A whole-slide foundation model for digital pathology from real-world data.

Digital pathology poses unique computational challenges, as a standard gigapixel slide may comprise tens of thousands of image tiles1-3. Prior models have often resorted to subsampling a small portion of tiles for each slide, thus missing the important slide-level context4. Here we present Prov-GigaPath, a whole-slide pathology foundation model pretrained on 1.3 billion 256 × 256 pathology image tiles in 171,189 whole slides from Providence, a large US health network comprising 28 cancer centres. The slides originated from more than 30,000 patients covering 31 major tissue types. To pretrain Prov-GigaPath, we propose GigaPath, a novel vision transformer architecture for pretraining gigapixel pathology slides. To scale GigaPath for slide-level learning with tens of thousands of image tiles, GigaPath adapts the newly developed LongNet5 method to digital pathology. To evaluate Prov-GigaPath, we construct a digital pathology benchmark comprising 9 cancer subtyping tasks and 17 pathomics tasks, using both Providence and TCGA data6. With large-scale pretraining and ultra-large-context modelling, Prov-GigaPath attains state-of-the-art performance on 25 out of 26 tasks, with significant improvement over the second-best method on 18 tasks. We further demonstrate the potential of Prov-GigaPath on vision-language pretraining for pathology7,8 by incorporating the pathology reports. In sum, Prov-GigaPath is an open-weight foundation model that achieves state-of-the-art performance on various digital pathology tasks, demonstrating the importance of real-world data and whole-slide modelling.

Datathons and Software to Promote Reproducible Research.

BACKGROUND: Datathons facilitate collaboration between clinicians, statisticians, and data scientists in order to answer important clinical questions. Previous datathons have resulted in numerous publications of interest to the critical care community and serve as a viable model for interdisciplinary collaboration. OBJECTIVE: We report on an open-source software called Chatto that was created by members of our group, in the context of the second international Critical Care Datathon, held in September 2015. METHODS: Datathon participants formed teams to discuss potential research questions and the methods required to address them. They were provided with the Chatto suite of tools to facilitate their teamwork. Each multidisciplinary team spent the next 2 days with clinicians working alongside data scientists to write code, extract and analyze data, and reformulate their queries in real time as needed. All projects were then presented on the last day of the datathon to a panel of judges that consisted of clinicians and scientists. RESULTS: Use of Chatto was particularly effective in the datathon setting, enabling teams to reduce the time spent configuring their research environments to just a few minutes-a process that would normally take hours to days. Chatto continued to serve as a useful research tool after the conclusion of the datathon. CONCLUSIONS: This suite of tools fulfills two purposes: (1) facilitation of interdisciplinary teamwork through archiving and version control of datasets, analytical code, and team discussions, and (2) advancement of research reproducibility by functioning postpublication as an online environment in which independent investigators can rerun or modify analyses with relative ease. With the introduction of Chatto, we hope to solve a variety of challenges presented by collaborative data mining projects while improving research reproducibility.

Machine learning for healthcare that matters: Reorienting from technical novelty to equitable impact.

Despite significant technical advances in machine learning (ML) over the past several years, the tangible impact of this technology in healthcare has been limited. This is due not only to the particular complexities of healthcare, but also due to structural issues in the machine learning for healthcare (MLHC) community which broadly reward technical novelty over tangible, equitable impact. We structure our work as a healthcare-focused echo of the 2012 paper "Machine Learning that Matters", which highlighted such structural issues in the ML community at large, and offered a series of clearly defined "Impact Challenges" to which the field should orient itself. Drawing on the expertise of a diverse and international group of authors, we engage in a narrative review and examine issues in the research background environment, training processes, evaluation metrics, and deployment protocols which act to limit the real-world applicability of MLHC. Broadly, we seek to distinguish between machine learning ON healthcare data and machine learning FOR healthcare-the former of which sees healthcare as merely a source of interesting technical challenges, and the latter of which regards ML as a tool in service of meeting tangible clinical needs. We offer specific recommendations for a series of stakeholders in the field, from ML researchers and clinicians, to the institutions in which they work, and the governments which regulate their data access.

Fine-tuning large neural language models for biomedical natural language processing.

Large neural language models have transformed modern natural language processing (NLP) applications. However, fine-tuning such models for specific tasks remains challenging as model size increases, especially with small labeled datasets, which are common in biomedical NLP. We conduct a systematic study on fine-tuning stability in biomedical NLP. We show that fine-tuning performance may be sensitive to pretraining settings and conduct an exploration of techniques for addressing fine-tuning instability. We show that these techniques can substantially improve fine-tuning performance for low-resource biomedical NLP applications. Specifically, freezing lower layers is helpful for standard BERT- B A S E models, while layerwise decay is more effective for BERT- L A R G E and ELECTRA models. For low-resource text similarity tasks, such as BIOSSES, reinitializing the top layers is the optimal strategy. Overall, domain-specific vocabulary and pretraining facilitate robust models for fine-tuning. Based on these findings, we establish a new state of the art on a wide range of biomedical NLP applications.

What are the Desired Characteristics of Calibration Sets? Identifying Correlates on Long Form Scientific Summarization.

Summarization models often generate text that is poorly calibrated to quality metrics because they are trained to maximize the likelihood of a single reference (MLE). To address this, recent work has added a calibration step, which exposes a model to its own ranked outputs to improve relevance or, in a separate line of work, contrasts positive and negative sets to improve faithfulness. While effective, much of this work has focused on how to generate and optimize these sets. Less is known about why one setup is more effective than another. In this work, we uncover the underlying characteristics of effective sets. For each training instance, we form a large, diverse pool of candidates and systematically vary the subsets used for calibration fine-tuning. Each selection strategy targets distinct aspects of the sets, such as lexical diversity or the size of the gap between positive and negatives. On three diverse scientific long-form summarization datasets (spanning biomedical, clinical, and chemical domains), we find, among others, that faithfulness calibration is optimal when the negative sets are extractive and more likely to be generated, whereas for relevance calibration, the metric margin between candidates should be maximized and surprise-the disagreement between model and metric defined candidate rankings-minimized. Code to create, select, and optimize calibration sets is available at https://github.com/griff4692/calibrating-summaries.

Toward structuring real-world data: Deep learning for extracting oncology information from clinical text with patient-level supervision.

Most detailed patient information in real-world data (RWD) is only consistently available in free-text clinical documents. Manual curation is expensive and time consuming. Developing natural language processing (NLP) methods for structuring RWD is thus essential for scaling real-world evidence generation. We propose leveraging patient-level supervision from medical registries, which are often readily available and capture key patient information, for general RWD applications. We conduct an extensive study on 135,107 patients from the cancer registry of a large integrated delivery network (IDN) comprising healthcare systems in five western US states. Our deep-learning methods attain test area under the receiver operating characteristic curve (AUROC) values of 94%-99% for key tumor attributes and comparable performance on held-out data from separate health systems and states. Ablation results demonstrate the superiority of these advanced deep-learning methods. Error analysis shows that our NLP system sometimes even corrects errors in registrar labels.

The role of machine learning in clinical research: transforming the future of evidence generation.

BACKGROUND: Interest in the application of machine learning (ML) to the design, conduct, and analysis of clinical trials has grown, but the evidence base for such applications has not been surveyed. This manuscript reviews the proceedings of a multi-stakeholder conference to discuss the current and future state of ML for clinical research. Key areas of clinical trial methodology in which ML holds particular promise and priority areas for further investigation are presented alongside a narrative review of evidence supporting the use of ML across the clinical trial spectrum. RESULTS: Conference attendees included stakeholders, such as biomedical and ML researchers, representatives from the US Food and Drug Administration (FDA), artificial intelligence technology and data analytics companies, non-profit organizations, patient advocacy groups, and pharmaceutical companies. ML contributions to clinical research were highlighted in the pre-trial phase, cohort selection and participant management, and data collection and analysis. A particular focus was paid to the operational and philosophical barriers to ML in clinical research. Peer-reviewed evidence was noted to be lacking in several areas. CONCLUSIONS: ML holds great promise for improving the efficiency and quality of clinical research, but substantial barriers remain, the surmounting of which will require addressing significant gaps in evidence.

A Review of Challenges and Opportunities in Machine Learning for Health.

Modern electronic health records (EHRs) provide data to answer clinically meaningful questions. The growing data in EHRs makes healthcare ripe for the use of machine learning. However, learning in a clinical setting presents unique challenges that complicate the use of common machine learning methodologies. For example, diseases in EHRs are poorly labeled, conditions can encompass multiple underlying endotypes, and healthy individuals are underrepresented. This article serves as a primer to illuminate these challenges and highlights opportunities for members of the machine learning community to contribute to healthcare.

Natural Language Processing for EHR-Based Computational Phenotyping.

This article reviews recent advances in applying natural language processing (NLP) to Electronic Health Records (EHRs) for computational phenotyping. NLP-based computational phenotyping has numerous applications including diagnosis categorization, novel phenotype discovery, clinical trial screening, pharmacogenomics, drug-drug interaction (DDI), and adverse drug event (ADE) detection, as well as genome-wide and phenome-wide association studies. Significant progress has been made in algorithm development and resource construction for computational phenotyping. Among the surveyed methods, well-designed keyword search and rule-based systems often achieve good performance. However, the construction of keyword and rule lists requires significant manual effort, which is difficult to scale. Supervised machine learning models have been favored because they are capable of acquiring both classification patterns and structures from data. Recently, deep learning and unsupervised learning have received growing attention, with the former favored for its performance and the latter for its ability to find novel phenotypes. Integrating heterogeneous data sources have become increasingly important and have shown promise in improving model performance. Often, better performance is achieved by combining multiple modalities of information. Despite these many advances, challenges and opportunities remain for NLP-based computational phenotyping, including better model interpretability and generalizability, and proper characterization of feature relations in clinical narratives.

Trends and Focus of Machine Learning Applications for Health Research.

Importance: The use of machine learning applications related to health is rapidly increasing and may have the potential to profoundly affect the field of health care. Objective: To analyze submissions to a popular machine learning for health venue to assess the current state of research, including areas of methodologic and clinical focus, limitations, and underexplored areas. Design, Setting, and Participants: In this data-driven qualitative analysis, 166 accepted manuscript submissions to the Third Annual Machine Learning for Health workshop at the 32nd Conference on Neural Information Processing Systems on December 8, 2018, were analyzed to understand research focus, progress, and trends. Experts reviewed each submission against a rubric to identify key data points, statistical modeling and analysis of submitting authors was performed, and research topics were quantitatively modeled. Finally, an iterative discussion of topics common in submissions and invited speakers at the workshop was held to identify key trends. Main Outcomes and Measures: Frequency and statistical measures of methods, topics, goals, and author attributes were derived from an expert review of submissions guided by a rubric. Results: Of the 166 accepted submissions, 58 (34.9%) had clinician involvement and 83 submissions (50.0%) that focused on clinical practice included clinical collaborators. A total of 97 data sets (58.4%) used in submissions were publicly available or required a standard registration process. Clinical practice was the most common application area (70 manuscripts [42.2%]), with brain and mental health (25 [15.1%]), oncology (21 [12.7%]), and cardiovascular (19 [11.4%]) being the most common specialties. Conclusions and Relevance: Trends in machine learning for health research indicate the importance of well-annotated, easily accessed data and the benefit from greater clinician involvement in the development of translational applications.

What's in a Note? Unpacking Predictive Value in Clinical Note Representations.

Electronic Health Records (EHRs) have seen a rapid increase in adoption during the last decade. The narrative prose contained in clinical notes is unstructured and unlocking its full potential has proved challenging. Many studies incorporating clinical notes have applied simple information extraction models to build representations that enhance a downstream clinical prediction task, such as mortality or readmission. Improved predictive performance suggests a "good" representation. However, these extrinsic evaluations are blind to most of the insight contained in the notes. In order to better understand the power of expressive clinical prose, we investigate both intrinsic and extrinsic methods for understanding several common note representations. To ensure replicability and to support the clinical modeling community, we run all experiments on publicly-available data and provide our code.

A "datathon" model to support cross-disciplinary collaboration.

In recent years, there has been a growing focus on the unreliability of published biomedical and clinical research. To introduce effective new scientific contributors to the culture of health care, we propose a "datathon" or "hackathon" model in which participants with disparate, but potentially synergistic and complementary, knowledge and skills effectively combine to address questions faced by clinicians. The continuous peer review intrinsically provided by follow-up datathons, which take up prior uncompleted projects, might produce more reliable research, either by providing a different perspective on the study design and methodology or by replication of prior analyses.

A Multivariate Timeseries Modeling Approach to Severity of Illness Assessment and Forecasting in ICU with Sparse, Heterogeneous Clinical Data.

The ability to determine patient acuity (or severity of illness) has immediate practical use for clinicians. We evaluate the use of multivariate timeseries modeling with the multi-task Gaussian process (GP) models using noisy, incomplete, sparse, heterogeneous and unevenly-sampled clinical data, including both physiological signals and clinical notes. The learned multi-task GP (MTGP) hyperparameters are then used to assess and forecast patient acuity. Experiments were conducted with two real clinical data sets acquired from ICU patients: firstly, estimating cerebrovascular pressure reactivity, an important indicator of secondary damage for traumatic brain injury patients, by learning the interactions between intracranial pressure and mean arterial blood pressure signals, and secondly, mortality prediction using clinical progress notes. In both cases, MTGPs provided improved results: an MTGP model provided better results than single-task GP models for signal interpolation and forecasting (0.91 vs 0.69 RMSE), and the use of MTGP hyperparameters obtained improved results when used as additional classification features (0.812 vs 0.788 AUC).

Unfolding Physiological State: Mortality Modelling in Intensive Care Units.

Accurate knowledge of a patient's disease state and trajectory is critical in a clinical setting. Modern electronic healthcare records contain an increasingly large amount of data, and the ability to automatically identify the factors that influence patient outcomes stand to greatly improve the efficiency and quality of care. We examined the use of latent variable models (viz. Latent Dirichlet Allocation) to decompose free-text hospital notes into meaningful features, and the predictive power of these features for patient mortality. We considered three prediction regimes: (1) baseline prediction, (2) dynamic (time-varying) outcome prediction, and (3) retrospective outcome prediction. In each, our prediction task differs from the familiar time-varying situation whereby data accumulates; since fewer patients have long ICU stays, as we move forward in time fewer patients are available and the prediction task becomes increasingly difficult. We found that latent topic-derived features were effective in determining patient mortality under three timelines: inhospital, 30 day post-discharge, and 1 year post-discharge mortality. Our results demonstrated that the latent topic features important in predicting hospital mortality are very different from those that are important in post-discharge mortality. In general, latent topic features were more predictive than structured features, and a combination of the two performed best. The time-varying models that combined latent topic features and baseline features had AUCs that reached 0.85, 0.80, and 0.77 for in-hospital, 30 day post-discharge and 1 year post-discharge mortality respectively. Our results agreed with other work suggesting that the first 24 hours of patient information are often the most predictive of hospital mortality. Retrospective models that used a combination of latent topic features and structured features achieved AUCs of 0.96, 0.82, and 0.81 for in-hospital, 30 day, and 1-year mortality prediction. Our work focuses on the dynamic (time-varying) setting because models from this regime could facilitate an on-going severity stratification system that helps direct care-staff resources and inform treatment strategies.

Making big data useful for health care: a summary of the inaugural mit critical data conference.

With growing concerns that big data will only augment the problem of unreliable research, the Laboratory of Computational Physiology at the Massachusetts Institute of Technology organized the Critical Data Conference in January 2014. Thought leaders from academia, government, and industry across disciplines-including clinical medicine, computer science, public health, informatics, biomedical research, health technology, statistics, and epidemiology-gathered and discussed the pitfalls and challenges of big data in health care. The key message from the conference is that the value of large amounts of data hinges on the ability of researchers to share data, methodologies, and findings in an open setting. If empirical value is to be from the analysis of retrospective data, groups must continuously work together on similar problems to create more effective peer review. This will lead to improvement in methodology and quality, with each iteration of analysis resulting in more reliability.