Many patients with cancer are not sent for genetic screening, largely because patient records often have incorrect or incomplete ICD codes. Patients are therefore missed for screening, therapies, and clinical trials when clinicians search based on ICD codes. Results of a study that examined whether the combination of artificial intelligence (AI) and clinical knowledge to mimic a clinician’s manual review process can identify more patients with lung and ovarian cancers were presented at the ESMO Targeted Anticancer Therapies Congress 2024.
The researchers configured an AI-driven product to find patients with lung and ovarian cancers who may have been miscoded with the incorrect ICD code, applying the product to two retrospective datasets, based on the clinical guidelines that clinicians use when manually reviewing patient records. One dataset consisted of 99 patients to identify those who were diagnosed with lung cancer, and the second consisted of 95 patients to find those diagnosed with ovarian cancer. Clinicians manually labeled the patient records in these two datasets for lung, ovarian, or no cancer recorded. Miscoded patients were those who were determined by manual review to have lung or ovarian cancer but lacked the correct ICD codes.
Results showed the AI product achieved 85% precision and 95% sensitivity by finding 63 patients with lung cancer; ICD codes achieved 96% precision and 73% sensitivity by finding 48 patients, translating to the AI product identifying 31% more patients with lung cancer who were miscoded. Additionally, the AI product achieved 87% precision and 100% sensitivity by finding 52 patients with ovarian cancer; ICD codes achieved 97% precision and 69% sensitivity by finding 36 patients with ovarian cancer. The AI product identified 44% more patients with ovarian cancer who were miscoded.
The researchers concluded that ICD code-based searches were able to find patients with high precision, but several patients were missed due to miscoding. “This study showed that combining clinical knowledge and AI can help mimic a clinician’s manual review and find more patients who would otherwise be missed,” they said. “The discovery of such patients allows them to be screened for lung and ovarian cancers and taken through the appropriate treatment involving therapy or a clinical trial.”
Source: Gupta V, Kayayan EN, Zhang J, Falakaflaki P, Gupta G. Using AI and clinical knowledge to find missed lung and ovarian cancer patients. Abstract 5P. Presented at the ESMO Targeted Anticancer Therapies Congress 2024; February 26-28, 2024; Paris, France.
© 2025 Mashup Media, LLC, a Formedics Property. All Rights Reserved.