Published Mar 2, 2018

ML at Sloan Kettering Cancer Center

Explore how Memorial Sloan Kettering Cancer Center leverages machine learning to classify genetic mutations, enhance patient care, and streamline clinical trial matching, as experts Alex Grigorenko and Iker Huerga delve into the transformative power of data-driven tools in oncology.
Episode Highlights
Data Skeptic logo

Popular Clips

Episode Highlights

  • Trial Matching

    The trial matching process at Memorial Sloan Kettering Cancer Center is a sophisticated system designed to connect patients with suitable clinical trials. explains that patients often arrive with complex cases, prompting the need for clinical trials as potential treatment options. The process involves a shared decision-making approach, where clinicians present standard therapies alongside trial opportunities, ensuring patients are informed of both risks and benefits 1. Grigorenko highlights the cognitive overload faced by oncologists, who must be aware of over 1,000 ongoing trials. This challenge is compounded by the unstructured nature of trial criteria, often written in PDFs, making it difficult to match patients effectively 2.

    Clinical trials are basically experiments that we do on people when we're trying to understand if one sets of interventions or treatments is going to be better for them in the long run than another.

    ---

    To address these issues, the center employs machine learning algorithms to recommend trials based on patient history and medical records, streamlining the matching process.

       

    Data-Driven Care

    Data-driven oncology at Sloan Kettering leverages patient data to enhance treatment outcomes and reduce cognitive load for physicians. emphasizes the importance of presenting accurate recommendations to oncologists, ensuring they can trust the system's suggestions without being overwhelmed 3. The collaboration between oncologists and data scientists is crucial, as notes the iterative process of developing models based on manually labeled datasets. This partnership aims to achieve high accuracy and precision in identifying patient populations for specific trials 4.

    We are at the cutting edge of cancer research. So for some of these oncologists, those are questions that have been in the cancer community for a long time that nobody has been able to answer.

    ---

    Ultimately, the goal is to provide patients with confidence in their treatment plans, knowing that decisions are informed by comprehensive data analysis.

Related Episodes