AI-Based Model May Guide Decision-Making in Kidney Cancer Based on CT, EGFR

Video

Data demonstrate the feasibility of automated glomerular filtration rate prediction to decide between partial nephrectomy and radical nephrectomy in kidney cancer, according to an expert from the Cleveland Clinic.

An artificial intelligence (AI)–based model was able to successfully predict post-operative glomerular filtration rates (GFR) in patients with kidney cancer in ways that clinicians may not be able to; it may also be easily implemented into the treatment landscape, according to Nour Abdallah, MD.

At the 2023 Genitourinary Cancers Symposium, CancerNetwork® spoke with Abdallah, a postdoctoral research fellow at Glickman Urological & Kidney Institute of Cleveland Clinic, about how AI-GFR prediction models calculated based on pre-operative CT scan and EGFR rate may impact the way patients are treated.

In study presented in a poster session at the symposium, Abdallah and her colleagues compared the AI and clinical model GFR estimations vs measured postoperative GFR based on correlation coefficients. When comparing the measured postoperative GFR, the correlation coefficient was 0.75 for the AI model and 0.77 for the clinical model, respectively. Additionally, the AI and clinical models performed similarly when predicting postoperative GFR below 45 ml/min/1.73m2 (Area under the curve, 0.89 and 0.9, respectively).

Based on the data, Abdallah stated that oncologists could potentially use these AI models for treatment decision-making without the need for additional clinical details or clinician involvement.

Transcript:

This model that we developed was able to behave as accurately as clinical models that were previously validated, whether to predict the post-operative EGFR in value or to predict the cutoff value which was 45.

We feel that this AI model would benefit the patients in a way that does not need any clinician time. It doesn't need any clinical details. It just needs the pre-operative EGFR value and the CT scan. We feel that it can be easily implemented in the health care system at the point of care to help the physician guide the patient towards the best surgery type for [them].

Reference

Abdallah N, Benidir T, Heller N, et al. Accuracy of fully automated, AI-generated models compared with validated clinical model to predict post-operative glomerular filtration rate after renal surgery. J Clin Oncol. 2023;41(suppl 6):693. doi:10.1200/JCO.2023.41.6_suppl.693

Recent Videos
Increasing patient awareness of modifiable risk factors for pancreatic cancer may help mitigate incidence of pancreatic cancers.
It may be crucial to test every patient for markers such as BRAF V600E mutations, NRG1 fusions, and KRAS G12C mutations to help manage pancreatic cancers.
Tanios S. Bekaii-Saab, MD, emphasizes the idea of moving targeted therapies to earlier lines of treatment to further improve outcomes in pancreatic cancer.
The Together for Supportive Cancer Care coalition may advance the national conversation in ensuring comprehensive care for all patients with cancer.
Health care organizations have come together to form the Together for Supportive Cancer Care coalition to address gaps in supportive cancer care services.
Further optimizing a PROTAC that targets MDM2 may lead to human clinical trials among patients with cancer harboring p53 mutations.
As patients are nearing the end of life, different management strategies, such as opioids, may be needed to help mitigate pain or fatigue.
Kelley A. Rone, DNP, RN, AGNP-c, highlights the importance of having end-of-life discussions early in a patient’s cancer treatment course.
Patients treated with BCMA-directed immunotherapies for myeloma may experience susceptibility to severe infections following treatment.
Related Content