Early Prediction of Progression to Alzheimer's Disease using Multi-Modality Neuroimages by a Novel Ordinal Learning Model ADPacer.

Machine learning has shown great promise for integrating multi-modality neuroimaging datasets to predict the risk of progression/conversion to Alzheimer's Disease (AD) for individuals with Mild Cognitive Impairment (MCI). Most existing work aims to classify MCI patients into converters versus non-converters using a pre-defined timeframe. The limitation is a lack of granularity in differentiating MCI patients who convert at different paces. Progression pace prediction has important clinical values, which allow from more personalized interventional strategies, better preparation of patients and their caregivers, and facilitation of patient selection in clinical trials. We proposed a novel ADPacer model which formulated the pace prediction into an ordinal learning problem with a unique capability of leveraging training samples with label ambiguity to augment the training set. This capability differentiates ADPacer from existing ordinal learning algorithms. We applied ADPacer to MCI patient cohorts from the Alzheimer's Disease Neuroimaging Initiative (ADNI) and the Australian Imaging, Biomarker & Lifestyle Flagship Study of Ageing (AIBL), and demonstrated the superior performance of ADPacer compared to existing ordinal learning algorithms. We also integrated the SHapley Additive exPlanations (SHAP) method with ADPacer to assess the contributions from different modalities to the model prediction. The findings are consistent with the AD literature.

Primary B cell lymphoma of the patella presenting as anterior knee pain: A case report.

Case: A 76-year-old female with a history of multiple falls on the patella presented with worsening knee pain, swelling, and reduced range of motion. Radiographs revealed tricompartmental osteoarthritis and subtle erosion of the posterior cortex of the patella that was missed by the referring orthopedist and radiologist. Her persistent pain prompted an MRI, revealing a mass later confirmed to be primary diffuse large B-cell lymphoma of the patella after biopsy and appropriate metastatic workup. Conclusion: Primary B-cell patellar lymphoma is a rarely described cause of knee pain requiring broad differential diagnosis and multidisciplinary workup in order to avoid erroneous treatment and ensure prompt oncologic treatment.

Self-Supervised Contrastive Learning to Predict the Progression of Alzheimer's Disease with 3D Amyloid-PET.

Early diagnosis of Alzheimer's disease (AD) is an important task that facilitates the development of treatment and prevention strategies, and may potentially improve patient outcomes. Neuroimaging has shown great promise, including the amyloid-PET, which measures the accumulation of amyloid plaques in the brain-a hallmark of AD. It is desirable to train end-to-end deep learning models to predict the progression of AD for individuals at early stages based on 3D amyloid-PET. However, commonly used models are trained in a fully supervised learning manner, and they are inevitably biased toward the given label information. To this end, we propose a selfsupervised contrastive learning method to accurately predict the conversion to AD for individuals with mild cognitive impairment (MCI) with 3D amyloid-PET. The proposed method, SMoCo, uses both labeled and unlabeled data to capture general semantic representations underlying the images. As the downstream task is given as classification of converters vs. non-converters, unlike the general self-supervised learning problem that aims to generate task-agnostic representations, SMoCo additionally utilizes the label information in the pre-training. To demonstrate the performance of our method, we conducted experiments on the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset. The results confirmed that the proposed method is capable of providing appropriate data representations, resulting in accurate classification. SMoCo showed the best classification performance over the existing methods, with AUROC = 85.17%, accuracy = 81.09%, sensitivity = 77.39%, and specificity = 82.17%. While SSL has demonstrated great success in other application domains of computer vision, this study provided the initial investigation of using a proposed self-supervised contrastive learning model, SMoCo, to effectively predict MCI conversion to AD based on 3D amyloid-PET.

A Mutual Knowledge Distillation-Empowered AI Framework for Early Detection of Alzheimer's Disease Using Incomplete Multi-Modal Images.

Early detection of Alzheimer's Disease (AD) is crucial to ensure timely interventions and optimize treatment outcomes for patients. While integrating multi-modal neuroimages, such as MRI and PET, has shown great promise, limited research has been done to effectively handle incomplete multi-modal image datasets in the integration. To this end, we propose a deep learning-based framework that employs Mutual Knowledge Distillation (MKD) to jointly model different sub-cohorts based on their respective available image modalities. In MKD, the model with more modalities (e.g., MRI and PET) is considered a teacher while the model with fewer modalities (e.g., only MRI) is considered a student. Our proposed MKD framework includes three key components: First, we design a teacher model that is student-oriented, namely the Student-oriented Multi-modal Teacher (SMT), through multi-modal information disentanglement. Second, we train the student model by not only minimizing its classification errors but also learning from the SMT teacher. Third, we update the teacher model by transfer learning from the student's feature extractor because the student model is trained with more samples. Evaluations on Alzheimer's Disease Neuroimaging Initiative (ADNI) datasets highlight the effectiveness of our method. Our work demonstrates the potential of using AI for addressing the challenges of incomplete multi-modal neuroimage datasets, opening new avenues for advancing early AD detection and treatment strategies.

Self-Supervised Contrastive Learning to Predict Alzheimer's Disease Progression with 3D Amyloid-PET.

Early diagnosis of Alzheimer's disease (AD) is an important task that facilitates the development of treatment and prevention strategies and may potentially improve patient outcomes. Neuroimaging has shown great promise, including the amyloid-PET which measures the accumulation of amyloid plaques in the brain - a hallmark of AD. It is desirable to train end-to-end deep learning models to predict the progression of AD for individuals at early stages based on 3D amyloid-PET. However, commonly used models are trained in a fully supervised learning manner and they are inevitably biased toward the given label information. To this end, we propose a self-supervised contrastive learning method to predict AD progression with 3D amyloid-PET. It uses unlabeled data to capture general representations underlying the images. As the downstream task is given as classification, unlike the general self-supervised learning problem that aims to generate task-agnostic representations, we also propose a loss function to utilize the label information in the pre-training. To demonstrate the performance of our method, we conducted experiments on the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset. The results confirmed that the proposed method is capable of providing appropriate data representations, resulting in accurate classification.

A template-free, more environmentally friendly approach for glass micro-texturing.

Micron and nanometer size textured silicate glass surfaces are of interest in consumer electronics, photovoltaics, and biosensing applications. Typically, texturing glass surfaces requires applying a patterned mask or a pre-etching treatment (e.g. sandblasting) on the glass substrate, followed by a mask transferring or etching process using a fluoride-containing compound. The major challenges of such a process are the complexity and cost of masking, and the safety and environmental concerns around the usage and disposal of hydrofluoric acid. Here, we describe a template-free method to construct micron-sized and submicron-sized texture on isotropic glass surfaces in one step. The new texturing mechanisms are well supported by experimental data and peridynamic simulations. With this novel strategy, the etchant uses fluoride-free chemicals such as citric acid to texture silicate glass. Etchant concentration, etch temperature, time, and additives are the primary parameters that dictate the texturing process. Surface feature size and depth can be independently controlled by tuning the leaching and chemical polishing process. We hope this study can trigger more research on novel and more environmentally friendly texturing of isotropic materials.

A Novel Transfer Learning Model for Predictive Analytics using Incomplete Multimodality Data.

Multimodality datasets are becoming increasingly common in various domains to provide complementary information for predictive analytics. One significant challenge in fusing multimodality data is that the multiple modalities are not universally available for all samples due to cost and accessibility constraints. This results in a unique data structure called Incomplete Multimodality Dataset (IMD). We propose a novel Incomplete-Multimodality Transfer Learning (IMTL) model that builds a predictive model for each sub-cohort of samples with the same missing modality pattern, and meanwhile couples the model estimation processes for different sub-cohorts to allow for transfer learning. We develop an Expectation-Maximization (EM) algorithm to estimate the parameters of IMTL and further extend it to a collaborative learning paradigm that is specifically valuable for patient privacy preservation in health care applications. We prove two advantageous properties of IMTL: the ability for out-of-sample prediction and a theoretical guarantee for a larger Fisher information compared with models without transfer learning. IMTL is applied to diagnosis and prognosis of the Alzheimer's Disease (AD) at an early stage called Mild Cognitive Impairment (MCI) using incomplete multimodality imaging data. IMTL achieves higher accuracy than competing methods without transfer learning. Supplementary materials are available for this article on the publisher's website.

Computed Tomography Findings in Vernix Caseosa Peritonitis.

INTRODUCTION: Vernix caseosa peritonitis (VCP) is a rare peripartum complication secondary to the introduction of fetal vernix into the maternal peritoneal cavity. Vernix caseosa peritonitis typically manifests a few hours to days after a cesarian section and is often initially misdiagnosed as a more common disease process resulting in delayed diagnosis. We report the computed tomography (CT) findings in 2 patients with VCP and reviewed the previously reported CT findings of VCP. CASES: Two patients, aged 17 and 24 years, presented with signs and symptoms of peritonitis within days of undergoing a cesarian section. In both cases, CT scans of the abdomen and pelvis demonstrated ascites and multiple small, well-defined, peripherally enhancing, cystic peritoneal nodules which were most prominent around the liver and became larger and more numerous over time. Antibiotic therapy was not effective, subsequent laparoscopic peritoneal biopsy demonstrated VCP, and patients were successfully treated with lavage and the addition of intravenous steroids. CONCLUSIONS: Vernix caseosa peritonitis is an underrecognized disorder that is most often mistaken for other more common causes of peritonitis. In the setting of peripartum peritonitis, the CT findings of ascites with multiple small, well-defined, peripherally enhancing, cystic peritoneal nodules, especially adjacent to the liver, which grow in size and number strongly suggests VCP.

The Use of Periareolar Topical Lidocaine Jelly After Needle Localization May Reduce Pain of Subsequent Lymphoscintigraphy.

Purpose Technetium Tc-99m sulfur colloid (99mTc-SC) breast lymphoscintigraphy is commonly performed to identify the sentinel lymph node (SLN) in patients diagnosed with breast carcinoma undergoing lumpectomy. The purpose of this report is to describe how the use of 2% topical lidocaine jelly immediately after the completion of needle localization and prior to scintigraphy may substantially reduce pain associated with the injection of 99mTc-SC. Materials and methods This was a quality improvement project. Patients were asked to score the severity of pain associated with the periareolar 99mTc-SC injections for sentinel node lymphoscintigraphy. In order to decrease the discomfort, topical lidocaine was applied to the periareolar skin after the completion of the needle localization, but prior to transferring the patient from the mammography room to the nuclear medicine department for the 99mTc-SC injections. At the time of 99mTc-SC injection, patients were asked to score the pain of injection from 0 (none) to 10 (worst). Results The average pain score of the women who did not receive topical lidocaine jelly was 8 (range: 5-9). In the 10 women who received topical lidocaine jelly after needle localization, the average pain score was 2.5 (range: 1-5). Interestingly, the pain score for women who discussed the possible use of lidocaine jelly with the radiologists but still did not receive topical lidocaine jelly was also low at 6.5. For patients who received the lidocaine jelly only five minutes prior to injection, the average pain score was 6. Conclusion The application of lidocaine jelly after the conclusion of needle localization, with a 15-40-minute delay prior to periareolar injections with 99mTc-SC for sentinel node lymphoscintigraphy, appears to substantially reduce the pain associated with the injection of 99mTc-SC.

The skinny on skin: MRI features of cutaneous and subcutaneous lesions detected on body MRI studies.

The skin and subcutaneous tissues are inevitably imaged as part of most body MRI studies. Incidental or even symptomatic skin lesions may, therefore, be detected and present a diagnostic challenge for the radiologist. We aim to provide a comprehensive review, with illustrative examples, of the skin abnormalities encountered on body MRI studies in our busy academic radiology department.

Use of multimodality imaging and artificial intelligence for diagnosis and prognosis of early stages of Alzheimer's disease.

Alzheimer's disease (AD) is a major neurodegenerative disease and the most common cause of dementia. Currently, no treatment exists to slow down or stop the progression of AD. There is converging belief that disease-modifying treatments should focus on early stages of the disease, that is, the mild cognitive impairment (MCI) and preclinical stages. Making a diagnosis of AD and offering a prognosis (likelihood of converting to AD) at these early stages are challenging tasks but possible with the help of multimodality imaging, such as magnetic resonance imaging (MRI), fluorodeoxyglucose (FDG)-positron emission topography (PET), amyloid-PET, and recently introduced tau-PET, which provides different but complementary information. This article is a focused review of existing research in the recent decade that used statistical machine learning and artificial intelligence methods to perform quantitative analysis of multimodality image data for diagnosis and prognosis of AD at the MCI or preclinical stages. We review the existing work in 3 subareas: diagnosis, prognosis, and methods for handling modality-wise missing data-a commonly encountered problem when using multimodality imaging for prediction or classification. Factors contributing to missing data include lack of imaging equipment, cost, difficulty of obtaining patient consent, and patient drop-off (in longitudinal studies). Finally, we summarize our major findings and provide some recommendations for potential future research directions.

Xanthogranulomatous pancreatitis: A review of the imaging characteristics of this rare and often misdiagnosed lesion of the pancreas.

Xanthogranulomatous pancreatitis (XGP) is an extremely rare cause of a cystic pancreatic mass. The pathophysiology of this process is not entirely clear but likely results from a combination of duct obstruction, infection, and repeated hemorrhage. It is difficult to differentiate this inflammatory lesion from a cystic neoplasm and, therefore, in the majority of cases XGP is misdiagnosed as a neoplasm on preoperative imaging. In this report, we describe a case of XGP, the imaging characteristics of XGP, and a differential diagnosis for a cystic pancreatic lesion.

Added value and limitations of amyloid-PET imaging: review and analysis of selected cases of mild cognitive impairment and dementia.

Amyloid-positron emission tomography (PET) imaging of the brain detects elevated amyloid-beta (amyloid-ß) neuritic plaques in vivo, which can be helpful in appropriately selected cases of mild cognitive impairment (MCI) and dementia, when Alzheimer's disease remains a possible etiology, after a comprehensive clinical evaluation. We reviewed cases of cognitively impaired patients who underwent amyloid-PET imaging because of diagnostic uncertainty. Pre- and post-PET elements of diagnosis and management were first compared, to assess impact of scan results on clinical decision-making, and then an analysis of those decisions was undertaken in appropriate clinical situations, to delineate the added value and limitations of amyloid-PET imaging. The potential benefits and limitations of this diagnostic tool are important to understand in an era when the utility of such scans in clinical practice is evolving.

Imaging Surveillance in Patients After a Benign Fine-Needle Aspiration Biopsy of the Thyroid: Associated Cost and Incidence of Subsequent Cancer.

OBJECTIVE: The objective of our study was to determine patterns and cost of imaging tumor surveillance in patients after a benign fine-needle aspiration (FNA) biopsy of the thyroid in a large teaching hospital as well as the rate of subsequent cancer detection. MATERIALS AND METHODS: This cohort study was approved by the appropriate institutional review board and complied with HIPAA. All patients who had a benign thyroid FNA biopsy between January 1, 1999, and December 31, 2003, were identified from an institutional pathology database. We gathered information from electronic medical records on imaging tumor surveillance and subsequent cancer detection. Cost was determined using the facility total relative value unit and the 2014 Hospital Outpatient Prospective Payment System conversion factor. RESULTS: Between January 1, 1999, and December 31, 2003, 1685 patients had a benign thyroid FNA biopsy, 800 (47.5%) of whom underwent follow-up imaging. These patients underwent 2223 thyroid ultrasound examinations, 606 ultrasound-guided thyroid FNA biopsies, 78 thyroid scintigraphy examinations, 168 neck CTs, and 53 neck MRIs at a cost of $529,874, $176,157, $39,622, $80,580, and $53,114, respectively, for a total cost of $879,347 or $1099 per patient. The mean length of follow-up was 7.3 years, during which time 19 (2.4%) patients were diagnosed with thyroid cancer at a cost of $46,281 per cancer. Seventeen (89.5%) were diagnosed with papillary carcinoma and two (10.5%) with Hurthle cell carcinoma. CONCLUSION: Over a 5-year period, about half of the patients who had a benign thyroid FNA biopsy underwent follow-up imaging at considerable cost with a small rate of subsequent malignancy.

Transient washout of hepatic hemangiomas: Potential pitfall mimicking malignancy.

Hemangiomas are the most common tumor of the liver and distinguishing them from malignancy is important. This is a report of 3 hemangiomas in 2 patients that exhibit transient washout of gadoxetate disodium (Eovist), relative to blood pool and liver parenchyma, a characteristic that is used to diagnose hepatocellular carcinoma in at-risk patients. It is important to recognize that high-flow hemangiomas can exhibit transient washout when using a small volume of injected contrast agent. This finding is unlikely to be present on CT examinations because of the larger volume of contrast administered.