RESUMEN
Cancer patients have a higher risk of cardiovascular disease (CVD) mortality than the general population. Low dose computed tomography (LDCT) for lung cancer screening offers an opportunity for simultaneous CVD risk estimation in at-risk patients. Our deep learning CVD risk prediction model, trained with 30,286 LDCTs from the National Lung Cancer Screening Trial, achieves an area under the curve (AUC) of 0.871 on a separate test set of 2,085 subjects and identifies patients with high CVD mortality risks (AUC of 0.768). We validate our model against ECG-gated cardiac CT based markers, including coronary artery calcification (CAC) score, CAD-RADS score, and MESA 10-year risk score from an independent dataset of 335 subjects. Our work shows that, in high-risk patients, deep learning can convert LDCT for lung cancer screening into a dual-screening quantitative tool for CVD risk estimation.
Asunto(s)
Enfermedades Cardiovasculares/epidemiología , Aprendizaje Profundo , Procesamiento de Imagen Asistido por Computador/métodos , Neoplasias Pulmonares/diagnóstico , Tamizaje Masivo/estadística & datos numéricos , Adulto , Anciano , Anciano de 80 o más Años , Enfermedades Cardiovasculares/diagnóstico , Enfermedades Cardiovasculares/etiología , Ensayos Clínicos como Asunto , Vasos Coronarios/diagnóstico por imagen , Conjuntos de Datos como Asunto , Electrocardiografía , Femenino , Estudios de Seguimiento , Humanos , Pulmón/diagnóstico por imagen , Neoplasias Pulmonares/complicaciones , Masculino , Tamizaje Masivo/métodos , Persona de Mediana Edad , Curva ROC , Estudios Retrospectivos , Medición de Riesgo/métodos , Medición de Riesgo/estadística & datos numéricos , Factores de Riesgo , Tomografía Computarizada por Rayos X/estadística & datos numéricosRESUMEN
PURPOSE: To assess and compare detectability of pneumothorax on unprocessed baseline, single-energy, bone-subtracted, and enhanced frontal chest radiographs (chest X-ray, CXR). METHOD AND MATERIALS: Our retrospective institutional review board-approved study included 202 patients (mean age 53 ± 24 years; 132 men, 70 women) who underwent frontal CXR and had trace, moderate, large, or tension pneumothorax. All patients (except those with tension pneumothorax) had concurrent chest computed tomography (CT). Two radiologists reviewed the CXR and chest CT for pneumothorax on baseline CXR (ground truth). All baseline CXR were processed to generate bone-subtracted and enhanced images (ClearRead X-ray). Four radiologists (R1-R4) assessed the baseline, bone-subtracted, and enhanced images and recorded the presence of pneumothorax (side, size, and confidence for detection) for each image type. Area under the curve (AUC) was calculated with receiver operating characteristic analyses to determine the accuracy of pneumothorax detection. RESULTS: Bone-subtracted images (AUC: 0.89-0.97) had the lowest accuracy for detection of pneumothorax compared to the baseline (AUC: 0.94-0.97) and enhanced (AUC: 0.96-0.99) radiographs (P < .01). Most false-positive and false-negative pneumothoraces were detected on the bone-subtracted images and the least numbers on the enhanced radiographs. Highest detection rates and confidence were noted for the enhanced images (empiric AUC for R1-R4 0.96-0.99). CONCLUSION: Enhanced CXRs are superior to bone-subtracted and unprocessed radiographs for detection of pneumothorax. CLINICAL RELEVANCE/APPLICATION: Enhanced CXRs improve detection of pneumothorax over unprocessed images; bone-subtracted images must be cautiously reviewed to avoid false negatives.
Asunto(s)
Neumotórax/diagnóstico por imagen , Radiografía Torácica/métodos , Adulto , Anciano , Área Bajo la Curva , Huesos/diagnóstico por imagen , Reacciones Falso Negativas , Reacciones Falso Positivas , Femenino , Humanos , Procesamiento de Imagen Asistido por Computador , Masculino , Persona de Mediana Edad , Curva ROC , Estudios Retrospectivos , Tomografía Computarizada por Rayos XRESUMEN
OBJECTIVE: Compared with other countries in the Middle East, Qatar has one of the highest breast cancer incidence and mortality rates. Poor quality mammography images may be associated with advanced stage breast cancer, however there is limited information about the quality of breast imaging in Qatar. Our purpose was to evaluate the clinical image quality of mammography examinations performed at a tertiary care center in Doha, Qatar using a standardized assessment tool. MATERIALS AND METHODS: Bilateral mammograms from consecutive patients from a tertiary care cancer center in Doha, Qatar were obtained. Proportions of examinations deemed adequate for interpretation were estimated. Standardized clinical image quality assessment form was utilized to evaluate image quality components. For each image, image quality components were given grades on a 1-5 scale (5-excellent, 4-good, 3-average, 2-fair, 1-poor). Mean scores with 95% confidence intervals were estimated for each component. RESULTS: Consecutive sample of 132 patients was obtained representing 528 mammographic images. Overall, 99.2% of patients underwent examinations rated as acceptable for interpretation. Mean scores for each image quality component ranged from 4.045 to 5.000 (lowest score for inframammary fold). Image quality component scores were 93.0% excellent, 5.2% good, 1.1% average, 0.6% fair, and 0.1% poor. CONCLUSION: Overall image quality at a tertiary care center in Doha, Qatar was acceptable for interpretation with minimal areas identified for improvement.
RESUMEN
PURPOSE: Prompt diagnosis and quantitation of pneumothorax impact decisions pertaining to patient management. The purpose of our study was to develop and evaluate the accuracy of a deep learning (DL)-based image classification program for detection of pneumothorax on chest CT. METHOD: In an IRB approved study, an eight-layer convolutional neural network (CNN) using constant-size (36*36 pixels) 2D image patches was trained on a set of 80 chest CTs, with (n = 50) and without (n = 30) pneumothorax. Image patches were classified based on their probability of representing pneumothorax with subsequent generation of 3D heat-maps. The heat maps were further defined to include 1) pneumothorax area size, 2) relative location of the region to the lung boundary, and 3) a shape descriptor based on regional anisotropy. A support vector machine (SVM) was trained for classification. RESULT: We assessed performance of our program in a separate test dataset of 200 chest CT examinations, with (160/200, 75%) and without (40/200, 25%) pneumothorax. Data were analyzed to determine the accuracy, sensitivity, specificity. The subject-wise sensitivity was 100% (all 160/160 pneumothoraces detected) and specificity was 82.5% (33 true negative/40). False positive classifications were primarily related to emphysema and/or artifacts in the test images. CONCLUSION: This deep learning-based program demonstrated high accuracy for automatic detection of pneumothorax on chest CTs. By implementing it on a high-performance computing platform and integrating the domain knowledge of radiologists into the analytics framework, our method can be used to rapidly pre-screen large numbers of cases for presence of pneumothorax, a critical finding.
Asunto(s)
Aprendizaje Profundo , Neumotórax/diagnóstico por imagen , Radiografía Torácica/métodos , Tomografía Computarizada por Rayos X/métodos , Adolescente , Adulto , Anciano , Anciano de 80 o más Años , Niño , Preescolar , Femenino , Humanos , Masculino , Persona de Mediana Edad , Redes Neurales de la Computación , Reproducibilidad de los Resultados , Sensibilidad y Especificidad , Máquina de Vectores de Soporte , Tiempo , Adulto JovenRESUMEN
OBJECTIVE: We assessed the effect of the forward projected model-based reconstruction technique (FIRST) on lesion detection of routine abdomen CT at <1 mSv. MATERIALS AND METHODS: Thirty-seven adult patients gave written informed consent for acquisition of low-dose CT (LDCT) immediately after their clinically-indicated, standard of care dose (SDCT), routine abdomen CT on a 640-slice MDCT (Aquillion One, Canon Medical System). The LDCT series were reconstructed with FIRST (at STD (Standard) and STR (Strong) levels), and SDCT series with filtered back projection (FBP). Two radiologists assessed lesions in LD-FBP and FIRST images followed by SDCT images. Then, SDCT and LDCT were compared for presence of artifacts in a randomized and blinded fashion. Patient demographics, size and radiation dose descriptors (CTDIvol, DLP) were recorded. Descriptive statistics and inter-observer variability were calculated for data analysis. RESULTS: Mean CTDIvol for SDCT and LDCT were 13 ± 4.7 mGy and 2.2 ± 0.8 mGy, respectively. There were 46 true positive lesions detected on SDCT. Radiologists detected 38/46 lesions on LD-FIRST-STD compared to 26/46 lesions on LD-FIRST-STR. The eight lesions (liver and kidney cysts, pancreatic lesions, sub-cm peritoneal lymph node) missed on LD-FIRST-STD were seen in patients with BMI > 25.8 kg/m2. Diagnostic confidence for lesion assessment was optimal in LD-FIRST-STD setting in most patients regardless of their size. The inter-observer agreement (kappa-value) for overall image quality were 0.98 and 0.84 for LD-FIRST-STD and STR levels, respectively. CONCLUSION: FIRST enabled optimal lesion detection in routine abdomen CT at less than 1 mSv radiation dose in patients with body mass less than ≤25.8 kg/m2.
RESUMEN
Commercial iterative reconstruction techniques help to reduce CT radiation dose but altered image appearance and artifacts limit their adoptability and potential use. Deep learning has been investigated for low-dose CT (LDCT). Here we design a modularized neural network for LDCT and compared it with commercial iterative reconstruction methods from three leading CT vendors. While popular networks are trained for an end-to-end mapping, our network performs an end-to-process mapping so that intermediate denoised images are obtained with associated noise reduction directions towards a final denoised image. The learned workflow allows radiologists-in-the-loop to optimize the denoising depth in a task-specific fashion. Our network was trained with the Mayo LDCT Dataset, and tested on separate chest and abdominal CT exams from Massachusetts General Hospital. The best deep learning reconstructions were systematically compared to the best iterative reconstructions in a double-blinded reader study. This study confirms that our deep learning approach performed either favorably or comparably in terms of noise suppression and structural fidelity, and is much faster than the commercial iterative reconstruction algorithms.