The survival rate for cancer patients is influenced by a host of factors, including lifestyle and genetic circumstances. For example, components like smoking, prolonged exposure to chemical toxins, cancer type, and comorbidities are studied by oncologists to retrospectively predict individual mortality rates with varying consistency.
In February 2023, researcher John-Jose Nunez of the British Columbia Cancer and Department of Psychiatry reported that “an individual patient’s survival could be used to improve cancer care […] it might suggest earlier referral to palliative care resources or consideration of more aggressive therapies upfront.” The new predictive models fueled by artificial intelligence (AI) are claimed to have outperformed previous tools in determining cancer patient survivability.
Natural language processing (NLP) is a component of AI that allows computers to comprehend human vocabulary and written text. Often called machine learning, the NLP system can process extensive quantities of unstructured text data in a logical format. For example, lead author Nunez and his colleagues used NLP to “predict the survival of patients with general cancer” by examining medical notes collected during the patient’s first visit with an oncologist and no additional data.
I use telehealth services?
A. Aww! It's so convenient.
B. No, strictly in-person for me.
C. Tele-who?
The NLP model was 80% accurate in predicting the survivability of cancer patients at the 6, 36, and 60-month markers based upon the initial consultation documents. Primitive machine learning software was constrained by requiring copious amounts of unstructured supplementasry data that was not quickly processed. However, Nunez alleges that the newly designed NLP outperforms oncologists who routinely overlook impactful patient information. Clinical notes often contain “many details relevant to survival” that are not tracked by healthcare teams, like age and marital status.
The National Cancer Institute predicts that the number of cancer survivors is projected to grow to 26 million by 2040. Quality of life for individuals is improved when evidenced-based screening, prevention, and early interventions are implemented. NLP has transformed limited patient information into an 80% accurate predictability offering. The technique complements traditional oncology treatment by creating a multilayered degree of data analysis for a multifaceted therapeutic plan.
-- Read the study: https://jamanetwork.com/journals/jamanetworkopen/fullarticle/2801709
Welcome to the MSM Peanut Gallery! Please take responsibility for your behavior as you enter this shared space. Keep conversations courteous and on-topic to foster productive, respectful communication. Review our Community Guidelines before posting.
Comentários