study, when evaluating the thrombophilia mutation subtypes in young ischemic stroke patients, 23.4% (n=11) of the patients were found to be heterozygous for the FVL mutation, while 76.6% (n=36) had no mutation. Concerning the SERPIN 1 mutation, 40.4% (n=19) of the patients were heterozygous, 23.4% (n=11) were homozygous, and 36.2% (n=17) exhibited no mutation. The distribution of patients according to MTHFR C677T mutation revealed that 57.4% (n=27) were heterozygous, 19.1% (n=9) were homozygous, and 23.4% (n=11) had no mutation. Considering the distribution of patients according to the MTHFR A1298C gene mutation, a total of 38.3% (n=18) of the patients were heterozygous, 14.9% (n=7) were homozygous, and 46.9% (n=22) had no mutation. Upon analysis of the distribution of patients according to factor XIII mutation, it was identified that 19.1% (n=9) were heterozygous, 2.1% (n=1) were homozygous, and 78.7% (n=37) of the patients exhibited no mutation. Upon analysis of the distribution of patients according to prothrombin (Factor II) mutation, it was identified that 8.5% (n=4) were heterozygous, while 91.5% (n=43) exhibited no mutation. Conclusion: Although stroke is less prevalent in young adults, the underlying etiology is highly varied. Further research, including the investigation of genetic and prothrombotic mutations, is vital for the prevention of recurrent strokes in young adults.

The Effect of the COVID-19 Pandemic on the Working Conditions and Efficiency of Radiographers at Ahvaz Jundishapur University of Medical Sciences' Radiology and CT Scan Imaging Centers

Background: Radiographers are highly skilled imaging technology users who work directly with COVID-19 cases. Objectives: This study aims to assess the impact of COVID-19 on the working conditions and efficiency of radiographers and their level of satisfaction with personal protective equipment (PPE), cleaning supplies, and facilities during the COVID-19 epidemic. Methods: A prospective observational study was conducted among radiographers registered in the radiology and CT scan imaging centers of Ahvaz Jundishapur University of Medical Sciences and some private centers between June and October 2022. A questionnaire with 28 multiple-choice questions was used to collect information. The questionnaire consisted of two sections; the first part contained demographic information. In the second part, participants answered questions about changes in working conditions, such as changes in stress level, work efficiency, the need for more rest, imaging time, use of equipment, and personal protective equipment at work, based on their experiences and observations during the COVID-19 pandemic. The data analysis was performed using Statistical Product Service Solutions (SPSS) software, version 26.0, with statistical significance assumed at a P-value < 0.05. Results: One hundred and fifty radiographers completed the questionnaire. The mean age and mean work experience of participants were 33.2 ± 8.3 years and 9 ± 7.2 years, respectively. The results showed that most participants of both sexes agreed that they needed more rest (P-value < 0.05); their work efficiency had not decreased significantly during the COVID-19 epidemic. The stress level of participants had significantly increased compared to before the epidemic. There were significant differences between government-owned hospitals and private radiology centers regarding the presence of a COVID-19-related protocol, the compulsory use of masks, the existence of decontamination instructions, the type of air decontamination, the contact with imaging equipment manufacturers, the timing of decontamination, and the type of decontamination applied to the imaging equipment (P-value < 0.05). Conclusions: This study demonstrated that the reduction in work efficiency among radiographers was not significant compared to before the COVID-19 pandemic. According to the study's results, radiographers' working conditions underwent numerous adjustments over the course of the COVID-19 epidemic.

Asbestos is still paid for dearly

Abstract Background In Italy, asbestos is a current and important problem in terms of health effects due to the long latency of diseases related to its exposure. The aim of the study is to analyze the case history of malignant pleural mesothelioma (MPM) in the University Hospital of Siena in the period 2019-2023, comparing the type of exposure (occupational/non-occupational), the mean conventional latencies (MCL) of diagnosis, the histotype and the most exposed categories. Methods The study was conducted from January to April 2024, on 69 patients (79.7% male and 20.3% female) with a diagnosis of pleural mesothelioma between 2019 and 2023. 48 patients lived in Tuscany and 21 outside the region. The data were collected by consulting the medical surveillance records and radiological images through the VuePacs platform, then organized in a filtered database. Results Of 69 patients, exposure was occupational in 54 cases and non-occupational in the remaining 15. In patients with occupational exposure, the MCL was 55.9 years and the mean age at diagnosis was 73.66 years; in the other group, the MCL was 47.3 years and the mean age at diagnosis was 66.8. The study revealed the prevalence of epithelioid histotype in 76.81% of cases, biphasic in 15.94% of cases and sarcomatoid in 7.25%. The most at-risk categories were: metalworking (10 cases), construction (8), electricity/gas production (6) and shipyards (4). The study also identified two cases of atypical occupational exposure. Conclusions The shorter average latency in cases with non-occupational exposure is in accordance with the current scientific literature, as well as the prevalence of epithelioid histotype. Are hereby confirmed the primary role of health surveillance in the in-depth investigation of occupational anamnesis and of the environmental sites’ survey as prevention tool, especially in those cases of MPM whose evaluation, deviating from the exposure standards, could lead to the assignment of a misdiagnosed exposure. Key messages • The effects of asbestos exposure are a current and important problem in our country, even after its disposal, due to the long latency of disease onset. • In-Department investigation of occupational anamnesis and survey of environmental sites as prevention tool play an essential role in MPM evaluation, especially in those cases of atypical exposure.

Unusual associations of bilateral Morgagni hernia in a child with Down’s syndrome: a case report

Morgagni hernia is a rare form of congenital diaphragmatic hernia. It is associated with Down’s syndrome as well as other congenital anomalies. Pediatric patients usually present with respiratory manifestations, the clinical presentation of Morgagni hernia is non-specific. Therefore, Morgagni hernia is mainly diagnosed by radiological imaging. The rarity of this type of hernia along with the vague clinical presentation can lead to missed diagnosis.

Elastic stable intramedullary nailing in paediatric diaphyseal forearm fractures – a retrospective analysis of 201 cases

BackgroundForearm shaft fractures are common injuries, often caused by falling from a fully-upright position or falling off a bike. They can be treated nonoperatively or surgically with intramedullary nailing or plates. The method of choice for treating pediatric forearm shaft fractures is the application of elastic stable intramedullary nailing (ESIN)|. The aim of the study was to evaluate ESIN in pediatric patients with forearm shaft fractures based on radiological images, and determine the etiology and complication rate associated with the injury.MethodsThe study included 201 patients, 30.5% female 69.5% male, aged 1 to 17 years (mean 9.1 years; SD = 3.2), all had been diagnosed with a fracture of the forearm shaft and had been treated surgically with ESIN. In addition, all possessed a complete set of X-ray images and had attended a minimum six-month follow-up examination of the forearm. Axial alignment was evaluated retrospectively in the anatomical (AP) and lateral (LAT) positions. In total, 402 radiographs were examined. Of the injuries, 68% occurred during sports activity and 75% involved both the radius and the ulna.ResultsUnion was observed in all cases. Mean axial alignment values in AP and LAT X-ray or both the ulna and radius were satisfactory. Axial alignment values were not influenced significantly by age, type of surgery, type of fracture or etiology. Plaster cast application (9.8% of cases) significantly influenced radius axial alignment. The complication rate was 11.4% (n = 23). Significantly more complications were observed in patients receiving open reduction internal fixation (ORIF) (p = 0.0025).ConclusionThe ESIN technique is an effective treatment for forearm diaphyseal fractures in children, with good results regarding reduction and bone healing, indicated by x-ray.

Association of hemoglobin, albumin, lymphocyte and platelet (HALP) score with testicular tumor aggressiveness and prognosis.

The hemoglobin, albumin, lymphocyte and platelet (HALP) score integrates readily available blood markers that reflect systemic inflammation, nutritional status, and immune response, all of which can influence cancer progression. This study investigated the association between the HALP score and clinicopathological characteristics in patients with testicular tumor. Data of patients who underwent radical orchiectomy for testicular tumors between January 2020 and January 2024 were reviewed. Preoperative serum tumor markers, hemogram parameters and albumin levels were recorded. Tumor stages were recorded from postoperative radiological imaging and serum tumor markers. The association between postoperative results and HALP score was analyzed. A total of 74 male patients were included in the study. The mean age of the patients was 30.27 ± 6.42 years. The mean HALP score in the patient group with metastasis and retroperitoneal lypmh node invasion (RPLNI) was statistically significantly lower than the patients without metastasis and RPLNI. HALP score decreased statistically significantly with increasing tumor T stage, N stage and M stage. In addition, the mean HALP score values of patients who received chemotherapy, developed progression and mortality were statistically significantly lower than those of patients who did not. Lower HALP scores are significantly associated with advanced disease and poorer prognosis in patients with testicular tumor. The HALP score, composed of routinely measured blood markers, may serve as a convenient and cost-effective prognostic tool to identify patients at higher risk and guide personalized management strategies.

A rare case of trapezium osteochondroma: unveiling a unique tumour with clinical implications

Osteochondroma is a benign bone tumour commonly found on the metaphysis of long bones. The presentation of osteochondroma on the trapezoid bone, a carpal bone in the wrist, is exceptionally rare. To our knowledge, no such cases have been reported in India. This report details a rare case of osteochondroma located on the trapezoid bone and its successful surgical management. A 55-year-old female presented with pain and swelling in the wrist. Radiological imaging, including X-rays and MRI, revealed a well-defined, pedunculated mass arising from the trapezoid bone. The patient underwent surgical resection of the tumour. Histopathological examination confirmed the diagnosis of osteochondroma. Postoperatively, the patient experienced significant relief from symptoms and showed no evidence of tumour recurrence during follow-up. This case contributes to the limited body of literature on osteochondromas in rare locations, particularly on the trapezoid bone. It illustrates the successful outcome of surgical treatment for rare bone tumours and emphasizes the need for heightened awareness and comprehensive diagnostic approaches in similar cases.

Rescuing the knee: arthroscopic extraction of a long-neglected glass fragment: a case report

The knee joint frequently encounters penetrating injuries and has ample space to harbor foreign objects. While foreign bodies in the knee joint are unusual, especially those not originating from surgical procedures, extensive literature exists on the extraction of loose bodies and metallic objects from this articulation. However, documentation on uncommon intruders remains scarce. This case study delves into the experience of a 22-year-old male who suffered persistent pain due to a neglected foreign object in his knee joint. After its removal, a complete resolution of symptoms was achieved. It highlights the importance of prompt arthroscopic intervention for optimal outcomes. Here we present a unique case involving a 22-year-old male who underwent arthroscopic retrieval of a neglected glass fragment from his knee joint. The patient sought medical attention five months after the initial injury due to sudden and severe pain onset. Arthroscopy stands as the favored minimally invasive method for extracting foreign bodies from the knee joint, especially from its deeper and less accessible regions. Utilizing pre-operative radiological imaging enhances understanding of the anatomical placement, facilitates surgical orientation, and reduces operation time. A keen sense of suspicion and thorough history-taking are crucial in achieving an accurate diagnosis.

Aggressiveness classification of clear cell renal cell carcinoma using registration-independent radiology-pathology correlation learning.

Renal cell carcinoma (RCC) is a common cancer that varies in clinical behavior. Clear cell RCC (ccRCC) is the most common RCC subtype, with both aggressive and indolent manifestations. Indolent ccRCC is often low-grade without necrosis and can be monitored without treatment. Aggressive ccRCC is often high-grade and can cause metastasis and death if not promptly detected and treated. While most RCCs are detected on computed tomography (CT) scans, aggressiveness classification is based on pathology images acquired from invasive biopsy orsurgery. CT imaging-based aggressiveness classification would be an important clinical advance, as it would facilitate non-invasive risk stratification and treatment planning. Here, we present a novel machine learning method, Correlated Feature Aggregation By Region (CorrFABR), for CT-based aggressiveness classification ofccRCC. CorrFABR is a multimodal fusion algorithm that learns from radiology and pathology images, and clinical variables in a clinically-relevant manner. CorrFABR leverages registration-independent radiology (CT) and pathology image correlations using features from vision transformer-based foundation models to facilitate aggressiveness assessment on CT images. CorrFABR consists of three main steps: (a) Feature aggregation where region-level features are extracted from radiology and pathology images at widely varying image resolutions, (b) Fusion where radiology features correlated with pathology features (pathology-informed CT biomarkers) are learned, and (c) Classification where the learned pathology-informed CT biomarkers, together with clinical variables of tumor diameter, gender, and age, are used to distinguish aggressive from indolent ccRCC using multi-layer perceptron-based classifiers. Pathology images are only required in the first two steps of CorrFABR, and are not required in the prediction module. Therefore, CorrFABR integrates information from CT images, pathology images, and clinical variables during training, but for inference, it relies solely on CT images and clinical variables, ensuring its clinical applicability. CorrFABR was trained with heterogenous, publicly-available data from 298 ccRCC tumors (136 indolent tumors, 162 aggressive tumors) in a five-fold cross-validation setup and evaluated on an independent test set of 74 tumors with a balanced distribution of aggressive and indolent tumors. Ablation studies were performed to test the utility of each component ofCorrFABR. CorrFABR outperformed the other classification methods, achieving an ROC-AUC (area under the curve) of 0.855±0.0005 (95% confidence interval: 0.775, 0.947), F1-score of 0.793±0.029, sensitivity of 0.741±0.058, and specificity of 0.876±0.032 in classifying ccRCC as aggressive or indolent subtypes. It was found that pathology-informed CT biomarkers learned through registration-independent correlation learning improves classification performance over using CT features alone, irrespective of the kind of features or the classification model used. Tumor diameter, gender, and age provide complementary clinical information, and integrating pathology-informed CT biomarkers with these clinical variables further improvesperformance. CorrFABR provides a novel method for CT-based aggressiveness classification of ccRCC by enabling the identification of pathology-informed CT biomarkers, and integrating them with clinical variables. CorrFABR enables learning of these pathology-informed CT biomarkers through a novel registration-independent correlation learning module that considers unaligned radiology and pathology images at widely varying imageresolutions.

Unmet Needs in Spondyloarthritis: Imaging in Axial Spondyloarthritis.

Imaging biomarkers in axial spondyloarthritis (axSpA) are currently the most specific biomarkers for the diagnosis of this condition. Despite advances in imaging, from plain radiographs-which detect only damage-to magnetic resonance imaging (MRI)-which identifies disease activity and structural change-there are still many challenges that remain. Imaging in sacroiliitis is characterized by active and structural changes. Current classification criteria stress the importance of bone marrow edema (BME); however, BME can occur in various diseases, mechanical conditions, and healthy individuals. Thus, the identification of structural lesions such as erosion, subchondral fat, backfill, and ankylosis is important to distinguish from mimics on differential diagnosis. Various imaging modalities are available to examine structural lesions, but computed tomography (CT) is considered the current reference standard. Nonetheless, recent advances in MRI allow for direct bone imaging and the reconstruction of CT-like images that can provide similar information. Therefore, the ability of MRI to detect and measure structural lesions is strengthened. Here, we present an overview of the spectrum of current and cutting-edge techniques for SpA imaging in clinical practice; namely, we discuss the advantages, disadvantages, and usefulness of imaging in SpA through radiography, low-dose and dual-energy CT, and MRI. Cutting-edge MRI sequences including volumetric interpolated breath-hold examination, ultrashort echo time, zero echo time, and deep learning-based synthetic CT that creates CT-like images without ionizing radiation, are discussed. Imaging techniques allow for quantification of inflammatory and structural lesions, which is important in the assessment of treatment response and disease progression. Radiographic damage is poorly sensitive to change. Artificial intelligence has already revolutionized radiology practice, including protocolization, image quality, and image interpretation.

Integrated ensemble CNN and explainable AI for COVID-19 diagnosis from CT scan and X-ray images

In light of the ongoing battle against COVID-19, while the pandemic may eventually subside, sporadic cases may still emerge, underscoring the need for accurate detection from radiological images. However, the limited explainability of current deep learning models restricts clinician acceptance. To address this issue, our research integrates multiple CNN models with explainable AI techniques, ensuring model interpretability before ensemble construction. Our approach enhances both accuracy and interpretability by evaluating advanced CNN models on the largest publicly available X-ray dataset, COVIDx CXR-3, which includes 29,986 images, and the CT scan dataset for SARS-CoV-2 from Kaggle, which includes a total of 2,482 images. We also employed additional public datasets for cross-dataset evaluation, ensuring a thorough assessment of model performance across various imaging conditions. By leveraging methods including LIME, SHAP, Grad-CAM, and Grad-CAM++, we provide transparent insights into model decisions. Our ensemble model, which includes DenseNet169, ResNet50, and VGG16, demonstrates strong performance. For the X-ray image dataset, sensitivity, specificity, accuracy, F1-score, and AUC are recorded at 99.00%, 99.00%, 99.00%, 0.99, and 0.99, respectively. For the CT image dataset, these metrics are 96.18%, 96.18%, 96.18%, 0.9618, and 0.96, respectively. Our methodology bridges the gap between precision and interpretability in clinical settings by combining model diversity with explainability, promising enhanced disease diagnosis and greater clinician acceptance.

Radiological, clinical, and molecular analyses reveal distinct subtypes of butterfly glioblastomas affecting the prognosis

Abstract Background Glioblastoma (GB) is known for its highly invasive nature. Images of butterfly GB (bGB) often illustrate this characteristic, but the molecular background and origins of bGB remain unknown. Methods We analyzed a cohort of 34 bGB patients from our dataset (K-cohort) and 46 bGB patients from publicly available datasets, including TCGA-GBM, CPTAC-GBM, IvyGAP, and UPENN-GBM. Results In the K-cohort, the median age was 66 years, and molecular analyses revealed TERT promoter mutations in 55.9% of cases, with no cases exhibiting H3F3A, HIST1H3B, or BRAF mutations. Sequential radiological imaging from the K-cohort provided unique insights, showing one case originating in the corpus callosum (CC) and three cases originating in the cerebral hemisphere before developing into bGB. Multi-regional sampling supported a mutational trajectory from the hemisphere to the CC. These observations indicate the presence of two distinct radiological origins for bGB. Consequently, we classified cases into CC-type and Hemispheric-type based on the tumor volume ratio within the CC. This subgrouping was clinically meaningful; the CC-type is an independent poor prognostic factor for overall survival, with a hazard ratio of 1.8 (95% confidence interval 1.1-3.0, P = 0.033), and is molecularly distinct by a higher frequency of methylated MGMTp (P = 0.0039) compared to the Hemispheric-type. Conclusions Our results highlight that the radiological features of bGB are not homogenous and can indicate two potential subtypes based on their origins. Further studies are mandatory, but CC-type and Hemispheric-type exhibit distinct clinical backgrounds, outcomes, and molecular features.

Foundational Segmentation Models and Clinical Data Mining Enable Accurate Computer Vision for Lung Cancer.

This study aims to assess the effectiveness of integrating Segment Anything Model (SAM) and its variant MedSAM into the automated mining, object detection, and segmentation (MODS) methodology for developing robust lung cancer detection and segmentation models without post hoc labeling of training images. In a retrospective analysis, 10,000 chest computed tomography scans from patients with lung cancer were mined. Line measurement annotations were converted to bounding boxes, excluding boxes < 1cm or > 7cm. The You Only Look Once object detection architecture was used for teacher-student learning to label unannotated lesions on the training images. Subsequently, a final tumor detection model was trained and employed with SAM and MedSAM for tumor segmentation. Model performance was assessed on a manually annotated test dataset, with additional evaluations conducted on an external lung cancer dataset before and after detection model fine-tuning. Bootstrap resampling was used to calculate 95% confidence intervals. Data mining yielded 10,789 line annotations, resulting in 5403 training boxes. The baseline detection model achieved an internal F1 score of 0.847, improving to 0.860 after self-labeling. Tumor segmentation using the final detection model attained internal Dice similarity coefficients (DSCs) of 0.842 (SAM) and 0.822 (MedSAM). After fine-tuning, external validation showed an F1 of 0.832 and DSCs of 0.802 (SAM) and 0.804 (MedSAM). Integrating foundational segmentation models into the MODS framework results in high-performing lung cancer detection and segmentation models using only mined clinical data. Both SAM and MedSAM hold promise as foundational segmentation models for radiology images.

A rare and misleading condition: isolated skeletal involvement of Erdheim-Chester Disease.

We are describing an unusual presentation of Erdheim-Chester Disease (ECD), a non -Langerhans histiocytosis provoked by abnormal aggregation of foamy histiocytes in several organs and systems. Our patient, already with osteoporosis, presented severe bilateral ankle pain with functional impairment, clinically mimicking "Complex Regional Pain Syndrome" (CRPS); the radiological images and subsequently the bone biopsy were fundamental for achieving the proper diagnosis. The diagnosis was challenging given the isolated bone localization which is uncommonly seen in ECD.

Reproducibility and interpretability in radiomics: a critical assessment.

Radiomics aims to improve clinical decision making through the use of radiological imaging. However, the field is challenged by reproducibility issues due to variability in imaging and subsequent statistical analysis, which particularly affects the interpretability of the model. In fact, radiomics extracts many highly correlated features that, combined with the small sample sizes often found in radiomics studies, result in high-dimensional datasets. These datasets, which are characterized by containing more features than samples, have different statistical properties than other datasets, thereby complicating their training by machine learning and deep learning methods. This review critically examines the challenges of both reproducibility issues and interpretability, beginning with an overview of the radiomics pipeline, followed by a discussion of the imaging and statistical reproducibility issues. It further highlights how limited model interpretability hinders clinical translation. The discussion concludes that these challenges could be mitigated by following best practices and by creating large, representative, and publicly available datasets.

Revolution or risk?-Assessing the potential and challenges of GPT-4V in radiologic image interpretation.

ChatGPT-4 Vision (GPT-4V) is a state-of-the-art multimodal large language model (LLM) that may be queried using images. We aimed to evaluate the tool's diagnostic performance when autonomously assessing clinical imaging studies. A total of 206 imaging studies (i.e., radiography (n = 60), CT (n = 60), MRI (n = 60), and angiography (n = 26)) with unequivocal findings and established reference diagnoses from the radiologic practice of a large university hospital were accessed. Readings were performed uncontextualized, with only the image provided, and contextualized, with additional clinical and demographic information. Responses were assessed along multiple diagnostic dimensions and analyzed using appropriate statistical tests. With its pronounced propensity to favor context over image information, the tool's diagnostic accuracy improved from 8.3% (uncontextualized) to 29.1% (contextualized, first diagnosis correct) and 63.6% (contextualized, correct diagnosis among differential diagnoses) (p ≤ 0.001, Cochran's Q test). Diagnostic accuracy declined by up to 30% when 20 images were re-read after 30 and 90 days and seemed unrelated to the tool's self-reported confidence (Spearman's ρ = 0.117 (p = 0.776)). While the described imaging findings matched the suggested diagnoses in 92.7%, indicating valid diagnostic reasoning, the tool fabricated 258 imaging findings in 412 responses and misidentified imaging modalities or anatomic regions in 65 images. GPT-4V, in its current form, cannot reliably interpret radiologic images. Its tendency to disregard the image, fabricate findings, and misidentify details, especially without clinical context, may misguide healthcare providers and put patients at risk. Question Can Generative Pre-trained Transformer 4 Vision (GPT-4V) interpret radiologic images-with and without clinical context? Findings GPT-4V performed poorly, demonstrating diagnostic accuracy rates of 8% (uncontextualized), 29% (contextualized, most likely diagnosis correct), and 64% (contextualized, correct diagnosis among differential diagnoses). Clinical relevance The utility of commercial multimodal large language models, such as GPT-4V, in radiologic practice is limited. Without clinical context, diagnostic errors and fabricated findings may compromise patient safety and misguide clinical decision-making. These models must be further refined to be beneficial.