Search

Your search keyword '"Ma, Martin"' showing total 1,034 results
Start Over Author "Ma, Martin"

Refine your results

more »

more »

more »

more »

more »
1,034 results on '"Ma, Martin"'

1. ELIXR: Towards a general purpose X-ray artificial intelligence system through alignment of large language models and radiology vision encoders

Author

Xu, Shawn, Yang, Lin, Kelly, Christopher, Sieniek, Marcin, Kohlberger, Timo, Ma, Martin, Weng, Wei-Hung, Kiraly, Atilla, Kazemzadeh, Sahar, Melamed, Zakkai, Park, Jungyeon, Strachan, Patricia, Liu, Yun, Lau, Chuck, Singh, Preeti, Chen, Christina, Etemadi, Mozziyar, Kalidindi, Sreenivasa Raju, Matias, Yossi, Chou, Katherine, Corrado, Greg S., Shetty, Shravya, Tse, Daniel, Prabhakara, Shruthi, Golden, Daniel, Pilgrim, Rory, Eswaran, Krish, and Sellergren, Andrew

Subjects
Computer Science - Computer Vision and Pattern Recognition, Electrical Engineering and Systems Science - Image and Video Processing
Abstract

In this work, we present an approach, which we call Embeddings for Language/Image-aligned X-Rays, or ELIXR, that leverages a language-aligned image encoder combined or grafted onto a fixed LLM, PaLM 2, to perform a broad range of chest X-ray tasks. We train this lightweight adapter architecture using images paired with corresponding free-text radiology reports from the MIMIC-CXR dataset. ELIXR achieved state-of-the-art performance on zero-shot chest X-ray (CXR) classification (mean AUC of 0.850 across 13 findings), data-efficient CXR classification (mean AUCs of 0.893 and 0.898 across five findings (atelectasis, cardiomegaly, consolidation, pleural effusion, and pulmonary edema) for 1% (~2,200 images) and 10% (~22,000 images) training data), and semantic search (0.76 normalized discounted cumulative gain (NDCG) across nineteen queries, including perfect retrieval on twelve of them). Compared to existing data-efficient methods including supervised contrastive learning (SupCon), ELIXR required two orders of magnitude less data to reach similar performance. ELIXR also showed promise on CXR vision-language tasks, demonstrating overall accuracies of 58.7% and 62.5% on visual question answering and report quality assurance tasks, respectively. These results suggest that ELIXR is a robust and versatile approach to CXR AI.

Published
2023

2. Factorized Contrastive Learning: Going Beyond Multi-view Redundancy

Author

Liang, Paul Pu, Deng, Zihao, Ma, Martin, Zou, James, Morency, Louis-Philippe, and Salakhutdinov, Ruslan

Subjects
Computer Science - Machine Learning, Computer Science - Artificial Intelligence, Computer Science - Computation and Language, Computer Science - Computer Vision and Pattern Recognition, Computer Science - Multimedia
Abstract

In a wide range of multimodal tasks, contrastive learning has become a particularly appealing approach since it can successfully learn representations from abundant unlabeled data with only pairing information (e.g., image-caption or video-audio pairs). Underpinning these approaches is the assumption of multi-view redundancy - that shared information between modalities is necessary and sufficient for downstream tasks. However, in many real-world settings, task-relevant information is also contained in modality-unique regions: information that is only present in one modality but still relevant to the task. How can we learn self-supervised multimodal representations to capture both shared and unique information relevant to downstream tasks? This paper proposes FactorCL, a new multimodal representation learning method to go beyond multi-view redundancy. FactorCL is built from three new contributions: (1) factorizing task-relevant information into shared and unique representations, (2) capturing task-relevant information via maximizing MI lower bounds and removing task-irrelevant information via minimizing MI upper bounds, and (3) multimodal data augmentations to approximate task relevance without labels. On large-scale real-world datasets, FactorCL captures both shared and unique information and achieves state-of-the-art results on six benchmarks, Comment: NeurIPS 2023. Code available at: https://github.com/pliang279/FactorCL

Published
2023

3. Understanding Masked Autoencoders via Hierarchical Latent Variable Models

Author

Kong, Lingjing, Ma, Martin Q., Chen, Guangyi, Xing, Eric P., Chi, Yuejie, Morency, Louis-Philippe, and Zhang, Kun

Subjects
Computer Science - Machine Learning, Computer Science - Computer Vision and Pattern Recognition
Abstract

Masked autoencoder (MAE), a simple and effective self-supervised learning framework based on the reconstruction of masked image regions, has recently achieved prominent success in a variety of vision tasks. Despite the emergence of intriguing empirical observations on MAE, a theoretically principled understanding is still lacking. In this work, we formally characterize and justify existing empirical insights and provide theoretical guarantees of MAE. We formulate the underlying data-generating process as a hierarchical latent variable model and show that under reasonable assumptions, MAE provably identifies a set of latent variables in the hierarchical model, explaining why MAE can extract high-level information from pixels. Further, we show how key hyperparameters in MAE (the masking ratio and the patch size) determine which true latent variables to be recovered, therefore influencing the level of semantic information in the representation. Specifically, extremely large or small masking ratios inevitably lead to low-level representations. Our theory offers coherent explanations of existing empirical observations and provides insights for potential empirical improvements and fundamental limitations of the masking-reconstruction paradigm. We conduct extensive experiments to validate our theoretical insights., Comment: CVPR 2023 Highlight

Published
2023

4. Mask R-CNN assisted 2.5D object detection pipeline of 68Ga-PSMA-11 PET/CT-positive metastatic pelvic lymph node after radical prostatectomy from solely CT imaging

Author

Xu, Di, Ma, Martin, Cao, Minsong, Kishan, Amar U, Nickols, Nicholas G, Scalzo, Fabien, and Sheng, Ke

Subjects
Prostate Cancer, Urologic Diseases, Cancer, Aging, Bioengineering, Biomedical Imaging, 4.2 Evaluation of markers and technologies, Detection, screening and diagnosis, Male, Humans, Positron Emission Tomography Computed Tomography, Prostate, Artificial Intelligence, Gallium Radioisotopes, Prostatic Neoplasms, Prostatectomy, Tomography, X-Ray Computed, Lymph Nodes
Abstract

Prostate-specific membrane antigen (PSMA) positron emission tomography (PET)/computed tomography (CT) is a molecular and functional imaging modality with better restaging accuracy over conventional imaging for detecting prostate cancer in men suspected of lymph node (LN) progression after definitive therapy. However, the availability of PSMA PET/CT is limited in both low-resource settings and for repeating imaging surveillance. In contrast, CT is widely available, cost-effective, and routinely performed as part of patient follow-up or radiotherapy workflow. Compared with the molecular activities, the morphological and texture changes of subclinical LNs in CT are subtle, making manual detection of positive LNs infeasible. Instead, we harness the power of artificial intelligence for automated LN detection on CT. We examined 68Ga-PSMA-11 PET/CT images from 88 patients (including 739 PSMA PET/CT-positive pelvic LNs) who experienced a biochemical recurrence after radical prostatectomy and presented for salvage radiotherapy with prostate-specific antigen

Published
2023

5. Face-to-Face Contrastive Learning for Social Intelligence Question-Answering

Author

Wilf, Alex, Ma, Martin Q., Liang, Paul Pu, Zadeh, Amir, and Morency, Louis-Philippe

Subjects
Computer Science - Machine Learning, Computer Science - Artificial Intelligence, Computer Science - Computer Vision and Pattern Recognition
Abstract

Creating artificial social intelligence - algorithms that can understand the nuances of multi-person interactions - is an exciting and emerging challenge in processing facial expressions and gestures from multimodal videos. Recent multimodal methods have set the state of the art on many tasks, but have difficulty modeling the complex face-to-face conversational dynamics across speaking turns in social interaction, particularly in a self-supervised setup. In this paper, we propose Face-to-Face Contrastive Learning (F2F-CL), a graph neural network designed to model social interactions using factorization nodes to contextualize the multimodal face-to-face interaction along the boundaries of the speaking turn. With the F2F-CL model, we propose to perform contrastive learning between the factorization nodes of different speaking turns within the same video. We experimentally evaluated the challenging Social-IQ dataset and show state-of-the-art results.

Published
2022

6. Conditional Contrastive Learning with Kernel

Author

Tsai, Yao-Hung Hubert, Li, Tianqin, Ma, Martin Q., Zhao, Han, Zhang, Kun, Morency, Louis-Philippe, and Salakhutdinov, Ruslan

Subjects
Computer Science - Machine Learning
Abstract

Conditional contrastive learning frameworks consider the conditional sampling procedure that constructs positive or negative data pairs conditioned on specific variables. Fair contrastive learning constructs negative pairs, for example, from the same gender (conditioning on sensitive information), which in turn reduces undesirable information from the learned representations; weakly supervised contrastive learning constructs positive pairs with similar annotative attributes (conditioning on auxiliary information), which in turn are incorporated into the representations. Although conditional contrastive learning enables many applications, the conditional sampling procedure can be challenging if we cannot obtain sufficient data pairs for some values of the conditioning variable. This paper presents Conditional Contrastive Learning with Kernel (CCL-K) that converts existing conditional contrastive objectives into alternative forms that mitigate the insufficient data problem. Instead of sampling data according to the value of the conditioning variable, CCL-K uses the Kernel Conditional Embedding Operator that samples data from all available data and assigns weights to each sampled data given the kernel similarity between the values of the conditioning variable. We conduct experiments using weakly supervised, fair, and hard negatives contrastive learning, showing CCL-K outperforms state-of-the-art baselines.

Published
2022

7. Conditional Contrastive Learning for Improving Fairness in Self-Supervised Learning

Author

Ma, Martin Q., Tsai, Yao-Hung Hubert, Liang, Paul Pu, Zhao, Han, Zhang, Kun, Salakhutdinov, Ruslan, and Morency, Louis-Philippe

Subjects
Computer Science - Machine Learning
Abstract

Contrastive self-supervised learning (SSL) learns an embedding space that maps similar data pairs closer and dissimilar data pairs farther apart. Despite its success, one issue has been overlooked: the fairness aspect of representations learned using contrastive SSL. Without mitigation, contrastive SSL techniques can incorporate sensitive information such as gender or race and cause potentially unfair predictions on downstream tasks. In this paper, we propose a Conditional Contrastive Learning (CCL) approach to improve the fairness of contrastive SSL methods. Our approach samples positive and negative pairs from distributions conditioning on the sensitive attribute, or empirically speaking, sampling positive and negative pairs from the same gender or the same race. We show that our approach provably maximizes the conditional mutual information between the learned representations of the positive pairs, and reduces the effect of the sensitive attribute by taking it as the conditional variable. On seven fairness and vision datasets, we empirically demonstrate that the proposed approach achieves state-of-the-art downstream performances compared to unsupervised baselines and significantly improves the fairness of contrastive SSL models on multiple fairness metrics.

Published
2021

8. A Large-scale Study on Unsupervised Outlier Model Selection: Do Internal Strategies Suffice?

Author

Ma, Martin Q., Zhao, Yue, Zhang, Xiaorong, and Akoglu, Leman

Subjects
Computer Science - Machine Learning
Abstract

Given an unsupervised outlier detection task, how should one select a detection algorithm as well as its hyperparameters (jointly called a model)? Unsupervised model selection is notoriously difficult, in the absence of hold-out validation data with ground-truth labels. Therefore, the problem is vastly understudied. In this work, we study the feasibility of employing internal model evaluation strategies for selecting a model for outlier detection. These so-called internal strategies solely rely on the input data (without labels) and the output (outlier scores) of the candidate models. We setup (and open-source) a large testbed with 39 detection tasks and 297 candidate models comprised of 8 detectors and various hyperparameter configurations. We evaluate 7 different strategies on their ability to discriminate between models w.r.t. detection performance, without using any labels. Our study reveals room for progress -- we find that none would be practically useful, as they select models only comparable to a state-of-the-art detector (with random configuration).

Published
2021

9. Self-supervised Representation Learning with Relative Predictive Coding

Author

Tsai, Yao-Hung Hubert, Ma, Martin Q., Yang, Muqiao, Zhao, Han, Morency, Louis-Philippe, and Salakhutdinov, Ruslan

Subjects
Computer Science - Machine Learning, Computer Science - Information Theory
Abstract

This paper introduces Relative Predictive Coding (RPC), a new contrastive representation learning objective that maintains a good balance among training stability, minibatch size sensitivity, and downstream task performance. The key to the success of RPC is two-fold. First, RPC introduces the relative parameters to regularize the objective for boundedness and low variance. Second, RPC contains no logarithm and exponential score functions, which are the main cause of training instability in prior contrastive objectives. We empirically verify the effectiveness of RPC on benchmark vision and speech self-supervised learning tasks. Lastly, we relate RPC with mutual information (MI) estimation, showing RPC can be used to estimate MI with low variance.

Published
2021

10. Multimodal Routing: Improving Local and Global Interpretability of Multimodal Language Analysis

Author

Tsai, Yao-Hung Hubert, Ma, Martin Q., Yang, Muqiao, Salakhutdinov, Ruslan, and Morency, Louis-Philippe

Subjects
Computer Science - Computation and Language
Abstract

The human language can be expressed through multiple sources of information known as modalities, including tones of voice, facial gestures, and spoken language. Recent multimodal learning with strong performances on human-centric tasks such as sentiment analysis and emotion recognition are often black-box, with very limited interpretability. In this paper we propose Multimodal Routing, which dynamically adjusts weights between input modalities and output representations differently for each input sample. Multimodal routing can identify relative importance of both individual modalities and cross-modality features. Moreover, the weight assignment by routing allows us to interpret modality-prediction relationships not only globally (i.e. general trends over the whole dataset), but also locally for each single input sample, meanwhile keeping competitive performance compared to state-of-the-art methods.

Published
2020

11. Complex Transformer: A Framework for Modeling Complex-Valued Sequence

Author

Yang, Muqiao, Ma, Martin Q., Li, Dongyu, Tsai, Yao-Hung Hubert, and Salakhutdinov, Ruslan

Subjects
Computer Science - Machine Learning, Computer Science - Sound, Electrical Engineering and Systems Science - Audio and Speech Processing, Statistics - Machine Learning
Abstract

While deep learning has received a surge of interest in a variety of fields in recent years, major deep learning models barely use complex numbers. However, speech, signal and audio data are naturally complex-valued after Fourier Transform, and studies have shown a potentially richer representation of complex nets. In this paper, we propose a Complex Transformer, which incorporates the transformer model as a backbone for sequence modeling; we also develop attention and encoder-decoder network operating for complex input. The model achieves state-of-the-art performance on the MusicNet dataset and an In-phase Quadrature (IQ) signal dataset.

Published
2019

12. A Framework for Sharing Radiation Dose Distribution Maps in the Electronic Medical Record for Improving Multidisciplinary Patient Management

Author

Savjani, Ricky R, Salamon, Noriko, Deng, Jie, Ma, Martin, Tenn, Steve, Agazaryan, Nzhde, Hegde, John, and Kaprealian, Tania

Subjects
Medical and Biological Physics, Biomedical and Clinical Sciences, Clinical Sciences, Physical Sciences, Oncology and Carcinogenesis, Brain Disorders, Biomedical Imaging, Rare Diseases, Neurosciences, Cancer, Brain Cancer, Good Health and Well Being, Electronic Health Records, Humans, Magnetic Resonance Imaging, Radiation Dosage, Radiology Information Systems, Software, Brain/Brain Stem, CNS, MRI, Neuro-Oncology, Radiation Effects, Radiation Therapy, Radiation Therapy/Oncology, Radiosurgery, Skull Base, Spine, Technology Assessment
Abstract

Radiation oncology practices use a suite of dedicated software and hardware that are not common to other medical subspecialties, making radiation treatment history inaccessible to colleagues. A radiation dose distribution map is generated for each patient internally that allows for visualization of the dose given to each anatomic structure volumetrically; however, this crucial information is not shared systematically to multidisciplinary medical, surgery, and radiology colleagues. A framework was developed in which dose distribution volumes are uploaded onto the medical center's picture archiving and communication system (PACS) to rapidly retrieve and review exactly where, when, and to what dose a lesion or structure was treated. The ability to easily visualize radiation therapy information allows radiology clinics to incorporate radiation dose into image interpretation without direct access to radiation oncology planning software and data. Tumor board discussions are simplified by incorporating radiation therapy information collectively in real time, and daily onboard imaging can also be uploaded while a patient is still undergoing radiation therapy. Placing dose distribution information into PACS facilitates central access into the electronic medical record and provides a succinct visual summary of a patient's radiation history for all medical providers. More broadly, the radiation dose map provides greater visibility and facilitates incorporation of a patient's radiation history to improve oncologic decision making and patient outcomes. Keywords: Brain/Brain Stem, CNS, MRI, Neuro-Oncology, Radiation Effects, Radiation Therapy, Radiation Therapy/Oncology, Radiosurgery, Skull Base, Spine, Technology Assessment Supplemental material is available for this article. © RSNA, 2021 See also commentary by Khandelwal and Scarboro in this issue.

Published
2021

16. Divergence excess and basic exotropia types of intermittent exotropia: a major review. Part 2: non-surgical and surgical treatment options.

Author

Ma, Martin Ming-Leung and Scheiman, Mitchell

Subjects
*EXOTROPIA, *CLINICAL trials, *REFRACTIVE errors, *VISUAL training, *QUALITY of life
Abstract

Introduction: Currently, there is no consensus regarding the management of intermittent exotropia (IXT), which includes both surgical and non-surgical treatment options. Nonsurgical management of IXT has been suggested and includes watchful observation, patching, overminus lenses, prism, and vision therapy/orthoptics. While a significant portion of IXT patients are treated by surgery, it is reported that there is a substantial tendency for reoperation or recurrence of IXT. This paper provides a comprehensive review of non-surgical and surgical treatment options for the IXT. Methods: Search strategies involving combination of keywords including intermittent exotropia, divergence excess, basic exotropia, refractive error, glasses, spectacles, natural history, untreated, observe, occlusion, patch, overminus, overcorrecting minus, prism, vision therapy, orthoptic, anti-suppression, fusion exercise, and surgery were used in Medline. All English articles from 01/01/1900 to 01/09/2020 were reviewed. The reference list of the identified articles was also checked for additional relevant articles. Studies focused on animal models or strabismus associated with neurologic disorders or injury were excluded. The following filters were used for surgical management due to the abundance of reports: full text, randomized controlled trial, review, in the last 5 years. Results: Appropriate optical correction of refractive error is generally the starting point for all management approaches, but there is a lack of randomized clinical trial data regarding this treatment modality. Randomized clinical trial data indicate that both observation and occlusion are reasonable management options for children 3–10 years old, and there were insufficient data to recommend occlusion for children 12–35 months old. While overminus lenses were found to improve the control of IXT when assessed wearing overminus spectacles, this improvement did not persist after the treatment ended. The result of the only randomized clinical trial on the effectiveness of base-in prism indicated that this treatment is no more effective than nonprism spectacles for improving control. A recent randomized clinical trial showed that vision therapy/orthoptics is effective in improving the control of IXT when compared to observation alone. Surgery was found to alter a number of clinical characteristics of IXT, including reducing the distance and near angle of deviation, reducing photophobia, improving health-related quality of life, stereopsis, and the Newcastle Control Score. However, there are no randomized clinical trial data comparing surgery with a control group such as placebo or a no treatment observation group. Discussion: Rigorously designed clinical trials to investigate the effectiveness of non-surgical and surgical treatments for intermittent exotropia are needed. [ABSTRACT FROM AUTHOR]

Published
2024
Full Text
View/download PDF

21. Office‐based vergence and anti‐suppression therapy for the treatment of small‐to‐moderate angle intermittent exotropia: A randomised clinical trial.

Author

Ma, Martin Ming‐Leung, Kang, Ying, Scheiman, Mitchell, Chen, Qiwen, Ye, Xuelian, Pan, Liuqing, Deng, Jiayu, Su, Guangxing, Zhang, Guohui, and Chen, Xiang

Abstract

Purpose: To evaluate the short‐term (1 week after completion of treatment) effect of office‐based vergence and anti‐suppression therapy (OBVAT) on the Office Control Score when compared to observation alone in children with small‐to‐moderate angle intermittent exotropia (IXT). Methods: In this single‐masked (examiner masked), two‐arm, single‐centre randomised clinical trial, 40 participants, 6 to <18 years of age with untreated IXT, were randomly assigned to OBVAT or observation alone. Participants assigned to therapy received 60 min of OBVAT with home reinforcement once per week for 16 weeks. Therapy included vergence, accommodation and anti‐suppression techniques. The primary outcome measure was the comparison of the distance Office Control Score between the two groups at the primary outcome visit (i.e., 17‐week follow‐up visit). Results: At the primary outcome visit, the OBVAT group (n = 20) had a significantly better distance Office Control Score (adjusted mean difference: −0.9; 95% CI: −0.2 to −1.5; p = 0.008; partial eta squared: 0.19) than the observation group (n = 16). Participants from the OBVAT group were more likely than those from the observation group to have ≥1 point of improvement at the 17‐week visit (OBVAT group: 75%; Observation group: 25%; p = 0.006). Conclusions: In this randomised clinical trial of participants aged 6 to <18 years with IXT, we found that the OBVAT group had a significantly better distance Office Control Score than the observation group at the 17‐week visit. This study provides the first data from a randomised clinical trial demonstrating the effectiveness of OBVAT for improving the control of IXT. Eye care practitioners should consider OBVAT as a viable, non‐surgical treatment option for IXT. A full‐scale randomised clinical trial investigating the long‐term effectiveness of OBVAT in treating IXT is warranted. [ABSTRACT FROM AUTHOR]

Published
2024
Full Text
View/download PDF

24. Effective use of mesenchymal stem cells in human skin substitutes generated by tissue engineering

Author

MA Martin-Piedra, CA Alfonso-Rodriguez, A Zapater, D Durand-Herrera, J Chato-Astrain, F Campos, MC Sanchez-Quevedo, M Alamino, and I Garzon

Subjects
Mesenchymal stem cells, bioartificial skin, tissue engineering, regenerative medicine., Diseases of the musculoskeletal system, RC925-935, Orthopedic surgery, RD701-811
Abstract

Mesenchymal stem cells (MSCs) can differentiate toward epithelial cells and may be used as an alternative source for generation of heterotypical artificial human skin substitutes, thus, enhancing their development and translation potential to the clinic. The present study aimed at comparing four types of heterotypical human bioengineered skin generated using MSCs as an alternative epithelial cell source. Adipose-tissue-derived stem cells (ADSCs), dental pulp stem cells (DPSCs), Wharton’s jelly stem cells (WJSCs) and bone marrow stem cells (BMSCs) were used for epidermal regeneration on top of dermal skin substitutes. Heterotypic human skin substitutes were evaluated before and after implantation in immune-deficient athymic mice for 30 d. Histological and genetic studies were performed to evaluate extracellular matrix synthesis, epidermal differentiation and human leukocyte antigen (HLA) molecule expression. The four cell types differentiated into keratinocytes, as shown by the expression of cytokeratin 10 and filaggrin 30 d post-grafting; also, they induced dermal fibroblasts responsible for the synthesis of extracellular fibrillar and non-fibrillar components, in a similar way among each other. WJSCs and BMSCs showed higher expression of cytokeratin 10 and filaggrin, suggesting these cells were more prone to epidermal regeneration. The absence of HLA molecules, even when the epithelial layer was differentiated, supports the future clinical use of these substitutes – especially ADSCs, DPSCs and WJSCs – with low rejection risk. MSCs allowed the generation of bioengineered human skin substitutes with potential clinical usefulness. According to their epidermal differentiation potential and lack of HLA antigens, WJSCs should preferentially be used.

Published
2019
Full Text
View/download PDF

33. Vitamin C improves gut Bifidobacteria in humans

Author

Sabine Hazan, Sonya Dave, Andreas J Papoutsis, Nirupama Deshpande, Mark C Howell, and Leisha MA Martin

Subjects
Microbiology (medical), Microbiology
Abstract

Aims: Numerous beneficial effects of vitamin C (ascorbic acid) supplementation have been reported in the literature. However, data on its effects toward the gut microbiome are limited. We assessed the effect of vitamin C supplementation on the abundance of beneficial bacterial species in the gut microbiome. Materials and methods: Stool samples were analyzed for relative abundance of gut microbiome bacteria using next-generation sequencing-based profiling and metagenomic shotgun analysis. Results: Supplementation with vitamin C increased the abundance of bacteria of the genus Bifidobacterium (p = 0.0001) and affected various species. Conclusion: The beneficial effects of vitamin C supplementation may be attributed to modulation of the gut microbiome and the consequent health benefits thereof.

Published
2022

37. Mask R-CNN assisted 2.5D object detection pipeline of 68Ga-PSMA-11 PET/CT-positive metastatic pelvic lymph node after radical prostatectomy from solely CT imaging.

Author

Xu, Di, Ma, Martin, Cao, Minsong, Kishan, Amar U., Nickols, Nicholas G., Scalzo, Fabien, and Sheng, Ke

Subjects
*OBJECT recognition (Computer vision), *COMPUTED tomography, *RADICAL prostatectomy, *POSITRON emission tomography, *LYMPH nodes, *PROSTATE-specific antigen
Abstract

Prostate-specific membrane antigen (PSMA) positron emission tomography (PET)/computed tomography (CT) is a molecular and functional imaging modality with better restaging accuracy over conventional imaging for detecting prostate cancer in men suspected of lymph node (LN) progression after definitive therapy. However, the availability of PSMA PET/CT is limited in both low-resource settings and for repeating imaging surveillance. In contrast, CT is widely available, cost-effective, and routinely performed as part of patient follow-up or radiotherapy workflow. Compared with the molecular activities, the morphological and texture changes of subclinical LNs in CT are subtle, making manual detection of positive LNs infeasible. Instead, we harness the power of artificial intelligence for automated LN detection on CT. We examined 68Ga-PSMA-11 PET/CT images from 88 patients (including 739 PSMA PET/CT-positive pelvic LNs) who experienced a biochemical recurrence after radical prostatectomy and presented for salvage radiotherapy with prostate-specific antigen < 1 ng/mL. Scans were divided into a training set (nPatient = 52, nNode = 400), a validation set (nPatient = 18, nNode = 143), and a test set (nPatient = 18, nNodes = 196). Using PSMA PET/CT as the ground truth and consensus pelvic LN clinical target volumes as search regions, a 2.5-dimensional (2.5D) Mask R-CNN based object detection framework was trained. The entire framework contained whole slice imaging pretraining, masked-out region fine-tuning, prediction post-processing, and "window bagging". Following an additional preprocessing step—pelvic LN clinical target volume extraction, our pipeline located positive pelvic LNs solely based on CT scans. Our pipeline could achieve a sensitivity of 83.351%, specificity of 58.621% out of 196 positive pelvic LNs from 18 patients in the test set, of which most of the false positives can be post-removable by radiologists. Our tool may aid CT-based detection of pelvic LN metastasis and triage patients most unlikely to benefit from the PSMA PET/CT scan. [ABSTRACT FROM AUTHOR]

Published
2023
Full Text
View/download PDF

40. COVID-19 crisis: epidemiological and social perspective of an unprecedented pandemic in our lives

Author

José Mª Martin-Moreno

Subjects
Economic growth, medicine.medical_specialty, Coronavirus disease 2019 (COVID-19), Social perspective, Political science, Pandemic, Epidemiology, medicine, General Medicine
Abstract

Increasing globalization drives complex human, biological and commercial interactions that affect the way we live and get sick. The pandemic risk that has appeared with the COVID-19 had already been anticipated by many institutions and experts, yet our governments and decision-makers have not prepared themselves sufficiently. This has meant that the devastating epidemic we have suffered has caught our entire system off guard. The SARS in 2003 and the MERS in 2012 were previous warnings of coronavirus mutations that presented a worrying infectivity and lethality, although they managed to be controlled with public health measures. On this occasion, the arrogance of many of our countries has meant that although the pandemic was coming progressively and announced, we did not do enough to prepare and strengthen the health system and public health. In this paper, we basically review the etiological agent, the incubation period, the mechanisms of transmission, the risk factors, the summarized clinical manifestations, the epidemiological analysis of the situation (including incidence or new cases, mortality and lethality, and estimating the excess mortality associated with the COVID-19), and the confinement and gradual stages of de-escalation. From the social point of view, the vulnerable groups that have been particularly affected by COVID-19 in Spain are identified. The article concludes by reviewing preventive medicine and public health measures in the face of COVID-19. This is articulated through the measures of basic hygiene and to investigate what we can expect from primary prevention through the mechanism of immunization that can provide us with vaccines. In the final part, the problem is discussed in order to develop, select, prioritize, and distribute these vaccines as a universal good, and it is emphasized that for now, and while we lack vaccines, our obligation is to continue insisting on hygiene and… on healthy behaviors. And to emphasize the importance of education for a better world.

Published
2020

44. PD08-09 META-ANALYSIS OF RADIATION THERAPY AND ANDROGEN DEPRIVATION THERAPY IN LOCALIZED PROSTATE CANCER

Author

Jiang, Tommy, primary, Markovic, Daniela, additional, Patel, Jay, additional, Juarez, Jesus, additional, Ma, Martin, additional, Shabsovich, David, additional, Zaorsky, Nicholas, additional, Nickols, Nicholas, additional, Reiter, Robert, additional, Elashoff, David, additional, Rettig, Matthew, additional, Spratt, Daniel, additional, and Kishan, Amar, additional

Published
2021
Full Text
View/download PDF

45. Analytical Pyrolysis of the Fungal Melanins from Ochroconis spp. Isolated from Lascaux Cave, France

Author

Pedro Mª Martin-Sanchez, Cesáreo Sáiz-Jiménez, José Antonio González-Pérez, Bernardo Hermosín, Sáiz-Jiménez, Cesáreo, Martin-Sanchez, Pedro Mª, González-Pérez, José Antonio, Hermosín, Bernardo, Sáiz-Jiménez, Cesáreo [0000-0003-0036-670X], Martin-Sanchez, Pedro Mª [0000-0003-2674-8650], González-Pérez, José Antonio [0000-0001-7607-1444], and Hermosín, Bernardo [0000-0002-2147-6636]

Subjects
0301 basic medicine, analytical pyrolysis, DOPA melanin, Chemical structure, 010501 environmental sciences, Polysaccharide, lcsh:Technology, fungal melanins, 01 natural sciences, lcsh:Chemistry, 03 medical and health sciences, chemistry.chemical_compound, Eurotium echinulatum, Chitin, Organic chemistry, General Materials Science, Homogentisic acid, lcsh:QH301-705.5, Instrumentation, 0105 earth and related environmental sciences, Fluid Flow and Transfer Processes, chemistry.chemical_classification, biology, lcsh:T, Process Chemistry and Technology, General Engineering, Lascaux Cave, Ochroconis lascauxensis, biology.organism_classification, lcsh:QC1-999, Computer Science Applications, pyomelanin, 030104 developmental biology, lcsh:Biology (General), lcsh:QD1-999, chemistry, Polymerization, lcsh:TA1-2040, Cryptococcus neoformans, DHN melanin, lcsh:Engineering (General). Civil engineering (General), Pyrolysis, lcsh:Physics, Macromolecule
Abstract

Analytical pyrolysis is a tool widely used for investigating the nature of macromolecules. This technique has been applied with success for solving the chemical structure of geo- and biopolymers. However, with complex macromolecules, such as fungal melanins, analytical pyrolysis presents some drawbacks due to the fact that they rarely can be obtained in a pure grade and contains a mixture of polysaccharides, chitin, proteins, nucleic acids, and lipids, among other materials. Analytical pyrolysis permits the characterization of the compounds co-extracted or linked to the melanin. Nevertheless, our data revealed that under the conditions usually employed in the pyrolysis of melanins, the technique presented serious limitations due to the complex structure and diversity of components that suffer extensive thermal degradation through secondary reactions with the production of considerable number of artifacts. To the light of pyrolysis data, the melanin of Ochroconis spp. is not based on 3,4-dihydroxyphenylalanine (DOPA) or 1,8-dihydroxynaphthalene (DHN), and it is suggested that it could be a type of pyomelanin, based on the polymerization of homogentisic acid and other phenolic compounds.

Published
2021

50. Efficiency of U.S. dialysis centers: an updated examination of facility characteristics that influence production of dialysis treatments

Author

Shreay, Sanatan, Ma, Martin, McCluskey, Jill, Mittelhammer, Ron C., Gitlin, Matthew, and Stephens, J. Mark

Subjects
Hemodialysis facilities -- Analysis, Health care industry -- Analysis -- Services, Hemodialysis -- Analysis, Health care industry, Business
Abstract

Objective. To explore the relative efficiency of dialysis facilities in the United States and identify factors that are associated with efficiency in the production of dialysis treatments. Data Sources/Study Setting. Medicare cost report data from 4,343 free-standing dialysis facilities in the United States that offered in-center hemodialysis in 2010. Study Design. A cross-sectional, facility-level retrospective database analysis, utilizing data envelopment analysis (DEA) to estimate facility efficiency. Data Collection/Extraction Methods. Treatment data and cost and labor inputs of dialysis treatments were obtained from 2010 Medicare Renal Cost Reports. Demographic data were obtained from the 2010 U.S. Census. Principal Findings. Only 26.6 percent of facilities were technically efficient. Neither the intensity of market competition nor the profit status of the facility had a significant effect on efficiency. Facilities that were members of large chains were less likely to be efficient. Cost and labor savings due to changes in drug protocols had little effect on overall dialysis center efficiency. Conclusions. The majority of free-standing dialysis facilities in the United States were functioning in a technically inefficient manner. As payment systems increasingly employ capitation and bundling provisions, these institutions will need to evaluate their efficiency to remain competitive. Key Words. Dialysis market, efficiency, data envelopment analysis, More than ever, health care providers in the United States are facing the challenge of meeting a growing demand for health care services in an environment of increasing pressures on [...]

Published
2014
Full Text
View/download PDF

Catalog

Books, media, physical & digital resources
See catalog results