MSGID: 1:317/3 64a8e690   
   PID: hpt/lnx 1.9.0-cur 2019-01-08   
   TID: hpt/lnx 1.9.0-cur 2019-01-08   
    AI tool decodes brain cancer's genome during surgery    
    Real-time tumor profiling can guide surgical, treatment decisions    
      
     Date:   
         July 7, 2023   
     Source:   
         Harvard Medical School   
     Summary:   
         New AI tool enables in-surgery genomic profiling of gliomas, the   
         most aggressive and most common brain tumors. This information   
         offers critical clues about how aggressive a cancer is, its   
         future behavior, and its likely response to treatment. The tool   
         can provide real-time guidance to surgeons on the optimal surgical   
         approach for removal of cancerous tissue.   
      
      
         Facebook Twitter Pinterest LinkedIN Email   
      
   ==========================================================================   
   FULL STORY   
   ==========================================================================   
   Scientists have designed an AI tool that can rapidly decode a brain   
   tumor's DNA to determine its molecular identity during surgery -- critical   
   information that under the current approach can take a few days and up   
   to a few weeks.   
      
   Knowing a tumor's molecular type enables neurosurgeons to make decisions   
   such as how much brain tissue to remove and whether to place tumor-killing   
   drugs directly into the brain -- while the patient is still on the   
   operating table.   
      
   A report on the work, led by Harvard Medical School researchers, is   
   published July 7 in the journalMed.   
      
   Accurate molecular diagnosis -- which details DNA alterations in a cell -   
   - during surgery can help a neurosurgeon decide how much brain tissue to   
   remove. Removing too much when the tumor is less aggressive can affect   
   a patient's neurologic and cognitive function. Likewise, removing too   
   little when the tumor is highly aggressive may leave behind malignant   
   tissue that can grow and spread quickly.   
      
   "Right now, even state-of-the-art clinical practice cannot profile   
   tumors molecularly during surgery. Our tool overcomes this challenge by   
   extracting thus-far untapped biomedical signals from frozen pathology   
   slides," said study senior author Kun-Hsing Yu, assistant professor of   
   biomedical informatics in the Blavatnik Institute at HMS.   
      
   Knowing a tumor's molecular identity during surgery is also valuable   
   because certain tumors benefit from on-the-spot treatment with drug-coated   
   wafers placed directly into the brain at the time of the operation,   
   Yu said.   
      
   "The ability to determine intraoperative molecular diagnosis in real   
   time, during surgery, can propel the development of real-time precision   
   oncology," Yu added.   
      
   The standard intraoperative diagnostic approach used now involves taking   
   brain tissue, freezing it, and examining it under a microscope. A major   
   drawback is that freezing the tissue tends to alter the appearance of   
   cells under a microscope and can interfere with the accuracy of clinical   
   evaluation.   
      
   Furthermore, the human eye, even when using potent microscopes, cannot   
   reliably detect subtle genomic variations on a slide.   
      
   The new AI approach overcomes these challenges.   
      
   The tool, called CHARM (Cryosection Histopathology Assessment and Review   
   Machine), is freely available to other researchers. It still has to be   
   clinically validated through testing in real-world settings and cleared   
   by the FDA before deployment in hospitals, the research team said.   
      
   Cracking cancer's molecular code Recent advances in genomics have   
   allowed pathologists to differentiate the molecular signatures -- and   
   the behaviors that such signatures portend - - across various types   
   of brain cancer as well as within specific types of brain cancer. For   
   example, glioma -- the most aggressive brain tumor and the most common   
   form of brain cancer -- has three main subvariants that carry different   
   molecular markers and have different propensities for growth and spread.   
      
   The new tool's ability to expedite molecular diagnosis could be   
   particularly valuable in areas with limited access to technology to   
   perform rapid cancer genetic sequencing.   
      
   Beyond the decisions made during surgery, knowledge of a tumor's   
   molecular type provides clues about its aggressiveness, behavior,   
   and likely response to various treatments. Such knowledge can inform   
   post-operative decisions.   
      
   Furthermore, the new tool enables during-surgery diagnoses aligned with   
   the World Health Organization's recently updated classification system   
   for diagnosing and grading the severity of gliomas, which calls for such   
   diagnoses to be made based on a tumor's genomic profile.   
      
   Training CHARM CHARM was developed using 2,334 brain tumor samples from   
   1,524 people with glioma from three different patient populations. When   
   tested on a never-before- seen set of brain samples, the tool   
   distinguished tumors with specific molecular mutations at 93 percent   
   accuracy and successfully classified three major types of gliomas with   
   distinct molecular features that carry different prognoses and respond   
   differently to treatments.   
      
   Going a step further, the tool successfully captured visual   
   characteristics of the tissue surrounding the malignant cells. It was   
   capable of spotting telltale areas with greater cellular density and   
   more cell death within samples, both of which signal more aggressive   
   glioma types.   
      
   The tool was also able to pinpoint clinically important molecular   
   alterations in a subset of low-grade gliomas, a subtype of glioma that   
   is less aggressive and therefore less likely to invade surrounding   
   tissue. Each of these changes also signals different propensity for   
   growth, spread, and treatment response.   
      
   The tool further connected the appearance of the cells -- the shape of   
   their nuclei, the presence of edema around the cells -- with the molecular   
   profile of the tumor. This means that the algorithm can pinpoint how a   
   cell's appearance relates to the molecular type of a tumor.   
      
   This ability to assess the broader context around the image renders the   
   model more accurate and closer to how a human pathologist would visually   
   assess a tumor sample, Yu said.   
      
   The researchers say that while the model was trained and tested on   
   glioma samples, it could be successfully retrained to identify other   
   brain cancer subtypes.   
      
   Scientists have already designed AI models to profile other types of   
   cancer - - colon, lung, breast -- but gliomas have remained particularly   
   challenging due to their molecular complexity and huge variation in   
   tumor cells' shape and appearance.   
      
   The CHARM tool would have to be retrained periodically to reflect new   
   disease classifications as they emerge from new knowledge, Yu said.   
      
   "Just like human clinicians who must engage in ongoing education and   
   training, AI tools must keep up with the latest knowledge to remain at   
   peak performance."  Authorship, funding, disclosures Coinvestigators   
   included MacLean P. Nasrallah, Junhan Zhao, Cheng Che Tsai, David   
   Meredith, Eliana Marostica, Keith L. Ligon, and Jeffrey A. Golden.   
      
   This work was supported in part by the National Institute of General   
   Medical Sciences grant R35GM142879, the Google Research Scholar Award,   
   the Blavatnik Center for Computational Biomedicine Award, the Partners   
   Innovation Discovery Grant, and the Schlager Family Award for Early-Stage   
   Digital Health Innovations.   
      
       * RELATED_TOPICS   
             o Health_&_Medicine   
                   # Brain_Tumor # Cancer # Lung_Cancer   
             o Mind_&_Brain   
                   # Brain-Computer_Interfaces # Intelligence # Brain_Injury   
             o Computers_&_Math   
                   # Neural_Interfaces # Computer_Modeling # Communications   
       * RELATED_TERMS   
             o Histology o Cancer o Robotic_surgery o Aggression o Surgery   
             o Colorectal_cancer o Breast_cancer o Bullying   
      
   ==========================================================================   
      
    Print   
      
    Email   
      
    Share   
   ==========================================================================   
   ****** 1 ****** ***** 2 ***** **** 3 ****   
   *** 4 *** ** 5 ** Breaking this hour   
   ==========================================================================   
       * Six_Foods_to_Boost_Cardiovascular_Health   
       * Cystic_Fibrosis:_Lasting_Improvement *   
       Artificial_Cells_Demonstrate_That_'Life_...   
      
       * Advice_to_Limit_High-Fat_Dairy_Foods_Challenged   
       * First_Snapshots_of_Fermion_Pairs *   
       Why_No_Kangaroos_in_Bali;_No_Tigers_in_Australia   
       * New_Route_for_Treating_Cancer:_Chromosomes *   
       Giant_Stone_Artefacts_Found:_Prehistoric_Tools   
       * Astonishing_Secrets_of_Tunicate_Origins *   
       Most_Distant_Active_Supermassive_Black_Hole   
      
   Trending Topics this week   
   ==========================================================================   
   HEALTH_&_MEDICINE Birth_Defects Cholesterol   
   Patient_Education_and_Counseling MIND_&_BRAIN Autism Creativity Depression   
   LIVING_&_WELL Healthy_Aging Fitness Nutrition   
      
      
   ==========================================================================   
      
   Strange & Offbeat   
   ==========================================================================   
   HEALTH_&_MEDICINE Holograms_for_Life:_Improving_IVF_Success   
   Grocery_Store_Carts_Set_to_Help_Diagnose_Common_Heart_Rhythm_Disorder_and   
   Prevent_Stroke DNA_Can_Fold_Into_Complex_Shapes_to_Execute_New_Functions   
   MIND_&_BRAIN AI_Tests_Into_Top_1%_for_Original_Creative_Thinking   
   Everyone's_Brain_Has_a_Pain_Fingerprint_--_New_Research_Has_Revealed_for_the   
   First_Time   
   Scientists_Discover_Spiral-Shaped_Signals_That_Organize_Brain_Activity   
   LIVING_&_WELL Illusions_Are_in_the_Eye,_Not_the_Mind   
   Amputees_Feel_Warmth_in_Their_Missing_Hand   
   Why_Do_Champagne_Bubbles_Rise_the_Way_They_Do?_Scientists'_New_Discovery_Is   
   Worthy_of_a_Toast Story Source: Materials provided by   
   Harvard_Medical_School. Original written by Ekaterina Pesheva. Note:   
   Content may be edited for style and length.   
      
      
   ==========================================================================   
   Journal Reference:   
      1. MacLean P. Nasrallah, Junhan Zhao, Cheng Che Tsai, David Meredith,   
      Eliana   
         Marostica, Keith L. Ligon, Jeffrey A. Golden, Kun-Hsing Yu. Machine   
         learning for cryosection pathology predicts the 2021 WHO   
         classification of glioma. Med, 2023; DOI: 10.1016/j.medj.2023.06.002   
   ==========================================================================   
      
   Link to news story:   
   https://www.sciencedaily.com/releases/2023/07/230707111646.htm   
      
   --- up 1 year, 18 weeks, 4 days, 10 hours, 50 minutes   
    * Origin: -=> Castle Rock BBS <=- Now Husky HPT Powered! (1:317/3)   
   SEEN-BY: 15/0 106/201 114/705 123/120 153/7715 218/700 226/30 227/114   
   SEEN-BY: 229/110 112 113 307 317 400 426 428 470 664 700 291/111 292/854   
   SEEN-BY: 298/25 305/3 317/3 320/219 396/45 5075/35   
   PATH: 317/3 229/426