Search

Purpose To develop and validate a deep learning algorithm that predicts the final diagnosis of Alzheimer disease (AD), mild cognitive impairment, or neither at fluorine 18 ( 18 F) fluorodeoxyglucose (FDG) PET of the brain and compare its performance to that of radiologic readers. Materials and Methods Prospective 18 F-FDG PET brain images from the Alzheimer's Disease Neuroimaging Initiative (ADNI) (2109 imaging studies from 2005 to 2017, 1002 patients) and retrospective independent test set (40 imaging studies from 2006 to 2016, 40 patients) were collected. Final clinical diagnosis at follow-up was recorded. Convolutional neural network of InceptionV3 architecture was trained on 90% of ADNI data set and tested on the remaining 10%, as well as the independent test set, with performance compared to radiologic readers. Model was analyzed with sensitivity, specificity, receiver operating characteristic (ROC), saliency map, and t-distributed stochastic neighbor embedding. Results The algorithm achieved area under the ROC curve of 0.98 (95% confidence interval: 0.94, 1.00) when evaluated on predicting the final clinical diagnosis of AD in the independent test set (82% specificity at 100% sensitivity), an average of 75.8 months prior to the final diagnosis, which in ROC space outperformed reader performance (57% [four of seven] sensitivity, 91% [30 of 33] specificity; P < .05). Saliency map demonstrated attention to known areas of interest but with focus on the entire brain. Conclusion By using fluorine 18 fluorodeoxyglucose PET of the brain, a deep learning algorithm developed for early prediction of Alzheimer disease achieved 82% specificity at 100% sensitivity, an average of 75.8 months prior to the final diagnosis. © RSNA, 2018 Online supplemental material is available for this article. See also the editorial by Larvie in this issue.

Purpose: To serially monitor bone remodeling in the swine femur after magnetic resonance (MR) imaging-guided high-intensity focused ultrasound (HIFU) ablation with MR imaging, computed tomography (CT), sodium fluorine 18 (Na(18)F)-positron emission tomography (PET), and histopathologic examination, as a function of sonication energy., Materials and Methods: Experimental procedures received approval from the local institutional animal care and use committee. MR imaging-guided HIFU was used to create distal and proximal ablations in the right femurs of eight pigs. The energy used at the distal target was higher (mean, 419 J; range, 390-440 J) than that used at the proximal target (mean, 324 J; range, 300-360 J). Imaging was performed before and after ablation with 3.0-T MR imaging and 64-section CT. Animals were reevaluated at 3 and 6 weeks with MR imaging (n = 8), CT (n = 8), Na(18)F-PET (n = 4), and histopathologic examination (n = 4). Three-dimensional ablation lengths were measured on contrast material-enhanced MR images, and bone remodeling in the cortex was measured on CT images., Results: Ablation sizes at MR imaging 3 and 6 weeks after MR imaging-guided HIFU ablation were similar between proximal (low-energy) and distal (high-energy) lesions (average, 8.7 × 21.9 × 16.4 mm). However, distal ablation lesions (n = 8) demonstrated evidence of subperiosteal new bone formation at CT, with a subtle focus of new ossification at 3 weeks and a larger focus of ossification at 6 weeks. New bone formation was associated with increased uptake at Na(18)F-PET in three of four animals; this was confirmed at histopathologic examination in four of four animals., Conclusion: MR imaging-guided HIFU ablation of bone may result in progressive remodeling, with both subcortical necrosis and subperiosteal new bone formation. This may be related to the use of high energies. MR imaging, CT, and PET are suitable noninvasive techniques to monitor bone remodeling after MR imaging-guided HIFU ablation., (© RSNA, 2014.)

Purpose: To assess the relationship between parameters measured on dynamic contrast material-enhanced (DCE) magnetic resonance (MR) imaging and fluorine 18 fluorodeoxyglucose (FDG) positron emission tomography (PET)/computed tomography (CT) in primary invasive breast cancer., Materials and Methods: This HIPAA-compliant study was a retrospective review of medical records and therefore approved by the institutional review board without the requirement for informed consent. Patients with a diagnosis of invasive breast cancer from January 2005 through December 2009 who underwent both DCE MR imaging and FDG PET/CT before treatment initiation were retrospectively identified. Fractional volumes were measured for ranges of signal enhancement ratio (SER) values from DCE MR imaging data and compared with maximum standardized uptake values (SUVmax) from FDG PET/CT data. Linear regression analysis was performed to clarify the relationship between SER and SUVmax, adjusting for tumor size, pathologic grade, and receptor status., Results: Analyzed were 117 invasive breast cancers in 117 patients. Overall, a higher percentage of high washout kinetics was positively associated with SUVmax (1.57% increase in SUVmax per 1% increase in high washout; P = .020), and a higher percentage of low plateau kinetics was negatively associated with SUVmax (1.19% decrease in SUVmax per 1% increase in low plateau; P = .003). These relationships were strongest among triple-negative (TN) tumors (4.34% increase in SUVmax per 1% increase in high washout and 2.65% decrease in SUVmax per 1% increase in low plateau; P = .018 and .004, respectively)., Conclusion: In invasive breast carcinoma, there is a positive relationship between the percentage of high washout and SUVmax and a negative relationship between the percentage of low plateau and SUVmax. These results are stronger in TN tumors., Supplemental Material: http://radiology.rsna.org/lookup/suppl/doi:10.1148/radiol.13130058/-/DC1., (RSNA, 2013)

Purpose: To evaluate the role of positron emission tomography (PET)/computed tomography (CT) in the differentiation of normal thymus from mediastinal lymphoma and lymphoma recurrence in pediatric patients., Materials and Methods: The study was approved by the institutional review board, and informed consent was waived. The study was HIPAA compliant. Two hundred eighty-two fluorine 18 fluorodeoxyglucose PET/CT studies in 75 pediatric oncology patients were reviewed retrospectively. Patients were divided into four groups: anterior mediastinal lymphoma (group A, n=16), anterior mediastinal lymphoma with subsequent recurrence (group B, n=5), lymphoma outside the mediastinum (group C, n=16), and other malignant tumors outside the thymus (group D, n=38). Analyses included measurements of the maximum anteroposterior and transverse dimensions of the anterior mediastinal mass or thymus on axial CT images and measurements of maximum standardized uptake values of anterior mediastinal mass, thymus (SUVt), and bone marrow at the level of the fifth lumbar vertebra (SUVb) on PET images. Quantitative parameters were compared by using an analysis of variance test., Results: Mean prechemotherapy SUVt was 4.82 for group A, 8.45 for group B, 2.00 for group C, and 2.09 for group D. Mean postchemotherapy SUVt for group B was 4.74. Thymic rebound (mean SUVt, 2.89) was seen in 44% of patients at a mean interval of 10 months from the end of chemotherapy. The differences between prechemotherapy SUVt of mediastinal lymphoma and normal thymus and postchemotherapy SUVt of lymphoma recurrence and thymic rebound were highly significant (P<.001)., Conclusion: SUVt is a sensitive predictor for differentiation of normal thymus or thymic rebound from mediastinal lymphoma. SUVt of 3.4 or higher is a strong predictor of mediastinal lymphoma., (© RSNA, 2011)

Purpose: To assess the usefulness of 2-[fluorine-18]fluoro-2-deoxy-D-glucose (FDG) positron emission tomography (PET) of the head and neck in locating occult primary lesions in patients with metastatic cervical adenopathy., Materials and Methods: Seventeen patients with metastatic cervical adenopathy of unknown primary origin were referred for FDG PET of the head and neck. All patients had undergone correlative anatomic imaging within 1 month of FDG PET. Surgical, clinical, and histopathologic findings were used to assess the performance of FDG PET., Results: Increased apical lung uptake at FDG PET led to a biopsy-proved diagnosis of primary lung cancer in two patients. Of the remaining 15 patients, 10 had a focus of increased activity; directed biopsy of these sites led to confirmation of a primary carcinoma in seven patients. Correlative anatomic imaging failed to demonstrate the primary sites of disease in two of these seven patients. None of the five patients with negative FDG PET studies have manifested evidence of a primary site of disease during follow-up of 8-42 months (mean, 29 months)., Conclusion: FDG PET allows effective localization of the unknown primary site of origin in metastatic head and neck cancer and can contribute substantially to patient care.

Purpose: To quantitate the changes induced in uptake of the glucose analog 2-[fluorine-18]fluoro-2-deoxy-D-glucose (FDG) in normal structures in the head and neck and compare these to the change in uptake in malignant structures in patients with head and neck tumors undergoing radiation therapy., Materials and Methods: Eleven patients with biopsy-confirmed squamous cell carcinoma of the head and neck were studied before, during, and after a 6-week course of radiation therapy with positron emission tomography (PET)-FDG imaging. A ratio of FDG uptake in the structure compared with that in the cerebellum (termed metabolic ratio) within and outside of the field of radiation was determined in the adenoids; lingual and palatine tonsils; parotid, submandibular, and sublingual glands; and nasal turbinates, soft palate, and gingiva., Results: The average metabolic ratio in the tonsils, nasal turbinates, soft palate, and gingiva did not change significantly with treatment., Conclusion: FDG uptake in normal structures does not change with radiation therapy. This fact is in marked contrast to the FDG uptake in squamous cell carcinomas in the head and neck, which decrease dramatically with treatment (P < .005).