Femtech World’s Post

AI revolutionizes metastatic cancer diagnosis. A team of scientists has developed an artificial intelligence (AI) tool capable of identifying the origins of metastatic cancer cells circulating in the body, surpassing even pathologists in accuracy. They trained their AI model with 30,000 images of cells from 21,000 people whose tumor origins were known and successfully tested it on 27,000 images, achieving an 83% accuracy in predicting the tumor's origin. Furthermore, the model could identify the tumor's origin within the top three predictions with 99% certainty. This tool could revolutionize the diagnosis and treatment of advanced cancer, allowing doctors to reduce the need for invasive additional tests. However, the predictions were limited to 12 common sources of cancer, excluding some types that typically do not spread to abdominal or lung fluids, such as prostate and kidney cancer. When tested on 500 images, the model outperformed human pathologists in predicting the tumor's origin, demonstrating a statistically significant improvement. Additionally, in a retrospective study of 391 cancer-treated participants, those whose treatment matched the model's prediction were more likely to survive and live longer. This finding supports the use of the AI model in clinical settings. Mahmood, who has previously used AI to predict cancer origin from tissue samples, suggests that combining cell, tissue, and genomic data could further improve outcomes for people with metastatic cancers of unknown origins. Source: https://lnkd.in/d5R4qeTR #AIcancerdiagnosis #MetastaticCancer #AIinHealthcare #CancerResearch #PrecisionMedicine #TheraResearch

Predicting cancer treatment response with AI 🤖   Using radiological images of patients’ tumours, a research team led by Dr. Benjamin Haibe-Kains developed an AI tool that predicts individual tumour response to chemotherapy in metastatic lung sarcoma:   https://lnkd.in/gwsA6wKY   When cancer cells spread to different sites in the body and the condition becomes metastatic, the tumours that develop at various locations may appear similar but often respond differently to treatments.   “We wanted to develop a reliable model for predicting how individual metastatic tumours will respond to standard treatments,” says Caryn Geady, a PhD candidate in Benjamin’s lab and the first author of the study.   Acquired from a clinical trial, CT images of lung metastases prior to treatment were compared to those taken after two cycles of systemic chemotherapy. The team evaluated the change in the tumour volume to develop a radiomic model that can predict tumour progression at each metastatic site.   The team found that, after treatment with doxorubicin, while most tumours shrank in size, 15.6% were larger than before, showing a progressive response. In patients with more than two metastatic sites, 18.5% had different tumour size responses at different sites. The machine-learning model was fine-tuned using a specialized form of logistic regression, and the team aims to validate it in other cancerous tissues.   “Our model confirmed that when cancer cells spread to multiple locations in lung tissues, they don’t all behave the same to therapy,” says Dr. David Shultz, a senior author of the study.   “It encourages us to address metastatic disease with a data-driven approach: if we can identify sites predicted to be progressive and remove them while applying systemic chemotherapy, it might bring us closer to eradicating the disease.” -------- Published in Computerized Medical Imaging and Graphics on Jun 25, 2024. Study funded by The Princess Margaret Cancer Foundation and The National Institutes of Health. Research at UHN UHN Office of Research Trainees #cancerresearch #lungcancer #metastatic #machinelearning #ai

Congratulations to PhD Student, Caryn Geady, on your new publication! Check out the paper to learn about a new AI tool for predicting cancer treatment response!

AI Revolutionizes Prostate Cancer Treatment: Breakthrough Discovery Leads to Personalized Medicine In a medical milestone that could save thousands of lives, researchers have harnessed the power of artificial intelligence to unlock a secret about prostate cancer – it's not just one disease. This groundbreaking discovery paves the way for personalized treatments, transforming the battle against a cancer that affects millions of men worldwide. Key Breakthroughs: Two Distinct Subtypes Identified: Using sophisticated AI algorithms to analyze DNA data, researchers discovered that prostate cancer actually consists of two distinct subtypes. The Promise of Tailored Therapy: This finding means treatments can be customized based on a patient's specific cancer type, dramatically increasing success rates and minimizing unnecessary side effects. AI as a Medical Game-Changer: The study demonstrates the immense potential of artificial intelligence to drive innovation in healthcare and revolutionize disease diagnosis and treatment. The Larger Picture Funded by Cancer Research UK, this breakthrough represents a shift towards precision medicine in prostate cancer care. Doctors hope that in the future, a simple genetic test aided by AI could provide a blueprint of a patient's cancer, guiding them to the most effective treatment. This "divide and conquer" approach, similar to strategies used in breast cancer treatment, could vastly improve outcomes. Professor Colin Cooper of the University of East Anglia, involved in the research, remarks, "Until now, we assumed prostate cancer was a single disease. Thanks to artificial intelligence, we now understand that it comes in two very different forms. This has profound implications for how we treat patients in the future." The potential of this breakthrough is immense, offering men with prostate cancer greater hope and the possibility of more effective, individualized treatments. #AI #artificialintelligence #AIinMedicine #DNAanalysis #MedicalBreakthrough #CancerResearch #prostatecancer #PersonalizedMedicine #PrecisionMedicine

Is AI the future of healthcare? Lung cancer is the leading cause of cancer-related deaths globally with 1.8 million deaths reported in 2020. But thanks to #generativeAI, screening and detecting lung cancer is going to be more efficient and accurate. Recently, Google has previously developed machine learning (ML) models for lung cancer detection, and have evaluated their ability to automatically detect and classify regions that show signs of potential cancer. Performance has been shown to be comparable to that of specialists in detecting possible cancer. As AI continues to evolve and integrate into healthcare systems, it offers the prospect of not only improving diagnostic accuracy but also optimizing treatment plans, enhancing patient care, and ultimately reshaping the landscape of medicine as we know it. Read more about it below. #AIinhealthcare #AIforAll

Using AI to beat up Cancer🤖🥊😵 AI can be a force for good. This week I want to talk about two groundbreaking examples of AI supporting the fight against cancer. Early cancer detection is a major factor in cancer survival rate - the sooner the cancer is identified, the sooner life-saving treatment can be administered. The challenge is that cancer can be difficult to identify at early stages. Step forward Mia, the AI-powered tool that can detect breast cancers. Developed by Kheiron Medical, Mia has been trialled by Dr Gerald Lip and his team within NHS Grampian.⚕️ Initial results based on 10000 patients show that Mia correctly assessed all the cancer cases identified by medical experts, with an additional 11 cases correctly identified by Mia only. That's 11 cases that would have been spotted at a later stage of the cancer. In parallel, we have Pi (Prostate Intelligence ©️) developed by Cambridge-based Lucida Medical Ltd and endorsed by Somerset NHS Foundation Trust . 🇬🇧 In what is thought to be an industry-pioneering move, Pi has been officially adopted by Dr Paul Burn and his team to assist with prostate cancer detection. It can accurately assess the probability of cancer as well as precise location of the cancer using an MRI scan. Does this replace expert radiologists? Of course it doesn't. Human oversight is essential to review and validate AI findings.📝✅ Amongst all the AI hype, it is refreshing to highlight actual examples of AI making a difference for the well-being of humans. ➡️Useful links in the comments, including NHS cancer screening info. Ps: special mention to Peter G. for bringing the Pi technology to my attention. #ai #cancer #NHS

AI Shows Promise in Early Lung Cancer Detection via Chest X-rays - HIT Consultant #AIinHealthcare #LungCancerDetection #ChestXrays #HealthcareIT AI technology is making significant strides in early lung cancer detection through the analysis of chest X-rays. By utilizing advanced algorithms, AI can accurately identify potential signs of lung cancer at an early stage, allowing for timely intervention and improved patient outcomes. #BenefitsOfAI #ImprovedAccuracy One of the key benefits of AI in lung cancer detection is its ability to enhance the accuracy of diagnosis. By leveraging machine learning capabilities, AI can analyze vast amounts of data from chest X-rays with precision, reducing the risk of human error and ensuring more reliable results. #ChallengesInAdoption #IntegrationIssues Despite the promising potential of AI in early lung cancer detection ai.mediformatica.com #cancer #lung #lungcancer #qure #study #qureai #chestxrays #thailand #advanced #aipowered #cancerdetection #conference #digitalhealth #healthit #healthtech #healthcaretechnology @MediFormatica (https://buff.ly/3XmS7Y7)

AI tool finds cancer signs missed by doctors In a groundbreaking development, an AI tool named Mia has demonstrated remarkable capabilities in detecting signs of cancer that were missed by human radiologists. Mia was put to the test alongside NHS clinicians in the UK, analyzing mammograms from a cohort of over 10,000 women. Despite most participants being cancer-free, Mia successfully flagged all cases showing symptoms of breast cancer, including an additional 11 cases overlooked by doctors. Only 81 out of the 10,889 women opted not to have their scans reviewed by Mia. Trained on a dataset of 6,000 previous breast cancer cases, Mia learned intricate patterns and imaging biomarkers associated with malignant tumors. Its performance in the trial was impressive, correctly predicting cancer presence with 81.6% accuracy and ruling it out accurately 72.9% of the time. Breast cancer affects millions globally, with early detection crucial for improved survival rates. While advancements have been made, patients often endure severe side effects post-treatment. Now, researchers are enhancing Mia to predict patients' risk of side effects like lymphoedema up to three years post-treatment. This personalized approach could lead to tailored care plans, mitigating risks for high-risk patients. A clinical trial named Pre-Act will validate Mia's risk prediction model in 780 breast cancer patients over two years, aiming for an AI system that comprehensively evaluates prognosis and treatment needs. This AI breakthrough signals a transformative shift in cancer care, empowering clinicians with advanced tools to improve detection, treatment, and patient outcomes. #CancerDetection #AIBreakthrough #PersonalizedMedicine #MedicalInnovation

"For cancer patients who have exhausted the first line of treatments, a clinical trial, in which new treatments are tested, can offer a best last hope. The hard part is finding one. The percentage of U.S. cancer patients participating in a trial is in the single digits, yet, ironically, lack of enrollment is a key reason clinical trials fail." At Providence, generative AI is starting to be used to comb through mountains of data to help match patients to clinical trials — a big barrier to enrollment — thus, unlocking new possibilities in advancing cancer care and research. This joint work between Providence and Microsoft is a compelling example of how AI is advancing precision medicine, and hopefully, helping us find cures that will lead to the eventual eradication of many diseases. #cancer #AI #research

Researchers at the University of British Columbia (UBC) have developed an artificial intelligence (AI) model capable of identifying patterns in images of endometrial cancer cancer cells. These patterns indicate a higher risk of recurrence and death, which could go unnoticed by traditional diagnostic methods. AI can assist doctors in treating high-risk patients, enabling more effective and comprehensive strategies. In collaboration with machine learning expert Ali Bashashati, PhD, the UBC team created an AI model to analyze images of samples collected from cervical cancer patients. The model was trained with 2,300 images of cancerous tissue to distinguish between different subtypes of endometrial cancer. This approach can help ensure that no patient misses the opportunity for potentially life-saving interventions. https://lnkd.in/gkBi-vGM