FDA 510K Clearance For Digital Pathology: What is it? What’s the process?
What is the FDA 510K clearance for a Software as a Medical Device (SaMD) for Digital Pathology? How does this compare to the IVDR/CE process in Europe?
Posts organized by Categories: Oncology, Technology and Data/AI/Machine-Learning
What is the FDA 510K clearance for a Software as a Medical Device (SaMD) for Digital Pathology? How does this compare to the IVDR/CE process in Europe?
Enlightening discussion on the application of AI in three categories for drug development: Biomarker Target Modulation, Molecule generation and Clinical deve...
Focused particularly on NSCLC and Melanoma, a discussion on the huge shift towards biomarker testing for targeted therapies and immunotherapies.
A retrospective over 5 years and prognosis for digital pathology.
Using a tumor derived immunotherapy composed of a patient’s own disease-fighting white blood cells known as T-cells - specific type called tumor-infiltrating...
Dr. Carl June’s pioneering work in CAR-T cell therapy which builds up the human immune system to recognize cells gone rogue, i.e., cancer!
Fascinating story of Dr. Daphne Koller on her journey from Stanford to founding Coursera to Insitro. Focus on early Biology and discovery using AI for the ca...
Amanda Coble (Director of Product Management at Proscia) shares her amazing journey from pathologist assistant > LIS admin > building LIS products >...
Brilliant insights on AI in healthcare from Ajit Singh, PhD, Managing Director at Artiman Ventures, who shares learnings from his personal journey starting a...
Dr Levenson talks about his work in slide free imaging and it’s potential to significantly decrease turnaround time, while providing more information than th...
Excellent overview of the vast landscape of digital pathology: Some history, benefits, Image Analysis and AI/ML.
From disease staging to pathologic complete response assessments, pathology guides the way toward precision medicine and targeted therapies.
An excellent podcast on how scientific discoveries are translated into clinical patient practice!
Agile can bring significant value to medical device software development and is supported by the FDA, regulators, and manufacturers alike. The key is to appl...
A discussion on Imaging based diagnostic and therapeutic strategies for Prostate Cancer including the use of AI. RSNA Podcasts
The field of spatial biology is becoming a critical driver of life science discovery. For decades we knew that cancer tissue was heterogeneous, with some cel...
Great discussion on the “obvious” ROI for Digital Pathology based on Dr. Andrew Evans’ talk on the roll out of digipath for MacKenzie Health. Plus, the key r...
A deep dive into the role of multiplex immunofluorescence for spatial biology and the interpretation of high dimensionality multimodal data. Digital Patholo...
Automating the laborious data processes and speed up new drug and healthcare developments for precision medicine.
The results suggest that the zero-shot SAM model achieves remarkable segmentation performance for large connected objects. However, it does not consistently ...
Fascinating discussion about the potential benefits of LLMs like ChatGPT in pathology and AI in automating slide scanning including quality control.
Great perspective from Dr. Welch at Pathgroup that AI is THE business case for digitization of pathology and AI algorithms will be required for a lab to be s...
Great discussion about the role of AI in early detection of Breast Cancer using imaging. And the role of AI with horizontal healthcare data including radiolo...
Dr. Marilyn Bui, is a patient-focused pathologist and emphasizes the patient-centricity of pathology work. Dr. Bui believes that pathology and laboratory med...
Digital pathology is transforming the accessibility of decision rules and decision augmentation, creating translational assets and changing how clinical deve...
Prof. Inti Zlobec (University of Bern) leads digital pathology and bridges the gap between pathologists, computer scientists, and data scientists. Digital P...
Throughout the evolution of the field of computational pathology the focus has shifted from research-oriented work to direct collaboration with clinicians to...
This podcast discusses the fundamentals of machine learning algorithm development from an industry perspective. Topics include choosing a project, defining h...
With advances in immunotherapy, we know that if you have cancer with MSI (Micro Satellite Instability), even if it’s late stage, it has a good chance of resp...
This podcast discusses the regulatory aspects of digital pathology and how it fits into translational medicine: The space between research and clinical use. ...
Article describes the advances in CAR T, a remarkable immunotherapy treatment dubbed a “living drug” and the foundations for understanding how CAR T works. ...
In the past decade, scientists are using CRISPR (Clustered Regularly Interspaced Short Palindromic Repeats) to edit cell genomes in multiple clinical trials ...
The modern biotech tech stack: For biotech companies, software is a means to an end—it’s science, not software, they seek to advance. As the need to analyze ...
The value proposition: Consultation with colleagues instantly, sharing screenshots of the images, put pins, do annotations and share other metadata about the...
A discussion of the immune system (Innate & Adaptive), the role of macrophages, dendritic cells, T-cells, B-cells and immune regulation. Finally, the use...
Is it possible to predict molecular biomarkers from H&E (Hematoxylin and Eosin) slides? Digital Pathology Today
Discussion on the state of Digitization of Pathology from presidents of digital pathology societies from 3 different continents. Includes some great advice f...
Dr. Pai from the Mayo Clinic discusses the supervised deep learning tissue image analysis model he developed to quantify visual features of colon cancer to p...
Amazing Podacat! The panel discusses the the ability of ChatGPT to be reasonably good at a variety of tasks like creating paragraphs for medical writing, can...
The 6 hallmarks for cancer: 1. They’re self-sufficient with growth signals 2. They’re insensitive to anti-growth signals. 3. They evade the immune system. 4....
Dr. Ramamurthy shares his thoughts on digital pathology and artificial intelligence as a tool for clinical trials and discusses the need for more guidance on...
An automatic end-to-end deep learning–based detection tool could detect pancreatic cancer on CT scans in a nationwide (Taiwan) real-world test data set with ...
A discussion of the current challenges with immuno-oncology with limited targeted therapies and increasing those using newer tools! Bio Eats World
Radiology 396, Cardiology 58 and Hematology 14 lead the pack. FDA has now cleared more than 500 healthcare AI algorithms Healthcare Exec
Stumbled upon this podcast to train participants in the peer-review process for a Radiological Society of North America (RSNA) AI Journal. Great concept! RS...
To dramatically reduce costs off-the-shelf CAR-T cells. The off-the shelf path would preempt the very costly ($500,000) , laborious and time consuming work o...
Building an Enterprise Digital Pathology operating system that is AI-computationally-native for Research, Diagnostics & Real-World-Data. StartupPhilly
A very timely and fascinating conversation on DICOM in Pathology supporting Whole-Slide-Images. Lots of details on image formats, metadata, viewers and APIs....
The Chair of Computational Pathology & Artificial Intelligence at the Mayo Clinic (Dr. Hipp) discusses his journey through digital pathology and the role...
How image analysis and artificial intelligence support digital pathology-enabled precision medicine today and what to expect in the future. AI Image Analysis
Dr. Garth Bryson and Dr. David Harrison, discuss ICAIRD (Industrial Center for Artificial Intelligence Research & Digital Diagnostics) - a program matchi...
New AI-based biomarker can help predict immunotherapy response for patients with lung cancer. AI Biomarker Response For Lung Cancer
The electrosurgical knife is connected to a mass spectrometer, an analytical instrument used to identify what chemicals are present in a sample. Different ty...
Dr. Ralf Huss shares his view on why and how digital pathology should be implemented into clinical practice. Digital Pathology Transformation Clinical
The role of artificial intelligence and why pathologists should embrace it in their work.
Great podcast on the role of the DICOM - Digital Imaging and Communications in Medicine standard from Radioloty in Pathology.
Session from American Society for Clinical Pathology (ASCP) 2022 annual meeting discussing the latest on emerging Biomarkers, Immuno-oncology resistance mech...
Enabling institutions to embark on the digital pathology journey in an incremental fashion would change the digital pathology landscape and significantly inc...
Podcast discussing the benefits of digitizing pathology include unlocking value in the gigapixels of data from images to enable precision medicine combining ...
Dr. Rishi Pai discusses the use of complex supervised deep learning tissue image analysis AI/ML model to quntify visual diagnostic features. Digital Patholo...
“These are immune cells that are found in the tumor, trying to kill the tumor, but obviously not doing a good enough job because the tumor is growing,” said ...
Amazing story: A T cell normally kills invaders, like viruses. Chimeric antigen receptor T (or CAR-T) cells had been modified in a lab to attack leukemia cel...
Researchers have developed a deep learning model that uses a single chest X-ray to predict the 10-year risk of death from a heart attack or stroke, stemming ...
New CRISPR cancer treatment tested in humans for first time One dose could potentially provide life-long protection against recurrence.
Great podcast: Pathologist shortage (retirement and lower incoming rate), windshield time for remote on-location work and AI/ML will drive digitization!
Breaking Down silos and Democratizing Access to Diagnostics: When will we have a Department of Diagnostic Medicine like in the TV series Dr. House? Will we a...
Bridging the Gap: The Critical Role of Regulatory Affairs and Clinical Affairs in the Total Product Life Cycle of Pathology Imaging Devices and Software.
Discusses the Joint Pathology Center’s approximately 55 million glass tissue slides from more than a century ago! Amazing initiative that’s helping unlock th...
Discusses history and emergence of IHC, biomarkers and genomics. Plus, the inevitable switch of pathology to digital including AI. Finally, perhaps a future ...
As the number of devices increases, the agency is looking to adapt its regulatory framework to the new technology, including faster approval of algorithm upd...
We may be entering a third wave of imaging AI’s evolution, bringing a shift from narrow point solutions to comprehensive multi-finding AI systems.
Miraculous breast cancer vaccine passes first human trials that went on for 20 years It is high time we killed the tumors.
Discusses the crossroads of Immunoncology, digital pathology, image analysis, artificial intelligence, and the increasing ability to multiplex.
A Digital Pathology Image Management Platform needs to support computationally-native applications using open APIs with scalability and ergonomically centere...
Application of AI in Radiology to predict cancer in the future based on past mammograms. AI In Radiology
This podcast discusses the need for using the Dicom standard for annotations, not just the images! Makes even more sense with the advent of AI/ML. Also, some...
Enlightening podcast by Prof Elfriede Noessner of Visiopharm about Immuno-oncology: Tapping the immune system to fight cancer. The dynamic nature of the tumo...
Brilliant podcast the discusses the benefits of digitizing pathology: Including deriving value from the gigapixels of data in each image that enable precisio...
Great podcast by Digital Pathology Today and Proscia co-founder Nathan Buchbinder discuss the new AMA/CPT (cat 3) codes for Digital Pathology. While this ste...
Loved the practical idea of using the analog-microscope and digital in parallel with the ability to fetch glass slides in < 1 hour for a period of time. I...
David West, Co-founder/CEO of Proscia talks about pathology’s shift from microscope to digital using software in the modern lab, the big data generated &...
Paul Graham’s blog discusses Founder mode versus Manager mode: What it takes to build a successful scalable company!
The already amazing AlphaFold 2 AI system only knew about amino acids, but now, AlphaFold 3 can in principle predict where a drug binds to a protein!
Great perspective and advice for engineers and tech leaders on growing, learning and contributing!
A great episode on significantly accelerating innovation using “Tiger Teams”. An OKRs focus is not sufficient!
An excellent explanation: Cross-origin resource sharing (CORS) is a mechanism for integrating applications. CORS defines a way for client web applications th...
A great overview of use cases that inform the selection of databases including Spatial DBs (like the PostGIS extension to postgres).
An amazing review of use cases that inform the selection of databases: Relational(SQL-like), Key-Value (Redis), Document (Mongo), Time-Series (Prometheus), G...
Podcast on various use cases for Generative AI including translation, image processing and data engineering.
Some real great and timely wisdom on what to do proactively and reactively!
The fascinating personal journey of one of the pioneers of visual AI. From immigrant beginnings to “CEO” of a laundromat to Professor at Stanford. Plus, the ...
Gregg Tavares (author of WebGL/WebGPU Fundamentals) gives us a tour of these low-level technologies that are pushing the web forward into the world of video ...
New idea generation methods need to look at innovation through the prisms of the individual, the team, and the organization.
Brilliant podcast on differentiating your SaaS offerings with category creating products: Digital Enterprise Pathology Platform (Open System).
Code Whisperer: Those using Microsoft’s GitHub-Copilot AI assistant completed the task on average 55% faster.
Intro-to-Large-Language-Models (LLMs): A behind-the-scenes look at Pre-training and Fine Tuning, practical mechanics of next-word-prediction by ChatGPT and t...
According to Albert, right now we’re both in the middle of an unprecedented hype cycle for LLMs and starting to witness a transformation as profound as that ...
Discussion around the evolution of node.js and emerging alternatives. Plus, optimizations required to scale node.js beyond its current capabilities.
GitHub Actions integrates nicely into GitHub.com and lets you define and automate software development workflows.
Brilliant video by Prof. Andrew Ng: An incredible teacher who takes complex subjects and makes them very simple to understand!
AI has the potential to save the world — that’s the thesis of Marc’s most recent article (link in the episode website below)
Explore the process of building an ML team, including reasons to build one and descriptions of the core roles of project manager, data engineer, data scienti...
Meta is giving its answer to OpenAI’s GPT-4 away for free. The move could intensify the generative AI boom by making it easier for entrepreneurs to build pow...
A 2-sided marketplace relies on the “Flywheel Effect”: You need one side to build and feed the other side!
A brilliant explanation of LLMs: Including word-vectors, transformers comprising attention and feed-forward layers. Plus, some examples of early OpenAI GPT m...
LLAMA2 from Meta is an AI model (transformer neural network) developed using unsupervised learning with 2 trillion tokens with up 7 Billion parameters. LLA...
Using AI/ML for automating microscopy image analysis using diverse data sets and validation for variability.
How fast is generative AI moving? The first fully A.I.-generated drug from Insilico Medicine enters clinical trials in human patients.
ChatGPT nearly passed a radiology board–style examination without images, highlighting the potential of large language models, but radiologists should be awa...
The application of Machine Learning to drug discovery by designing, making and testing molecules.
Instead of the typical blame game, Etsy decided to take a more empathetic, constructive approach. They recognized that mistakes and system failures were more...
Best practices for continuous integration and delivery of software application code to production is discussed including the use of Trunk-Based-Development.
Loved this: Fascinating walk through history with the story of JSON, XML, SOAP, and Javascript is NOT Java!
Microservices architectures are more complex, however, more scalable to develop and deploy. A Monolithic architecture is simpler and easier to develop, deplo...
Repetitive automatable jobs will disappear: Jobs with creative applications to complex workflows will thrive in the post-ChatGPT world!
Security shouldn’t be an afterthought. Releasing code filled with exploits and bugs is a recipe for disaster. This is why more and more organizations are loo...
A conversation around GitHub Copilot X and the latest AI tools for software development. Plus, TypeScript 5 & continue the debate about the future of Rea...
Excellent discussion about React’s place in the frontend ecosystem, Single Page Application fatigue to Server Components and the future of the framework.
Excellent write up on using a “Reverse-Conway” maneuver to design your organization to meet the desired architecture for the business.
Docusaurus is a static site generator that builds a single-page app with fast client-side navigation using React. It provides out of the box documentation fe...
Conway’s Law is an IT theory created by computer scientist/programmer Melvin Conway in 1967. Conway’s Law states that “Organizations, who design systems, are...
DLT (Distributed Ledger Technology) leverages decentralized peer-to-peer computing and cryptography to verify and propagate a chain of transaction records ac...
Monorepos are a way to manage and organize your code used for software projects together in whatever way you find to be most logically consistent, and not ju...
Helpful guidelines for launching an AWS EC2 (Elastic Compute) virtual machine instance for running your applications.
Large Javascript frameworks continue to dominate, will there be a trend towards smaller ones?
Challenges with multiobject data including ACID atomic transactions that guarantee atomicity and consistency.
The evolution of Google’s automation for Site Reliability Engineering (SRE) for consistency, faster action and repairs.
New requirements for software security recommend the use of Software Bill of Materials (SBOM) to track dependencies, vulnerabilities and software licensing.
Shifting left on security for software developers is critical for delivering high quality code. Both static and dynamic analysis tools are necessary. This po...
An excellent guide for Software as a Medical Device (SaMD): Guidelines for Categorization, Risk, Software Development Life Cycle, Verification and Validation...
Good tutorial on node.js for beginners: Describes the “Single Threaded Event Loop” architecture of node: The Node.js processing model is based on the JavaScr...
People do mentor other people, they do support learning. And that is what creates 10x results in software development.
The tech team at LinkedIn working on protecting LinkedIn’s Members from bots and abusive scripted behavior, pioneering the use of deep learning to detect abu...
In October 2020, the University of Vermont Medical (UVM) Center fell victim to a cyberattack, and the leadership team intentionally shut down their system to...
A code review checklist, as well as clear rules and guidelines around code reviews, are crucial. A code review checklist can make your code review practice s...
The ChatGPT transformer is neural-network based with self-attention layers, allowing the model to weigh different words/phrases. See Vaswani et al. 2017 pape...
Embedding security analysis tools directly into the software developer’s workflow. “Shifting-Left” on security! Software Engineering Unlocked
Effective testing, domain driven tests, writing testable code and maintainable test code. Software Engineering Unlocked
Collaborative software debugging tools to improve developer productivity and quality. Software Engineering Unlocked
A practical guide for accelerating code to production using a serverless architecture and a Domain Driven Design (DDD) approach: Understand events, involve e...
Excellent description of the basics of a typical web based system: DNS, Load Balancer, Web App Servers, Databases, Caching, Job Queues and Services, Search, ...
Decent podcast on CSS in 2023. There is a nice discussion here on grid and flexbox with some ideas learning CSS in 2023! Enjoy
The author didn’t write this story, an AI algorithm did…
In this post you’ll learn how to configure a development environment with live-reload enabled. This will allow you to convert a legacy application so it uses...
Great advice for software engineers from a perspective of culture, self-advocacy and mentorship.
How to incorporate Agile-Scrum in a Software-As-A-Medical device environment? Some ideas suggested: Follow agile-scrum within the overall release cycle re...
Podcast discussing container registries: Amazon ECR, Docker Hub and GitHub’s container registry. Covers a few items: Immutable tag settings, Lifecycle manag...
Importance of Test Driven Development Balancing unit, integration and contract testing Pair programming
Wonderful guidelines for selecting gitflow versus trunk-based-development for software code source control. Really depends on your team and business model. @...
Security Guidelines: Prevention against Sql injection using Parametrized queries, ORM usage, sanitize inputs, and escaped inputs.
Great review of Google’s SRE book that includes the 4 golden signal’s. If you do noting else, monitor the following: Latency, Traffic (e.g. http requests per...
This blog from NGNIX does a great job of adapting the Twelve-Factor App to suit microservices-based applications.
Interesting podcast on the use of GitHub Copilot for code reviews!
Happy developers are more productive. What makes them happy? To mention a few: Collaborative team, good manager, solid architecture and your personal impact ...
High level: DevOps is more on the CI/CD side and SRE (Site Reliability Engineering) focused more on the operations side of the fence. Conway’s law applies, s...
Meta’s Segment Anything Model (SAM) is a new and open-source computer vision model for image segmentation. It is as revolutionary to images as ChatGPT is to ...
Great Blog on how transfomers work. They are formed by several blocks, each one with its own function, working together to understand text and generate the n...
The 5 mistakes to avoid while using Machine Learning for Digital Pathology: Not understanding your data, Not researching prior art carefully, Starting with a...
The use of Real World Data (RWD) to create targeted therapies for chronic conditions. Fascinating discussion about its ability to transform biopharma researc...
The research team used a newly released version of artificial intelligence software called AlphaFold (DeepMind/Google). AlphaFold predicts the three-dimensio...
There are plenty of indicators demonstrating the effectiveness of Digital Pathology (DP) and Augmented Intelligence (AI) on the modern Pathology practice. DP...
GPT-4 (Generative Pretrained Transformer) is the engine behind the latest version of ChatGPT and the updated Bing search engine from Microscoft. This Technic...
The intersection of image analysis, AI/Machine Learning and pathology! Methods and techniques to extract insights from image analysis.
An engineering founder shares his views on search and ChatGPT, plus the need for enterprise search. Fascinating journey from Silicon Valley to Telaviv to Ind...
A data scientist turned career coach shares her personal journey to finding a job in this area and helping others find their path!
AI models can only reach their full potential if a few other key components are in place: 1) quality control, 2) generalizable models, and 3) proper validati...
Are we ready for Neuro-AI and Brain Computer Interfaces? The Data Scientist Show
Personal journey towards Data Science and things that only 5 experienced data scientist know! The Data Scientist Show
The work shows that AI has the potential to make a time-sensitive, critical diagnosis easier and more accessible to pathologists.
Applying AI/ML in Healthcare and scalability of these solutions.
Excellent overview of value proposition of dbt as a data-transform tool in an ELT process.
Discussion on the importance of quality control systems in an ML pipeline, model generalizability, and how the regulatory process affects ML development.
Using synthetic data and models to advance AI with innovative federated data models that balance privacy and availability
As the number of devices increases, the agency is looking to adapt its regulatory framework to the new technology, including faster approval of algorithm upd...
Interesting podcast focused on computational pathology using the semi and weakly supervised approaches with Whole Slide Images (WSI). It turns out that those...
This podcast describes an incredible journey starting with Image Analysis to including AI/ML deep learning capabilities
Good overview of the application of Deep MIL to Image Based Diagnosis Deep MIL
Every Data Scientist needs to be a “Product Manager”: Understand the full picture: user needs and pushing back on requirements! Asking question. Finally, whi...
AI/Machine-Learning Convolution Neural Networks (CNNs) are at the heart of many major image classification systems. I found a great blog on FreeCodeCamp (lin...
Great advice on what it takes to be great data scientist and data engineer: A data-centric inquiring mind, Desire to pick up some tech skills over time and,...
Brilliant conversation on data-oriented programming with JSON schema front-and-center versus Object Oriented Programming including strongly typed systems. D...
A great story: A tenacious and humble engineer @loganKilpatrick describes his journey from the Apple store to an ML engineer at Apple! Apple-Retail-to-ML