According to Arturo Loaiza-Bonilla, MD, MSEd, FACP, trends in complementing radiology practices with artificial intelligence (AI) include mammography, early screening for other diseases, and the integration of AI systems into RECIST standards. Loaiza-Bonilla discussed these topics in a conversation with CancerNetwork®.
First, she suggested that traditional mammography can take a lot of time and effort, and that AI could be used to expedite this process while reducing the need for frequent double screenings. Professor Loaiza Bonilla explained that after data was collected in Europe, a trial was recently launched in the United States to assess the demographics of patients at risk of breast cancer in their home country.
He also highlighted ongoing research investigating early lung cancer detection through screening, in addition to X-rays used in emergency departments. He further highlighted Friends of Cancer Research’s efforts to integrate the use of AI systems into RECIST guidelines to embed algorithms to assess tumor burden and characterize disease in a scalable manner.1
Finally, Loaiza-Bonilla discussed the application of AI to predictive models, such as the predictive model within the PANORAMA trial, to explore the role of AI in supporting early disease detection through CT scans in pancreatic cancer. He concluded by citing the very high risks of cancer as a disease and hoping that these models will begin to be implemented in real practice.
Loaiza-Bonilla is the overall director of the Hematology-Oncology Department at St. Luke’s University.
Transcription:
Currently, radiologists, especially in oncology, are primarily focused on two major trends. Mammography is one of them. [we] There are many patients who need to be screened and tested, so try to optimize your time. [it’s a] a considerable amount [time and] work. We are currently using AI to reduce double mammogram screening, and it is also being implemented in clinical trials. We have data from Europe, but there’s also a lot of work being done in the U.S. right now because the population is different in terms of things like breast density. I’m looking forward to seeing the results when these algorithms are implemented.
other [trends] what we are looking for [at] We work closely on finding lung nodules and detecting them in real time, allowing screening to detect lung cancer early, even from X-rays found in the emergency department. Some efforts are being made [that]. The use of RECIST criteria in clinical trials has progressed further. The Friends of Cancer Research initiative is piloting the ai.RECIST standard, which allows researchers to measure lesions in a more effective and scalable way by simply embedding an algorithm. Radiologists play an important role in this.
lastly, [AI is being used in] some predictive models. There are several efforts showing that CT scans can detect pancreatic cancer early, such as the results of the PANORAMA trial.2 This is one of those diseases that you want to catch early, before it actually metastasizes, because most are locally advanced/resectable or metastatic. It is very promising in terms of radiology.
What I want is to actually start using these models. Currently, many of the models we use in radiology are limited to these use cases, such as fracture detection, and emergency departments, but there is an opportunity here because oncology is one of the most dangerous diseases.
References
- Ai.RECIST Project: Artificial Intelligence-Enabled Response Evaluation Criteria in Solid Tumor Projects. A friend of cancer research. Accessed January 28, 2026. https://tinyurl.com/rx26459z
- Alves N, Schuurmans M, Rutkowski D, et al. Artificial intelligence and radiologists in the detection of pancreatic cancer using standard-of-care CT scans (PANORAMA): An international paired non-inferiority confirmatory observational study. Lancet Oncor. 2026;21(1):116-124. doi:10.1016/S1470-2045(25)00567-4
