• AI Grand Rounds
  • Posts
  • Big Tech's AI penetration in different Healthcare sectors: Hospitals to Industry

Big Tech's AI penetration in different Healthcare sectors: Hospitals to Industry

AI's benefit in cancer diagnosis

Fellow Healthcare Champions,

Are you overwhelmed by all the fluff and hype around AI and not sure how to identify meaningful information? We get it. As busy clinicians ourselves, our newsletter, "AI Grand Rounds," is here to provide clinically relevant AI information.

No matter who you are—a healthcare provider, inventor, investor, or curious reader—we PROMISE one thing: you will always find information that is relevant, meaningful, and balanced.

Let’s join our journey to share clinically meaningful and relevant knowledge on healthcare AI.

Sincerely,

Your fellow physicians!

Table of Contents

🚨 Pulse of Innovation 🚨

Breaking news in the healthcare AI

Big Tech, Big Hospitals, Big Healthcare/Pharma: How are they partnering to improve care

The fusion of artificial intelligence (AI) with healthcare is gaining momentum, fueled by strategic partnerships between big tech companies, hospitals, and major healthcare organizations.

  • These collaborations combine advanced AI technologies, vast computational resources, and deep clinical expertise to revolutionize patient care, optimize hospital operations, and innovate medical solutions.

  • This convergence is unlocking new possibilities, though it also sparks debates over data security, ethical implications, and accessibility. Big tech giants such as Google, Microsoft, Amazon, and IBM are at the forefront, offering AI algorithms, cloud platforms, and data analytics.

  • Hospitals provide real-world patient data and operational insights, while big healthcare companies—spanning insurers, health systems, and medical device makers—bring scale, infrastructure, and industry-specific knowledge.

  • Together, they aim to enhance diagnostics, personalize treatments, and improve healthcare delivery.

A standout example is Google’s partnership with Ascension, one of the largest U.S. healthcare systems. Launched in 2019 and expanded since, this collaboration uses Google Cloud to analyze patient data from Ascension’s network of hospitals and clinics. AI models process electronic health records (EHRs), lab results, and imaging to predict patient deterioration, streamline workflows, and support clinical decisions. This effort demonstrates how tech can amplify a healthcare system’s ability to manage complex patient populations efficiently. Google’s has since partnered with Mayo clinic and HCA Healthcare to bring generative AI solutions to hospitals for patient care as well as clinical decision making.

Microsoft’s AI for Health program offers another compelling case. Through this initiative, Microsoft collaborates with Providence, a major healthcare system operating dozens of hospitals across the U.S. West Coast. By leveraging Azure’s AI capabilities, Providence enhances its telehealth services and optimizes resource allocation—critical during surges like the COVID-19 pandemic. Microsoft also partners with healthcare giant Novartis to refine drug development, showing its dual focus on hospital operations and broader healthcare innovation. More recently Microsoft worked with University of Washington Cancer Center for improved cancer diagnosis using AI (see below for similar AI tools). Microsoft has also partnered with Johns Hopkins University of Pancreatic cancer diagnosis.

Amazon has entered the fray via AWS, partnering with healthcare organizations like Intermountain Healthcare, a Utah-based system with hospitals, clinics, and insurance plans. AWS’s AI tools analyze patient data to improve care coordination, predict readmissions, and reduce costs. Additionally, Amazon collaborates with GE Healthcare, a leader in medical imaging, to integrate AI into diagnostic equipment. This partnership enhances the accuracy of scans like MRIs and CTs, enabling earlier detection of conditions such as cancer or neurological disorders.

NVIDIA is also at forefront of this partnership with collaboration with multiple organizations like IQVIA, Mayo clinic, PHILIPS, CHULALONGKORN University, HYPERFINE. The HYPERFINE is bringing portable MRI solutions using AI to bedside. CHULALONGKORN University is using AI tools for improved endoscopy based diagnosis.

Not all of these partnerships have been fruitful. The IBM’s Watson Health collaborated with Memorial Sloan Kettering Cancer Center, a top-tier hospital, to train AI on oncology data, aiding doctors in crafting personalized cancer treatment plans. However it was pre-generative AI period and could have been the reason for not a successful partnership.

These alliances are transformative but not without challenges. Privacy concerns loom large as sensitive patient data moves across tech platforms, requiring strict adherence to regulations like GDPR and HIPAA. The high cost of AI adoption also risks excluding smaller healthcare providers, potentially deepening inequities. Ethical questions about AI-driven decisions in life-and-death scenarios persist, demanding transparent governance. Nevertheless, the collaboration between big tech, hospitals, and big healthcare companies is reshaping medicine. From Google’s predictive analytics to Amazon’s diagnostic enhancements, these examples highlight AI’s potential to elevate care quality and efficiency. As this ecosystem evolves, balancing innovation with accountability will be key to ensuring that AI’s benefits extend across the healthcare landscape, reaching patients and providers alike.

🧑🏼‍🔬 Bench to Bedside👨🏽‍🔬

Developments in healthcare AI research and innovations

Machine Learning in Diagnosis and Treatment of Lung Nodules: Approaching Targeted Personalized Care

Lung cancer, a major malignant tumor with highest mortality rate and economic burden, has been a concern for healthcare providers globally. With use of low-dose computed tomography (LDCT), the early screening, timely diagnosis and treatment has been possible. Every pulmonary nodule detected during chest imaging must be distinguished whether it is a malignant one or has potential to be a malignant one in future.

Study findings:

Total of 45064 patients were included in the medical checkup cohort (MCC)- training and internal testing and 14,437 participants in a mobile screening cohort (MSC)- independent testing. A triage driven Chinese Lung Nodules Reporting and Data System (C-Lung-RADS) was used. The study was subdivided in three phases: Phase 1- Initial Risk Classification: In this, pulmonary nodules on the basis of size and density were sub-classified as low, mid, high and extremely high risk and routine follow up was done at 1 year, 6 months, 3 months and immediate clinical assessment accordingly. Phase 2 -AI based Risk Stratification- check for malignancy probability and Phase 2+: Refined risk with additional information- distinguish whether the pulmonary nodule is malignant or not.

C-Lung-RADS (phase 1 classification tree) outperformed Lung-RADS v2022 in classifying nodules, with an AUC of 0.899, identifying 2.9% of high-risk cases and 19.3% of extremely high-risk cases, compared to 1.4% and 13.6% for Lung-RADS v2022 respectively. The accuracy of diagnosis could be increased exponentially using molecular biomarkers.

Benefit of use of LDCT:

  • Accurate estimations of malignancy risk of pulmonary nodules.

  • Avoid missed diagnosis, late diagnosis and unnecessary biopsy procedures.

Limitations:

  • AI in healthcare could widen the gap between those with and without access to advanced care.

  • Early stage lung cancer nodules which turn out to be slow growing Lung Ca can be missed.

  • Follow-up times (3, 6, and 12 months) need to be more precise.

  • The system was tested in Western China, so it needs to be checked in other areas as well.

In a nutshell, with the use of LDCT, a multi-step method of assessing the cancer risk of lung nodules is possible which improves the early detection of lung cancer, optimization of healthcare resources, enhances accuracy as well as minimizes the rate of delay diagnosis eventually leading to a better health of those in needs.

🧑🏽‍⚕️ AI in Clinic 🏥

Developments in healthcare AI research and innovations

Transforming Lung Nodule Diagnosis with FDA cleared AI tools

Lung cancer remains one of the leading causes of cancer-related mortality worldwide, with early detection being crucial for improving patient outcomes. Advances in artificial intelligence (AI) are revolutionizing the identification and evaluation of pulmonary nodules.

F.A.S.T.® aiCockpit®: A Game Changer in Lung Nodule Diagnosis

F.A.S.T.® aiCockpit® is an AI-driven diagnostic tool designed to assist radiologists in detecting, analyzing, and reporting lung nodules more efficiently. The software integrates seamlessly into existing Picture Archiving and Communication Systems (PACS), AI platforms, and reporting systems, allowing radiologists to access AI-generated insights within their current workflows.

Key Features & Clinical Advantages

1. AI-Powered Nodule

  • The software automatically identifies and highlights lung nodules on CT scans, aiding radiologists in detecting even subtle abnormalities.

  • It provides 2D and 3D visualizations, allowing for a more comprehensive analysis of nodule morphology and growth over time.

2. Seamless Workflow Integration

  • Radiologists can modify AI-generated findings, accept or reject results, and make adjustments without switching between applications.

3. Automated and Interactive Reporting

  • Allows interaction with AI-generated findings in real-time, refining measurements and descriptions.

  • Once finalized, reports can be automatically sent to archives and clinical reporting systems, reducing documentation time.

How is it Different from Other AI Solutions?

  1. Universal AI Viewer

    -It can support multiple AI algorithms in a single installation.

    -Unlike other solutions requiring separate integrations for different AI tools, this software consolidates multiple algorithms, enhancing efficiency and reducing costs.

  2. Real-Time Interactivity & Customization

    -F.A.S.T.® aiCockpit® enables radiologists to fine-tune AI-generated findings, ensuring that the final report aligns with their clinical judgment.

  3. FDA 510(k) Clearance & Clinical Readiness

    -The software has recently received FDA 510(k) clearance, ensuring it meets regulatory standards for clinical use.

Qure.ai’s AI Solutions for Lung Nodule Detection

Qure.ai has developed a suite of AI-powered tools aimed at improving the detection and analysis of lung nodules:

qXR-LN: AI for Chest X-rays

  1. Functionality:

    -qXR-LN is an AI-driven software that analyzes chest X-rays to identify and localize lung nodules ranging from 6 to 30 mm in size.

    -It serves as a computer-aided detection tool, assisting radiologists, pulmonologists, and emergency room physicians in the early detection of potential malignancies.

  2. Regulatory Clearance: 

    -In January 2024, qXR-LN received FDA clearance.

    -Notably, it is the only FDA-cleared solution for detecting and localizing lung nodules utilizing computer vision, with radiologists, pulmonologists, and ER physicians as intended users.

  3. Clinical Validation: https://www.sciencedirect.com/science/article/pii/S1076633224008481

qCT LN Quant: AI for Chest CT Scans

  1. Functionality:

    • IT analyzes non-contrast chest CT scans, providing advanced quantitative characterization of solid lung nodules including average, short-axis, long-axis, and effective diameters, and tracks volumetric growth over time.

    • The tool also generates 2D and 3D reconstructions, Brock malignancy risk scores, and management suggestions based on Fleischner Society guidelines.

  2. Regulatory Clearance:

    -In August 2024, qCT LN Quant received FDA clearance.

  3. Clinical Application:

    -The tool assists clinicians in analyzing morphological data across single or multiple thoracic studies, including estimating volume doubling time and tracking nodules over several time points, thereby supporting consistent and reliable clinical decision-making.

qTrack: Comprehensive Lung Nodule Management

  1. Functionality:

    -It enables care coordination through smart prompts, hardware-agnostic image viewing for AI-annotated chest X-rays and CTs, cross-departmental scan sharing, and custom notifications for suspect cases.

In summary, while both Fovia AI and Qure.ai offer AI solutions aimed at improving lung nodule detection and management, their approaches differ in terms of imaging modalities, integration capabilities, and the range of tools provided to clinicians. Fovia AI’s solution focuses on enhancing CT scan analysis with seamless integration into existing systems, whereas Qure.ai provides a comprehensive suite that includes tools for both chest X-rays and CT scans, along with a dedicated platform for ongoing patient management.

🤖 Patient First, AI Second🤖

Ethical and Regulatory Landscape of Healthcare AI

AI based assessment of patient reported outcomes: Better evaluation for improved care

The quality of care provided and safety are the key indicators of patient experience in hospitals. Patient-reported experience measures (PREMs) is commonly used as a tool to assess patient satisfaction in a hospital. Manually analyzing the free-text data is a cumbersome task which is unsustainable for the rapidly increasing acute care populations. In the digital age, Natural language processing (NLP) and machine learning (ML) are key tools for analyzing patient feedback data.

Study findings:

In order to focus on AI- powered systems to routinely analyze PREMs in hospitals and health services in Australia, this cross-sectional study was conducted. Further, this study addressed the research and practice gap under various subtopics.

Total statewide public hospital (n=114) and health service (n=16) network were included in the study. It was found that the manual technique was time consuming, had a higher rate of errors, needed a large task force for management, breach in data and privacy as well as less feasible.

Benefits of AI in PREMs

·       Technically feasible - faster data sharing between the health service provider.       

·       Economic feasibility - Reduced labor, reduce data duplication, and improve efficiency which will eventually be cost effective.

·       Legal feasibility- Higher data security, transparency and increased patients trust, tamper proof records.

·       Operational feasibility- Risk as well as error reduction, improved the outcome and user friendly.

·       Schedule feasibility- Improve project planning and reduce risk of delay.

To sum up, AI use in PREMs can act as a positive catalyst to enhance the clinical assessment and increase the patients’ satisfaction which will eventually improve the quality of service provided.

Disclaimer: This newsletter contains opinions and speculations and is based solely on public information. It should not be considered medical, business, or investment advice. This newsletter's banner and other images are created for illustrative purposes only. All brand names, logos, and trademarks are the property of their respective owners. At the time of publication of this newsletter, the author has no business relationships, affiliations, or conflicts of interest with any of the companies mentioned except as noted. ** OPINIONS ARE PERSONAL AND NOT THOSE OF ANY AFFILIATED ORGANIZATIONS!

Reply

or to participate.