Deep Learning Using Multiple Degrees of Maximum-Intensity Projection for PET/CT Image Classification in Breast Cancer

Deep learning (DL) has become a remarkably powerful tool for image processing recently. However, the usefulness of DL in positron emission tomography (PET)/computed tomography (CT) for breast cancer (BC) has been insufficiently studied. This study investigated whether a DL model using images with multiple degrees of PET maximum-intensity projection (MIP) images contributes to increase diagnostic accuracy for PET/CT image classification in BC. We retrospectively gathered 400 images of 200 BC and 200 non-BC patients for training data. For each image, we obtained PET MIP images with four different degrees (0°, 30°, 60°, 90°) and made two DL models using Xception. One DL model diagnosed BC with only 0-degree MIP and the other used four different degrees. After training phases, our DL models analyzed test data including 50 BC and 50 non-BC patients. Five radiologists interpreted these test data. Sensitivity, specificity, and area under the receiver operating characteristic curve (AUC) were calculated. Our 4-degree model, 0-degree model, and radiologists had a sensitivity of 96%, 82%, and 80–98% and a specificity of 80%, 88%, and 76–92%, respectively. Our 4-degree model had equal or better diagnostic performance compared with that of the radiologists (AUC = 0.936 and 0.872–0.967, p = 0.036–0.405). A DL model similar to our 4-degree model may lead to help radiologists in their diagnostic work in the future.

OCTAVA: An open-source toolbox for quantitative analysis of optical coherence tomography angiography images

Optical coherence tomography angiography (OCTA) performs non-invasive visualization and characterization of microvasculature in research and clinical applications mainly in ophthalmology and dermatology. A wide variety of instruments, imaging protocols, processing methods and metrics have been used to describe the microvasculature, such that comparing different study outcomes is currently not feasible. With the goal of contributing to standardization of OCTA data analysis, we report a user-friendly, open-source toolbox, OCTAVA (OCTA Vascular Analyzer), to automate the pre-processing, segmentation, and quantitative analysis of en face OCTA maximum intensity projection images in a standardized workflow. We present each analysis step, including optimization of filtering and choice of segmentation algorithm, and definition of metrics. We perform quantitative analysis of OCTA images from different commercial and non-commercial instruments and samples and show OCTAVA can accurately and reproducibly determine metrics for characterization of microvasculature. Wide adoption could enable studies and aggregation of data on a scale sufficient to develop reliable microvascular biomarkers for early detection, and to guide treatment, of microvascular disease.

Account for the Full Extent of Esophagus Motion in Radiation Therapy Planning: A Preliminary Study of the IRV of the Esophagus

ObjectiveAccounting for esophagus motion in radiotherapy planning is an important basis for accurate assessment of toxicity. In this study, we calculated how much the delineations of the esophagus should be expanded based on three-dimensional (3D) computed tomography (CT), four-dimensional (4D) average projection (AVG), and maximum intensity projection (MIP) scans to account for the full extent of esophagus motion during 4D imaging acquisition.Methods and MaterialsThe 3D and 4D CT scans of 20 lung cancer patients treated with conventional radiotherapy and 20 patients treated with stereotactic ablative radiation therapy (SBRT) were used. Radiation oncologists contoured the esophagus on the 3DCT, AVG, MIP and 25% exhale scans, and the combination of the esophagus in every phase of 4DCT. The union of all 4D phase delineations (U4D) represented the full extent of esophagus motion during imaging acquisition. Surface distances from U4D to 3D, AVG, and MIP volumes were calculated. Distances in the most extreme surface points (1.5 cm most superoinferior, 10% most right/left/anteroposterior) were used to derive margins accounting only for systematic (delineation) errors.ResultsEsophagus delineations on the MIP were the closest to the full extent of motion, requiring only 6.9 mm margins. Delineations on the AVG and 3D scans required margins up to 7.97 and 7.90 mm, respectively. The largest margins were for the inferior, right, and anterior aspects for the delineations on the 3D, AVG, and MIP scans, respectively.ConclusionDelineations on 3D, AVG, or MIP scans required extensions for representing the esophagus’s full extent of motion, with the MIP requiring the smallest margins. Research including daily imaging to determine the random components for the margins and dosimetric measurements to determine the relevance of creating a planning organ at risk volume (PRV) of the esophagus is required.

Diagnosis of Breast Cancer Using Radiomics Models Built Based on Dynamic Contrast Enhanced MRI Combined With Mammography

ObjectiveTo build radiomics models using features extracted from DCE-MRI and mammography for diagnosis of breast cancer.Materials and Methods266 patients receiving MRI and mammography, who had well-enhanced lesions on MRI and histologically confirmed diagnosis were analyzed. Training dataset had 146 malignant and 56 benign, and testing dataset had 48 malignant and 18 benign lesions. Fuzzy-C-means clustering algorithm was used to segment the enhanced lesion on subtraction MRI maps. Two radiologists manually outlined the corresponding lesion on mammography by consensus, with the guidance of MRI maximum intensity projection. Features were extracted using PyRadiomics from three DCE-MRI parametric maps, and from the lesion and a 2-cm bandshell margin on mammography. The support vector machine (SVM) was applied for feature selection and model building, using 5 datasets: DCE-MRI, mammography lesion-ROI, mammography margin-ROI, mammography lesion+margin, and all combined.ResultsIn the training dataset evaluated using 10-fold cross-validation, the diagnostic accuracy of the individual model was 83.2% for DCE-MRI, 75.7% for mammography lesion, 64.4% for mammography margin, and 77.2% for lesion+margin. When all features were combined, the accuracy was improved to 89.6%. By adding mammography features to MRI, the specificity was significantly improved from 69.6% (39/56) to 82.1% (46/56), p<0.01. When the developed models were applied to the independent testing dataset, the accuracy was 78.8% for DCE-MRI and 83.3% for combined MRI+Mammography.ConclusionThe radiomics model built from the combined MRI and mammography has the potential to provide a machine learning-based diagnostic tool and decrease the false positive diagnosis of contrast-enhanced benign lesions on MRI.

Short-Vessel Occlusion Might Indicate Higher Possibility of Success in Reperfusion following Mechanical Thrombectomy in Acute Middle Cerebral Artery Occlusion

<b><i>Background:</i></b> The impact of the length of the occluded vessel in acute large-vessel occlusion on successful reperfusion by mechanical thrombectomy remains unclear. This study evaluated whether diameter and length of the occluded vessel in acute middle cerebral artery (MCA) occlusion might relate to successful reperfusion following mechanical thrombectomy. <b><i>Methods:</i></b> This retrospective study included patients with acute MCA occlusion who underwent intra-aortic injection of contrast medium to obtain maximum intensity projection (MIP) images acquired by flat-panel detector computed tomography (FD-CT) equipped with an angiographic system. All patients received mechanical thrombectomy and were divided into two groups: those with successful reperfusion (Thrombolysis in Cerebral Infarction [TICI] 2b/3) and those without. We compared the diameter and length of the occluded vessel between the groups. In the sub-analysis of patients with stent retriever use, ratio of length of occluded vessel to length of the active zone was compared. <b><i>Results:</i></b> We enrolled 29 patients (median age: 73, M1 occlusion: 51%, stent retriever use: 72%). Eighteen patients achieved TICI 2b/3 with significantly larger distal end diameter (1.7 [interquartile range: 1.5–1.9] vs. 1.2 [1.2–1.5] mm, <i>p</i> = 0.007) and shorter length (7.1 [4.9–9.7] vs. 12.3 [7.2–15.8] mm, <i>p</i> = 0.043) of the occluded vessel. Sub-analysis of 21 patients showed that the cut-off value for TICI 2b/3 reperfusion was 0.32 as the ratio between the occluded vessel and stent retriever active zone (receiver operating characteristic area under the curve: 0.90). <b><i>Conclusion:</i></b> In acute MCA occlusion, larger diameter of the distal end and shorter length of the occluded vessel on FD-CT MIP images might indicate a higher possibility of achieving TICI 2b/3 following mechanical thrombectomy.

Radiographic optimization of the lateral position of the knee joint aided by CT images and the maximum intensity projection technique

Background Standard lateral knee-joint X-ray images are crucial for the accurate diagnosis and treatment of many knee-joint-related conditions. However, it is difficult to obtain standard lateral knee-joint X-ray images in the current knee-joint lateral radiography position. Purpose To optimize the lateral position of knee joint for radiography aided by computed tomography (CT) images and the maximum intensity projection technique. Materials and methods One hundred cases of anteroposterior and lateral radiographs of knee joints were included. Of these, 50 cases were for lateral radiography in conventional position, and the other 50 cases were for lateral radiography in optimized position. The optimized position was acquired by a retrospective analysis of one hundred cases of knee-joint CT images. The quality of the X-ray images in optimized group was compared with those in conventional group. The data were statistically analyzed using the Mann–Whitney U test. Results There were differences in the optimized position between males and females. The posterior condyles of the femoral epiphysis in optimized group were in perfect superimposition for most patients. However, the ones in conventional group were not. The average quality score of the lateral knee-joint X-ray images in optimized position was 3.76 ± 0.98, which is much higher than the 1.84 ± 1.15 score in conventional position. Moreover, the difference in the average quality score was statistically significant (P < 0.05). Conclusion Optimization of the lateral position of knee joint for radiography is possible with the aid of CT images and the maximum intensity projection technique.

3D-Printed Coronary Plaques to Simulate High Calcification in the Coronary Arteries for Investigation of Blooming Artifacts

The diagnostic value of coronary computed tomography angiography (CCTA) is significantly affected by high calcification in the coronary arteries owing to blooming artifacts limiting its accuracy in assessing the calcified plaques. This study aimed to simulate highly calcified plaques in 3D-printed coronary models. A combination of silicone + 32.8% calcium carbonate was found to produce 800 HU, representing extensive calcification. Six patient-specific coronary artery models were printed using the photosensitive polyurethane resin and a total of 22 calcified plaques with diameters ranging from 1 to 4 mm were inserted into different segments of these 3D-printed coronary models. The coronary models were scanned on a 192-slice CT scanner with 70 kV, pitch of 1.4, and slice thickness of 1 mm. Plaque attenuation was measured between 1100 and 1400 HU. Both maximum-intensity projection (MIP) and volume rendering (VR) images (wide and narrow window widths) were generated for measuring the diameters of these calcified plaques. An overestimation of plaque diameters was noticed on both MIP and VR images, with measurements on the MIP images close to those of the actual plaque sizes (<10% deviation), and a large measurement discrepancy observed on the VR images (up to 50% overestimation). This study proves the feasibility of simulating extensive calcification in coronary arteries using a 3D printing technique to develop calcified plaques and generate 3D-printed coronary models.

Faster Region-Based Convolutional Neural Network in the Classification of Different Parkinsonism Patterns of the Striatum on Maximum Intensity Projection Images of [18F]FP-CIT Positron Emission Tomography

The aim of this study was to compare the performance of a deep-learning convolutional neural network (Faster R-CNN) model to detect imaging findings suggestive of idiopathic Parkinson’s disease (PD) based on [18F]FP-CIT PET maximum intensity projection (MIP) images versus that of nuclear medicine (NM) physicians. The anteroposterior MIP images of the [18F]FP-CIT PET scan of 527 patients were classified as having PD (139 images) or non-PD (388 images) patterns according to the final diagnosis. Non-PD patterns were classified as overall-normal (ONL, 365 images) and vascular parkinsonism with definite defects or prominently decreased dopamine transporter binding (dVP, 23 images) patterns. Faster R-CNN was trained on 120 PD, 320 ONL, and 16 dVP pattern images and tested on the 19 PD, 45 ONL, and seven dVP patterns images. The performance of the Faster R-CNN and three NM physicians was assessed using receiver operating characteristics curve analysis. The difference in performance was assessed using Cochran’s Q test, and the inter-rater reliability was calculated. Faster R-CNN showed high accuracy in differentiating PD from non-PD patterns and also from dVP patterns, with results comparable to those of NM physicians. There were no significant differences in the area under the curve and performance. The inter-rater reliability among Faster R-CNN and NM physicians showed substantial to almost perfect agreement. The deep-learning model accurately differentiated PD from non-PD patterns on MIP images of [18F]FP-CIT PET, and its performance was comparable to that of NM physicians.