RESUMEN
Introduction: This study explores the use of the latest You Only Look Once (YOLO V7) object detection method to enhance kidney detection in medical imaging by training and testing a modified YOLO V7 on medical image formats. Methods: Study includes 878 patients with various subtypes of renal cell carcinoma (RCC) and 206 patients with normal kidneys. A total of 5657 MRI scans for 1084 patients were retrieved. 326 patients with 1034 tumors recruited from a retrospective maintained database, and bounding boxes were drawn around their tumors. A primary model was trained on 80% of annotated cases, with 20% saved for testing (primary test set). The best primary model was then used to identify tumors in the remaining 861 patients and bounding box coordinates were generated on their scans using the model. Ten benchmark training sets were created with generated coordinates on not-segmented patients. The final model used to predict the kidney in the primary test set. We reported the positive predictive value (PPV), sensitivity, and mean average precision (mAP). Results: The primary training set showed an average PPV of 0.94 ± 0.01, sensitivity of 0.87 ± 0.04, and mAP of 0.91 ± 0.02. The best primary model yielded a PPV of 0.97, sensitivity of 0.92, and mAP of 0.95. The final model demonstrated an average PPV of 0.95 ± 0.03, sensitivity of 0.98 ± 0.004, and mAP of 0.95 ± 0.01. Conclusion: Using a semi-supervised approach with a medical image library, we developed a high-performing model for kidney detection. Further external validation is required to assess the model's generalizability.
RESUMEN
BACKGROUND: Pathology grading is an essential step for the treatment and evaluation of the prognosis in patients with clear cell renal cell carcinoma (ccRCC). PURPOSE: To investigate the utility of texture analysis in evaluating Fuhrman grades of renal tumors in patients with Von Hippel-Lindau (VHL)-associated ccRCC, aiming to improve non-invasive diagnosis and personalized treatment. STUDY TYPE: Retrospective analysis of a prospectively maintained cohort. POPULATION: One hundred and thirty-six patients, 84 (61%) males and 52 (39%) females with pathology-proven ccRCC with a mean age of 52.8 ± 12.7 from 2010 to 2023. FIELD STRENGTH AND SEQUENCES: 1.5 and 3 T MRIs. Segmentations were performed on the T1-weighted 3-minute delayed sequence and then registered on pre-contrast, T1-weighted arterial and venous sequences. ASSESSMENT: A total of 404 lesions, 345 low-grade tumors, and 59 high-grade tumors were segmented using ITK-SNAP on a T1-weighted 3-minute delayed sequence of MRI. Radiomics features were extracted from pre-contrast, T1-weighted arterial, venous, and delayed post-contrast sequences. Preprocessing techniques were employed to address class imbalances. Features were then rescaled to normalize the numeric values. We developed a stacked model combining random forest and XGBoost to assess tumor grades using radiomics signatures. STATISTICAL TESTS: The model's performance was evaluated using positive predictive value (PPV), sensitivity, F1 score, area under the curve of receiver operating characteristic curve, and Matthews correlation coefficient. Using Monte Carlo technique, the average performance of 100 benchmarks of 85% train and 15% test was reported. RESULTS: The best model displayed an accuracy of 0.79. For low-grade tumor detection, a sensitivity of 0.79, a PPV of 0.95, and an F1 score of 0.86 were obtained. For high-grade tumor detection, a sensitivity of 0.78, PPV of 0.39, and F1 score of 0.52 were reported. DATA CONCLUSION: Radiomics analysis shows promise in classifying pathology grades non-invasively for patients with VHL-associated ccRCC, potentially leading to better diagnosis and personalized treatment. LEVEL OF EVIDENCE: 1 TECHNICAL EFFICACY: Stage 2.
Asunto(s)
Carcinoma de Células Renales , Neoplasias Renales , Imagen por Resonancia Magnética , Clasificación del Tumor , Humanos , Carcinoma de Células Renales/diagnóstico por imagen , Carcinoma de Células Renales/patología , Femenino , Masculino , Persona de Mediana Edad , Neoplasias Renales/diagnóstico por imagen , Neoplasias Renales/patología , Imagen por Resonancia Magnética/métodos , Estudios Retrospectivos , Adulto , Anciano , Enfermedad de von Hippel-Lindau/diagnóstico por imagen , Enfermedad de von Hippel-Lindau/complicaciones , Curva ROC , Procesamiento de Imagen Asistido por Computador/métodos , PronósticoRESUMEN
INTRODUCTION: Classification of clear cell renal cell carcinoma (ccRCC) growth rates in patients with Von Hippel-Lindau (VHL) syndrome has several ramifications for tumor monitoring and surgical planning. Using two separate machine-learning algorithms, we sought to produce models to predict ccRCC growth rate classes based on qualitative MRI-derived characteristics. MATERIAL AND METHODS: We used a prospectively maintained database of patients with VHL who underwent surgical resection for ccRCC between January 2015 and June 2022. We employed a threshold growth rate of 0.5 cm per year to categorize ccRCC tumors into two distinct groups-'slow-growing' and 'fast-growing'. Utilizing a questionnaire of qualitative imaging features, two radiologists assessed each lesion on different MRI sequences. Two machine-learning models, a stacked ensemble technique and a decision tree algorithm, were used to predict the tumor growth rate classes. Positive predictive value (PPV), sensitivity, and F1-score were used to evaluate the performance of the models. RESULTS: This study comprises 55 patients with VHL with 128 ccRCC tumors. Patients' median age was 48 years, and 28 patients were males. Each patient had an average of two tumors, with a median size of 2.1 cm and a median growth rate of 0.35 cm/year. The overall performance of the stacked and DT model had 0.77 ± 0.05 and 0.71 ± 0.06 accuracies, respectively. The best stacked model achieved a PPV of 0.92, a sensitivity of 0.91, and an F1-score of 0.90. CONCLUSION: This study provides valuable insight into the potential of machine-learning analysis for the determination of renal tumor growth rate in patients with VHL. This finding could be utilized as an assistive tool for the individualized screening and follow-up of this population.
Asunto(s)
Carcinoma de Células Renales , Carcinoma , Neoplasias Renales , Masculino , Humanos , Persona de Mediana Edad , Femenino , Carcinoma de Células Renales/diagnóstico por imagen , Carcinoma de Células Renales/patología , Riñón/diagnóstico por imagen , Riñón/patología , Neoplasias Renales/diagnóstico por imagen , Neoplasias Renales/cirugía , Imagen por Resonancia Magnética , Aprendizaje AutomáticoRESUMEN
INTRODUCTION: Accurate diagnosis and treatment of kidney tumors greatly benefit from automated solutions for detection and classification on MRI. In this study, we explore the application of a deep learning algorithm, YOLOv7, for detecting kidney tumors on contrast-enhanced MRI. MATERIAL AND METHODS: We assessed the performance of YOLOv7 tumor detection on excretory phase MRIs in a large institutional cohort of patients with RCC. Tumors were segmented on MRI using ITK-SNAP and converted to bounding boxes. The cohort was randomly divided into ten benchmarks for training and testing the YOLOv7 algorithm. The model was evaluated using both 2-dimensional and a novel in-house developed 2.5-dimensional approach. Performance measures included F1, Positive Predictive Value (PPV), Sensitivity, F1 curve, PPV-Sensitivity curve, Intersection over Union (IoU), and mean average PPV (mAP). RESULTS: A total of 326 patients with 1034 tumors with 7 different pathologies were analyzed across ten benchmarks. The average 2D evaluation results were as follows: Positive Predictive Value (PPV) of 0.69 ± 0.05, sensitivity of 0.39 ± 0.02, and F1 score of 0.43 ± 0.03. For the 2.5D evaluation, the average results included a PPV of 0.72 ± 0.06, sensitivity of 0.61 ± 0.06, and F1 score of 0.66 ± 0.04. The best model performance demonstrated a 2.5D PPV of 0.75, sensitivity of 0.69, and F1 score of 0.72. CONCLUSION: Using computer vision for tumor identification is a cutting-edge and rapidly expanding subject. In this work, we showed that YOLOv7 can be utilized in the detection of kidney cancers.
Asunto(s)
Carcinoma de Células Renales , Aprendizaje Profundo , Neoplasias Renales , Humanos , Algoritmos , Carcinoma de Células Renales/diagnóstico por imagen , Neoplasias Renales/diagnóstico por imagen , Imagen por Resonancia Magnética , Distribución AleatoriaRESUMEN
OBJECTIVE: Radiology reporting is an essential component of clinical diagnosis and decision-making. With the advent of advanced artificial intelligence (AI) models like GPT-4 (Generative Pre-trained Transformer 4), there is growing interest in evaluating their potential for optimizing or generating radiology reports. This study aimed to compare the quality and content of radiologist-generated and GPT-4 AI-generated radiology reports. METHODS: A comparative study design was employed in the study, where a total of 100 anonymized radiology reports were randomly selected and analyzed. Each report was processed by GPT-4, resulting in the generation of a corresponding AI-generated report. Quantitative and qualitative analysis techniques were utilized to assess similarities and differences between the two sets of reports. RESULTS: The AI-generated reports showed comparable quality to radiologist-generated reports in most categories. Significant differences were observed in clarity (p = 0.027), ease of understanding (p = 0.023), and structure (p = 0.050), favoring the AI-generated reports. AI-generated reports were more concise, with 34.53 fewer words and 174.22 fewer characters on average, but had greater variability in sentence length. Content similarity was high, with an average Cosine Similarity of 0.85, Sequence Matcher Similarity of 0.52, BLEU Score of 0.5008, and BERTScore F1 of 0.8775. CONCLUSION: The results of this proof-of-concept study suggest that GPT-4 can be a reliable tool for generating standardized radiology reports, offering potential benefits such as improved efficiency, better communication, and simplified data extraction and analysis. However, limitations and ethical implications must be addressed to ensure the safe and effective implementation of this technology in clinical practice. CLINICAL RELEVANCE STATEMENT: The findings of this study suggest that GPT-4 (Generative Pre-trained Transformer 4), an advanced AI model, has the potential to significantly contribute to the standardization and optimization of radiology reporting, offering improved efficiency and communication in clinical practice. KEY POINTS: ⢠Large language model-generated radiology reports exhibited high content similarity and moderate structural resemblance to radiologist-generated reports. ⢠Performance metrics highlighted the strong matching of word selection and order, as well as high semantic similarity between AI and radiologist-generated reports. ⢠Large language model demonstrated potential for generating standardized radiology reports, improving efficiency and communication in clinical settings.
RESUMEN
BACKGROUND: The CT findings of cirrhosis caused by primary sclerosing cholangitis (PSC) differ from cryptogenic cirrhosis. PSC could become complicated with biliary cirrhosis and cholangiocarcinoma. This study aimed at augmenting the information on the role of the three-phasic-abdominopelvic CT scan in PSC. MATERIAL AND METHODS: A total of 185 CT scans were retrospectively reviewed, including 100 patients with cryptogenic cirrhosis and 85 patients with PSC-cirrhosis. Different morphologic criteria were compared, including segmental atrophy/hypertrophy, hepatic contour, portal-hypertension, perihilar lymphadenopathy, biliary tree dilatation, gallbladder appearance. Inflammatory-bowel-disease (IBD) and cholangiocarcinoma frequency, presence of perihilar lymph nodes (LNs), and their size during end-stage PSC cirrhosis are investigated. RESULTS: Six findings occur more frequently with PSC than those diagnosed with cryptogenic cirrhosis. Modified caudate/right lobe (m-CRL) ratio >0.73, moderate and severe lobulated liver contour, lateral left lobe atrophy, over distended gallbladder (GB), biliary tree dilatation and wall thickening, and LN sizes were higher in PSC patients as compared to cryptogenic cirrhosis (P < 0.005). Ascites and portosystemic collateral formations were significant in cryptogenic cirrhosis compared to PSC patients (P < 0.005). Cholangiocarcinoma frequency in PSC patients was 14.7%, and the frequency of inflammatory bowel disease (IBD) was 57.6%. Further, 22.4% of the patients were diagnosed with IBD and PSC simultaneously. The LN number and size in PSC patients were not different between those with or without cholangiocarcinoma. CONCLUSION: Using three-phasic CT scans and PSC characteristics could be considered as an additional suggestion besides pathology measures. Diagnosis of PSC based on histological findings could be a last resort due to its invasive essence and specific characteristics of PSC in imaging.
RESUMEN
PURPOSE: The purpose of this study was to establish normative values for the thoracic aorta diameter in pediatric patients from birth to 18 years of age using computed tomography (CT) measurements and to create nomograms related to body surface area (BSA). METHODS: A total of 623 pediatric patients without cardiovascular disease (42.1% females; from 3 d to 18 y old) with high-quality, non-electrocardiogram-gated, contrast-enhanced CT imaging of the chest were retrospectively evaluated. Systematic measurements of the aortic diameter at predetermined levels were recorded, and demographic data including age, sex, ethnicity, and BSA were collected. Reference graphs plotting BSA over aortic diameter included the mean and Z -3 to Z +3, where Z represents SDs from the mean. RESULTS: The study population was divided into 2 groups (below 2 and greater than or equal to 2 y old). There were no significant differences in average aortic measurements between males and females. Both age groups exhibited significant positive correlations among all size-related metrics (all P <0.001) with BSA having the highest correlation. For both groups, the average orthogonal thoracic aortic diameters at each level of the thoracic aorta were used to create nomograms. CONCLUSION: This study establishes clinically applicable, BSA-specific reference values of the normal thoracic aorta for the pediatric population from CT imaging.
Asunto(s)
Aorta Torácica , Tomografía Computarizada por Rayos X , Factores de Edad , Aorta Torácica/diagnóstico por imagen , Superficie Corporal , Niño , Femenino , Humanos , Masculino , Valores de Referencia , Estudios Retrospectivos , Factores Sexuales , Tomografía Computarizada por Rayos X/métodosRESUMEN
Background There are considerable differences in the prevalence of coronary artery disease (CAD) and its cardiovascular risk factors between men and women. Due to the significance of gender as a factor that potentially affects cardiovascular disorders and patient outcomes, the present study aimed to assess the baseline characteristics and outcomes of CAD patients in terms of gender distribution. Methods All consecutive patients diagnosed with ST-elevation myocardial infarction (MI) who had undergone primary percutaneous coronary intervention (PCI) in the previous two years in a comprehensive cardiology center were included. Data were retrospectively collected from the hospital record files. Color Doppler echocardiography, valvular involvement, and the type of coronary vessel involvement were also evaluated. Results In total, 557 consecutive patients (437 men and 120 women) were included with a mean age of 59.37 ± 26.23 years and 64.07 ± 11.60 years for men and women, respectively (p = 0.004). The prevalence of mitral regurgitation (MR) and tricuspid regurgitation (TR) was significantly higher among women than men. Conclusion Female patients who suffered from CAD and underwent PCI were older than men. Also, ischemic mitral regurgitation (MR) and tricuspid regurgitation (TR) were more prevalent among women, while smoking was more prevalent among men.
RESUMEN
PURPOSE: The current literature reports a wide range of diagnostic accuracy of non-contrast magnetic resonance coronary angiography (NC-MRCA) for the assessment of coronary artery disease (CAD). We aimed to compare the clinical effectiveness of NC-MRCA with that of invasive coronary angiography (ICA) in patients with suspected CAD using a systematic review and meta-analysis. METHODS: Two investigators independently extracted 36 published manuscripts between 2010 and 2019. Databases including Medline, Web of Knowledge, Google Scholar, Scopus, and Cochrane were searched using pre-established keywords. Analysis of the data followed the PRISMA statement for reporting systematic reviews and meta-analyses and primary analysis followed the Mantel-Hansel methodology. Correctness of classification for detecting coronary artery stenosis ≥50% (CAS) was measured using ICA as the gold standard. RESULTS: A total of five studies met inclusion criteria, with a total of 417 patients and 2883 coronary segments. The pooled per patient sensitivity and specificity of NC-MRCA for CAS in suspected patients was 90.3% (95% CI 85.6-95.1%) and 77.9% (95% CI 69.5-86.3%). Pooled per vessel assessment of NC- MRCA revealed a sensitivity of 83.7% (95%CI 79.7-87.8%) and specificity of 90.0% (95%CI 86.7-93.4%). Per-segment assessment of NC-MRCA showed a pooled sensitivity of 81.6% (95% CI 76.8-86.4) and specificity of 97.0% (95% CI 95.5-98.5). Mild to moderate heterogeneity was noted in most diagnostic parameters with larger heterogeneity noted in the per-segment analyses. There was less heterogeneity in sensitivity and NPV than specificity and PPV. CONCLUSION: According to this meta-analysis, non-contrast coronary MRA resulted in adequate screening in patients with suspected CAD with high sensitivity and specificity. This result was true for per-patient, per-vessel, and per-segment assessment.
RESUMEN
BACKGROUND: Marijuana is the most popular drug of abuse in the United States. The association between its use and coronary artery disease has not yet been fully elucidated. This study aims to determine the frequency of coronary artery disease among young to middle aged adults presenting with chest pain who currently use marijuana as compared to nonusers. METHODS: In this retrospective study, 1,420 patients with chest pain or angina equivalent were studied. Only men between 18 and 40 years and women between 18 and 50 years of age without history of cardiac disease were included. All patients were queried about current or prior cannabis use and underwent coronary CT angiography. Each coronary artery on coronary CT angiography was assessed based on the CAD-RADS reporting system. RESULTS: A total of 146 (10.3%) out of 1,420 patients with chest pain were identified as marijuana users. Only 6.8% of the 146 marijuana users had evidence of coronary artery disease on coronary CT angiography. In comparison, the rate was 15.0% among the 1,274 marijuana nonusers (p = 0.008). After accounting for other cardiac risk factors in a multivariate analysis, the negative association between marijuana use and coronary artery disease on coronary CT angiography diminished (p = 0.12, 95% CI 0.299-1.15). A majority of marijuana users were younger than nonusers and had a lower frequency of hypertension and diabetes than nonusers. There was no statistical difference in lipid panel values between the two groups. Only 2 out of 10 marijuana users with coronary artery disease on coronary CT angiography had hemodynamically significant stenosis. CONCLUSION: Among younger patients being evaluated for chest pain, self-reported cannabis use conferred no additional risk of coronary artery disease as detected on coronary CT angiography.
Asunto(s)
Enfermedad de la Arteria Coronaria/epidemiología , Uso de la Marihuana/epidemiología , Adulto , Estudios de Casos y Controles , Angiografía por Tomografía Computarizada , Enfermedad de la Arteria Coronaria/diagnóstico por imagen , Femenino , Humanos , Masculino , Uso de la Marihuana/efectos adversos , Persona de Mediana Edad , Estudios Retrospectivos , Estados Unidos/epidemiología , Adulto JovenRESUMEN
Objective: Breast cancer is the most common cancer and leading cause of mortality in women worldwide. Diagnosis of axillary lymph node involvement, frequently by axillary ultrasonography (AUS), is an important step in patients with primary breast cancer, while the gold standard pathological examination is only applicable during surgery. Thus, we aimed to evaluate the predictive value of AUS in detection of lymph node involvement. Methods: A total of 56 patients with primary breast cancer, who were candidate of mastectomy were included in this study. A single radiologist performed pre-operative axillary sonography for all the patients and reported the results in a checklist. The results were then compared with intra-operative pathological results of lymph node dissection to evaluate the sensitivity, specificity, positive and negative predictive values of AUS in detection of the disease. Result: The results showed that pre-operative AUS had a sensitivity of 63.3%, specificity of 84.6%, positive predictive value of 82.6%, and negative predictive value of 66.6%. Correlation between axillary lymph node characteristics revealed that the absence of fatty hilum, cortical thickness, and loss of ovality of the lymph nodes were the most specific factors in detection of lymph nodes' involvement. Conclusion: AUS has a low sensitivity and cannot determine the presence of micro-metastases and suspicious lymph nodes accurately in early stages of the disease, and it seems that defining new parameters may help for increasing the sensitivity and specificity of the axillary lymph node sonography. Apparently, pathological examination remains the gold standard diagnostic method.