Artificial Intelligence Helps to Predict Recurrence and Mortality for Prostate Cancer using Histology Images | bioRxiv


Artificial Intelligence Helps to Predict Recurrence and Mortality for Prostate Cancer using Histology Images | bioRxiv

Besides grading, deep learning could improve expert consensus to predict prostate cancer (PCa) recurrence. We developed a novel PCa recurrence prediction system based on artificial intelligence (AI). We validated it using multi-institutional and international datasets comprising 2,647 PCa patients with at least a 10-year follow-up. Survival analyses were performed and goodness-of-fit of multivariate models was evaluated using partial likelihood ratio tests, Akaike's test, or Bayesian information criteria to determine the superiority of our system over existing grading systems. Comprehensive survival analyses demonstrated the effectiveness of our AI-system in categorizing PCa into four distinct risk groups. The system was independent and superior to the existing five grade groups for malignancies. A high consensus level was observed among five blinded genitourinary pathology experts in ranking images according to our prediction system. Therefore, AI may help develop an accurate and clinically interpretable PCa recurrence prediction system, facilitating informed decision-making for PCa patients. 

This paper describes the development and validation of a novel #AI-based system for predicting #biochemicalrecurrence and #cancer-specific #mortality in #prostatecancer patients using #histology images. The key findings are:

The authors developed a #deeplearning model trained on histology images from prostate cancer samples to predict 10-year biochemical recurrence probability. The model was trained without using Gleason patterns to avoid biasing it towards reproducing the existing Gleason grading system.


The authors used histology image data from prostate cancer samples obtained from multiple sources:

The development cohort consisted of 600 #radicalprostatectomy (RP) cases from the #Canadian Prostate Cancer #Biomarker Network (#CPCBN) obtained from 2 institutions.

1) The first external validation cohort was 889 RP cases from the CPCBN framework from 3 different institutions.

2) The second external validation cohort was 989 cases from the #PROCURE prostate cancer #biobank, using tissue microarrays.

3) The third validation cohort was 1,502 whole slide images from 861 RP cases from the Prostate, Lung, Colorectal, and Ovarian (#PLCO) Cancer Screening Trial.

So in summary, the histology image data was obtained from multiple institutional cohorts and trials involving radical prostatectomy specimens and tissue microarrays. The samples spanned different institutions in Canada and the United States. Each cohort underwent expert review and quality control to ensure representative sampling and reliable clinicopathological data.

The model was validated on multiple independent cohorts totaling over 2500 patients. It stratified patients into four distinct risk groups with clear separation of biochemical recurrence-free and cancer-specific survival probabilities.

The risk groups were independent of and superior to the standard #Gleason grade groups in predicting outcomes, based on survival analyses and model performance metrics.

Pathologists blinded to outcomes were able to accurately rank histology image clusters according to the model's risk groups, indicating the AI-based system produces interpretable results consistent with human assessment.

The four risk groups showed a histopathological gradient from well-formed glands (lowest risk) to poorly-formed/absent glands (highest risk), providing further biological plausibility.

In summary, the authors have developed a novel AI-based prostate cancer grading system that more accurately predicts outcomes compared to standard pathology methods. The risk groups are calibrated and interpretable, addressing key limitations of black-box AI models. This could eventually allow better personalized prognosis and treatment decisions.

Comments

Post a Comment

Popular posts from this blog

A 10-Second Steam Blast: The New Weapon Against Prostate Cancer?

Image
Vanquish® Water Vapor Ablation System, Francis Medical Steams Out Prostate Cancer The Vanquish® Water Vapor Ablation System, developed by Francis Medical, Inc. , represents an innovative approach to treating prostate cancer using water vapor (steam) to ablate cancerous tissue. This minimally invasive technique aims to target and destroy prostate cancer cells while minimizing damage to surrounding healthy tissue, potentially reducing common side effects associated with traditional treatments like surgery and radiation, Mechanism of Action:  The Vanquish system utilizes the thermal energy stored in sterile water vapor to ablate cancerous prostate tissue. During the procedure, a catheter is inserted through the urethra into the prostate, delivering water vapor directly to the targeted tissue. As the vapor condenses back into liquid, it releases stored thermal energy, effectively destroying the cancerous cells. This method is designed to respect the prostate's natural boundaries, ...
Read more

Researchers develop low-cost device that detects cancer in an hour | ScienceDaily

Image
Summary Sanjay S. Timilsina , currently Lead Scientist, StataDx, a spinoff from Wyss Institute, former Postdoctoral Fellow at Harvard Medical School and the Wyss Institute at Harvard University, PhD from UTEP Department of Chemistry & Biochemistry, and XiuJun Li full Professor in the Department of Chemistry and Biochemistry and of Forensic Science, & Environmental Science and Engineering, at the University of Texas at El Paso (UTEP) have developed a new portable device for cancer detection that offers several advantages over traditional methods: Key Features: - Low cost (a few dollars per device, <8 cents per assay) - Provides results in about 1 hour (vs. 12-16 hours for traditional methods) - Approximately 10 times more sensitive than traditional methods - Portable and doesn't require specialized instruments - Can detect both CEA colorectal and PSA prostate cancer biomarkers - Uses a "paper-in-polymer-pond" structure to analyze blood samples How it works: - U...
Read more

Cancer patients and doctors team up to change how cancer drugs are tested | Fox News

Cancer patients and doctors team up to change how cancer drugs are tested | Fox News foxnews.com Associated Press For cancer patients, the harsh side effects of powerful drugs have long been the trade-off for living longer. Now, patients and doctors are questioning whether all that suffering is necessary. They’ve ignited a movement to radically change how new cancer drugs are tested, with the U.S. Food and Drug Administration urging drugmakers to do a better job at finding the lowest effective dose, even if it takes more time. Advances in treatment mean millions of people are surviving for years with incurable cancers. Jill Feldman, 54, of Deerfield, Illinois, has lived 15 years with lung cancer , thanks to that progress. Her parents both died of lung cancer months after their diagnoses. But her cancer drug causes joint pain, fatigue and mouth sores that m...
Read more