Generative Adversarial Networks Can Create High Quality Artificial Prostate Cancer Magnetic Resonance Images.

The recent integration of open-source data with machine learning models, especially in the medical field, has opened new doors to studying disease progression and/or regression. However, the ability to use medical data for machine learning approaches is limited by the specificity of data for a particular medical condition. In this context, the most recent technologies, like generative adversarial networks (GANs), are being looked upon as a potential way to generate high-quality synthetic data that preserve the clinical variability of a condition. However, despite some success, GAN model usage remains largely minimal when depicting the heterogeneity of a disease such as prostate cancer. Previous studies from our group members have focused on automating the quantitative multi-parametric magnetic resonance imaging (mpMRI) using habitat risk scoring (HRS) maps on the prostate cancer patients in the BLaStM trial. In the current study, we aimed to use the images from the BLaStM trial and other sources to train the GAN models, generate synthetic images, and validate their quality. In this context, we used T2-weighted prostate MRI images as training data for Single Natural Image GANs (SinGANs) to make a generative model. A deep learning semantic segmentation pipeline trained the model to segment the prostate boundary on 2D MRI slices. Synthetic images with a high-level segmentation boundary of the prostate were filtered and used in the quality control assessment by participating scientists with varying degrees of experience (more than ten years, one year, or no experience) to work with MRI images. Results showed that the most experienced participating group correctly identified conventional vs. synthetic images with 67% accuracy, the group with one year of experience correctly identified the images with 58% accuracy, and the group with no prior experience reached 50% accuracy. Nearly half (47%) of the synthetic images were mistakenly evaluated as conventional. Interestingly, in a blinded quality assessment, a board-certified radiologist did not significantly differentiate between conventional and synthetic images in the context of the mean quality of synthetic and conventional images. Furthermore, to validate the usability of the generated synthetic images from prostate cancer MRIs, we subjected these to anomaly detection along with the original images. Importantly, the success rate of anomaly detection for quality control-approved synthetic data in phase one corresponded to that of the conventional images. In sum, this study shows promise that high-quality synthetic images from MRIs can be generated using GANs. Such an AI model may contribute significantly to various clinical applications which involve supervised machine-learning approaches.

Where is My Spleen? - A Case of Splenosis Diagnosed Years Later after Splenectomy.

Hepatic splenosis was first described in 1939 and is a rare condition that results from splenic trauma or splenectomy. A 43-year-old man with a past medical history significant for a prior splenectomy was admitted to the hospital due to right upper quadrant pain for two days. Magnetic resonance imaging (MRI) of the abdomen suggested features of hepatic adenoma, however, a percutaneous biopsy showed the mass within the liver to be a discrete collection of splenic tissue, apparently the result of a traumatic splenic rupture years ago. Hepatic splenosis is a rare entity, and due to the asymptomatic nature of this condition, most cases are found incidentally after different imaging modalities are done. The management of this entity is based on conservative measures. We report this case to emphasize that in the appropriate clinical setting, hepatic splenosis should be considered in the differential diagnosis of a patient with a homogenous well-circumscribed liver mass.

Small pulmonary artery defects are not reliable indicators of pulmonary embolism.

OBJECTIVES: To evaluate the rate of agreement of pulmonary embolism diagnosis in computed tomography (CT) pulmonary angiogram studies and to evaluate the rate of inaccurate interpretations in the community hospital setting. METHODS: Using the keywords "pulmonary embolism/embolus/emboli," the radiology information system was searched for CT pulmonary angiograms performed over a 3-year period at three U.S. community hospitals. Studies containing probable or definite pulmonary emboli were independently reviewed by four subspecialty thoracic radiologists. MEASUREMENTS AND MAIN RESULTS: Agreement about the presence of pulmonary embolism progressively decreased with decreasing diameter of pulmonary vascular lesions (P < 0.0001). There was a sharp fall in observer agreement for pulmonary embolism of subsegmental lesions (P < 0.0001). The frequency of agreement decreased with decreasing quality of the imaging examination (P < 0.0001). Community radiologists were prone to false-positive pulmonary embolism diagnosis of subsegmental and/or small pulmonary arterial defects. The probability of a false-positive diagnosis and indeterminate examinations progressively increased with: (1) more peripheral location of the lesion, (2) decreased size (short-axis diameter) of the lesion, and (3) diminishing quality of the CT examination. Forty-eight of 177 (27%) of subsegmental vascular defects identified by community radiologists were deemed indeterminate, and 27 of 177 (15%) of subsegmental vascular defects were judged to be false positive for pulmonary embolism by the consensus diagnosis. Fifty-four of 274 (20%) vascular defects with short axis less than 6 mm were indeterminate for pulmonary embolism, and 37 of 274 (14%) of vascular defects with short axis less than 6 mm were false positive for pulmonary embolism. Eleven of 13 (85%) of vascular lesions identified as pulmonary emboli on the lowest-quality CT examinations were false positive or indeterminate for pulmonary embolism. False-positive examinations were most often due to respiratory motion artifact (19/38, 50%). CONCLUSIONS: There is relatively poor interobserver agreement for subsegmental and/or small pulmonary artery defects, especially in CT pulmonary angiograms degraded by technical artifacts. These factors can lead to an increased frequency of inaccurate interpretation or indeterminate diagnosis of subsegmental and/or small defects. Caution is indicated in interpreting the significance of small vascular defects in CT pulmonary angiograms.

Assessment of age-related morphological and functional changes of selected structures of the head and neck by computed tomography, magnetic resonance imaging, and positron emission tomography.

The head and neck is a complex anatomical region that can be evaluated using many imaging modalities. It is important to discern normal structures from ones that are affected by disease and to study how these structures change in their morphological and functional properties with aging. Therefore, using magnetic resonance imaging (MRI), we retrospectively evaluated volumes of the parotid glands, submandibular glands, thyroid gland, tongue, soft palate, and lingual tonsils in 64 subjects ages 13 to 81 years. Volume, attenuation (HU), and metabolic activity (maximum SUV) of the parotid, submandibular, and thyroid glands were assessed retrospectively using positron emission tomography/computed tomography (PET/CT) imaging in 35 subjects ages 10 to 76 years. Metabolic activity (maximum SUV) of the parotid, submandibular, and sublingual glands; tongue; adenoids; and tonsils (lingual and palatine) were evaluated retrospectively using PET imaging in 15 subjects ages 6 to 20 years. Metabolic volumetric products of the parotid, submandibular, and thyroid gland were calculated and analyzed with increasing age in subjects who underwent PET/CT imaging. Structures that exhibited statistically significant changes (P < 0.05) with increasing age included the submandibular glands, thyroid gland, soft palate, and adenoids. The CT volume of the submandibular glands increased with age, and the attenuation decreased with age with statistical significance. The thyroid gland volume, as measured using MRI, showed a statistically significant decrease with aging. The volume of the soft palate and lingual tonsils, as measured by MRI, exhibited a statistically significant decrease in volume with aging. The maximum SUV of the adenoids demonstrated a statistically significant decrease with aging. In conclusion, CT, MRI, and PET may be used to quantitatively and qualitatively assess structures of the head and neck and are useful in the assessment of structural and functional changes of these structures with aging.

Detection of age-related changes in thoracic structure and function by computed tomography, magnetic resonance imaging, and positron emission tomography.

It is useful to understand the normal changes in structure and function in the thorax that occur with age. Thus, we present the following quantitative preliminary data obtained from retrospective quantitative analysis of computed tomography (CT) and positron emission tomography (PET) examinations in subjects 0 to 90 years of age: Mean lung standard uptake values were found to significantly increase with increasing age and with increasing body mass index (BMI). Mean lung attenuation was seen to statistically significantly decrease with increasing age in subjects who had a CT scan, had a nonsignificant tendency to decrease with increasing age in subjects with a PET/CT scan, had a nonsignificant tendency to increase with increasing BMI, and was seen to significantly increase with increasing mean lung standard uptake values. Mean lung volumes were not noted to significantly change with increasing age in adult subjects whether or not they were normalized to the craniocaudal thoracic lengths, although mean lung volumes significantly increased with increasing age in pediatric subjects. Mean lung volumes had a nonsignificant tendency to decrease with increasing BMI, although normalized mean lung volumes significantly decreased with increasing BMI. Lung metabolic volumetric products were not noted to significantly change with increasing BMI or with increasing age. In this work, we also review the literature regarding normal structural and functional changes in the thorax with age.

Potential clinical role of fluorodeoxyglucose-positron emission tomography in assessing primary or secondary lymphomas of the parotid gland.

BACKGROUND: We have reviewed the utility of fluorodeoxyglucose-positron emission tomography (FDG-PET) imaging for clinical management of primary parotid lymphoma (by applying strict criteria of primary extranodal involvement by this disease) as well as of cases in which spread of the disease to the parotid gland had occurred as a consequence of primary nodal disease. PATIENTS AND METHODS: A total of 9 cases of parotid lymphoma (5 primary parotid lymphoma and 4 cases of combined nodal/extranodal lymphoma with parotid involvement) were identified and analyzed for this study. A retrospective review of the clinical records, radiologic data, and pathology results was carried out for assessment of the natural course and FDG-PET results in this disease. These patients had undergone conventional whole-body FDGPET or PET/computed tomography for initial or posttherapy monitoring purposes. RESULTS: All cases in both subgroups had unilateral parotid involvement. Fluorodeoxyglucose uptake was focal, and visual assessment was sufficient to detect the disease in all the cases with a sensitivity of 100% in primary and secondary lymphoma of the parotid, independent of the histologic subtypes. The maximum standard uptake value in untreated cases of diffuse large B-cell lymphoma and follicular non-Hodgkin lymphoma of parotid was higher (10.4 and 10.2, respectively) than that of primary parotid marginal zone lymphoma (5.2). Postchemotherapeutic remission was correctly determined by PET in all 3 patients who underwent chemotherapy. The parotid involvement was noted at diagnosis in 2 cases and in the remaining ones up to 30 months after initial diagnosis. Fluorodeoxyglucose uptake was focal and distinct in all cases, and in 1 patient with a parotid nodule as small as 0.5 cm, the lesion was clearly visualized. The maximum standard uptake values in the posttreatment scenario varied from 1.3 to 1.9, which are within the range of what has been observed in the normal parotids. In 4 of 5 patients who underwent treatment monitoring, complete metabolic response in PET was noted in advance of size criteria by radiologic techniques for complete response. CONCLUSION: Despite the known physiologic FDG uptake in parotid glands, FDG-PET appears to be of potential value in managing patients with parotid lymphoma in various stages of disease, including diagnosis and monitoring for therapeutic response.