Building Flexible, Scalable, and Machine Learning-Ready Multimodal Oncology Datasets

The advancements in data acquisition, storage, and processing techniques have resulted in the rapid growth of heterogeneous medical data. Integrating radiological scans, histopathology images, and molecular information with clinical data is essential for developing a holistic understanding of the disease and optimizing treatment. The need for integrating data from multiple sources is further pronounced in complex diseases such as cancer for enabling precision medicine and personalized treatments. This work proposes Multimodal Integration of Oncology Data System (MINDS)-a flexible, scalable, and cost-effective metadata framework for efficiently fusing disparate data from public sources such as the Cancer Research Data Commons (CRDC) into an interconnected, patient-centric framework. MINDS consolidates over 41,000 cases from across repositories while achieving a high compression ratio relative to the 3.78 PB source data size. It offers sub-5-s query response times for interactive exploration. MINDS offers an interface for exploring relationships across data types and building cohorts for developing large-scale multimodal machine learning models. By harmonizing multimodal data, MINDS aims to potentially empower researchers with greater analytical ability to uncover diagnostic and prognostic insights and enable evidence-based personalized care. MINDS tracks granular end-to-end data provenance, ensuring reproducibility and transparency. The cloud-native architecture of MINDS can handle exponential data growth in a secure, cost-optimized manner while ensuring substantial storage optimization, replication avoidance, and dynamic access capabilities. Auto-scaling, access controls, and other mechanisms guarantee pipelines' scalability and security. MINDS overcomes the limitations of existing biomedical data silos via an interoperable metadata-driven approach that represents a pivotal step toward the future of oncology data integration.

Revolutionizing Digital Pathology With the Power of Generative Artificial Intelligence and Foundation Models

Digital pathology has transformed the traditional pathology practice of analyzing tissue under a microscope into a computer vision workflow. Whole-slide imaging allows pathologists to view and analyze microscopic images on a computer monitor, enabling computational pathology. By leveraging artificial intelligence (AI) and machine learning (ML), computational pathology has emerged as a promising field in recent years. Recently, task-specific AI/ML (eg, convolutional neural networks) has risen to the forefront, achieving above-human performance in many image-processing and computer vision tasks. The performance of task-specific AI/ML models depends on the availability of many annotated training datasets, which presents a rate-limiting factor for AI/ML development in pathology. Task-specific AI/ML models cannot benefit from multimodal data and lack generalization, eg, the AI models often struggle to generalize to new datasets or unseen variations in image acquisition, staining techniques, or tissue types. The 2020s are witnessing the rise of foundation models and generative AI. A foundation model is a large AI model trained using sizable data, which is later adapted (or fine-tuned) to perform different tasks using a modest amount of task-specific annotated data. These AI models provide in-context learning, can self-correct mistakes, and promptly adjust to user feedback. In this review, we provide a brief overview of recent advances in computational pathology enabled by task-specific AI, their challenges and limitations, and then introduce various foundation models. We propose to create a pathology-specific generative AI based on multimodal foundation models and present its potentially transformative role in digital pathology. We describe different use cases, delineating how it could serve as an expert companion of pathologists and help them efficiently and objectively perform routine laboratory tasks, including quantifying image analysis, generating pathology reports, diagnosis, and prognosis. We also outline the potential role that foundation models and generative AI can play in standardizing the pathology laboratory workflow, education, and training.

836 Presentation, Clinical Outcomes and Prognosis of Primary Appendiceal Adenocarcinoma; Single Centre Experience

Background

Appendiceal adenocarcinoma is a rare condition. The aim of the study is to determine the pattern of clinical picture, histological diagnosis and prognosis of the disease on patients in our area.

Method

A retrospective data analysis of patients who underwent appendicectomies from January 2011 to December 2019 in Northumbria NHS Trust.

Records of patients with histologically diagnosed appendiceal adenocarcinoma were studied for their demographics, clinical features, laboratory, radiological results, operative notes and histological types.

Results

13 out of 3138 appendicectomies (0.41%: 6 Males, 7 Females with mean age of 69.84 years) were found to have primary appendiceal adenocarcinoma. Almost all the patients presented with clinical picture in keeping of acute appendicitis.

9 patients underwent emergency appendicectomies. 4 patients were treated conservatively.

Appendiceal adenocarcinoma was diagnosed based on colonoscopy in 1 patient, CT scan in 1, post-appendicectomy histology in 9 patients, 2 post RHC histologies.

Survival rate was 71.4% after 1 year, 64.3% after 2 years and 42.9% after 3 years.

Overall median survival time for all TNM stages was 39.0 (95% CI, 15.72 to 62.27) months. Mean survival time was best in stage II 56.33 (95% CI, 48.86 to 63.80) months.

Conclusions

Preoperative diagnosis is very difficult. Overall prognosis is poor.

One treatment fits all: Effectiveness of a multicomponent cognitive behavioral therapy program in data-driven subtypes of perinatal depression

Introduction

It has been well established that depressive disorders including perinatal depression are very heterogeneous, which partly explain the ineffectiveness of available treatments for many patients. Recent innovations in data science can help elucidate the nature of perinatal depression especially the heterogeneity in its presentation.

Objectives

The present study aime to elucidate heterogeneous subtypes of PND and assess the effectiveness of a multicomponent cognitive behavioral therapy (CBT) across heterogenous subtypes of PND.

Methods

This study was conducted in 2005 in two rural areas of Rawalpindi, Pakistan. Out of a total of 3,898 women, 903 pregnant women were identifed with PND (using DSM-IV) and randomly assigned to intervention and control group. Baseline assessments included interviewer admininstered Hamilton Depression Scale (HDS) and social risk factors. Follow-up assessments were conducted at 6 months and 12 months post-intervention. Principle component analysis was run to reduce dimensionality of the HDS. Two step cluster analysis was then run to elucidate subtypes of PND using the dimensional scores. Thereafter, effectiveness of CBT was compared across these subtypes of PND using multilevel modelling.

Results

Principle component analysis revealed a four component solution for the Hamilton depression rating scale. Using these dimensional scores, cluster analysis (average silhouette= 0.5) revealed a parsimonius four cluster soultion of participants with mild PND symptoms (n=326); predominant sleep problems (n=311) c) predominant atypical symptoms (n=80) and d) comorbid depressive and anxiety symptoms (n=186). CBT yielded moderate effect sizes across all these subtypes of PND (cohen’s d > 0.8).

Conclusions

Multicomponent CBT is effective across hetergeneous presentations of PND.

Disclosure

No significant relationships.

Full genome sequence of a peste des petits ruminants virus (PPRV) from Ghana

The full genome of a peste des petits ruminants virus (PPRV) isolated from a sheep lung sample collected in Ghana, Western Africa, in 2010, has been sequenced. Phylogenetic analysis demonstrated that the virus clustered within the lineage II clade while comparison of its full genome with those of other PPRV strains revealed the highest identity (96.6 %) at a nucleotide level with the PPRV strain Nigeria/76/1. This is the first full genome sequence generated for a PPRV lineage II isolated since 1976.