An interview with Dr. Rajan Gupta, researcher and radiologist
As the medical community prepares for the Radiological Society of North America (RSNA), the world’s largest stage for breakthroughs in medical imaging, prostate cancer continues to be a major focus of innovation. In order to explore how these advances are shaping the future of prostate cancer treatment, we spoke with Dr. Rajan Gupta, Professor of Radiology at Duke University School of Medicine.
Personalized Prostate Cancer Risk Stratification
We are approaching RSNA, the global stage for advances in medical imaging. How far are we from achieving truly personalized image-based risk stratification for prostate cancer?
We are in a very exciting phase for medical imaging. With ongoing improvements in prostate MRI acquisition techniques and image analysis, we are steadily moving closer to personalized, image-based risk stratification for prostate cancer.
However, achieving this vision requires one essential ingredient: standardization. Consistent image interpretation across readers and centers remains critical. Without that consistency, it becomes much harder to correctly assess the likelihood of clinically significant prostate cancer and to confidently guide patients toward the most appropriate treatment options.
So while the progress is real and encouraging, standardization is the foundation that will allow personalized imaging to reach its full potential.
Radiomics and Imaging Biomarkers: Unlocking What the Eye Can’t See
Quantitative imaging biomarkers, and particularly radiomics, are gaining traction in prostate MRI. How do you see their potential to support clinical decision-making in the coming years?
Prostate MRI has always relied on the combination of high-quality imaging and expert interpretation to evaluate the risk of clinically significant cancer. Radiologists use both qualitative impressions and traditional quantitative metrics when assessing a prostate MRI.
Imaging biomarkers and radiomics are now opening a new frontier. These tools allow us to extract information from MRI data that goes beyond what the human eye can perceive—patterns in lesion texture, gland heterogeneity, and other micro-features invisible in standard review.
If validated and integrated effectively, these advanced metrics could deepen our understanding of tumor behavior, offering clearer insights into cancer type and grade. Ultimately, radiomics may play a central role in supporting more informed and individualized treatment decisions.
AI & Collaboration Across Specialties
Artificial intelligence is becoming increasingly present in clinical workflows. How do you think it can help radiologists collaborate more effectively with urologists and pathologists in the prostate cancer diagnosis pathway?
Liquid biopsy, particularly ctDNA analysis, is now central to detecting minimal residual disease (MRD) and treatment resistance in gastrointestinal cancers.
There is an extraordinary amount of buzz around how artificial intelligence is reshaping the way radiologists interpret medical images. But the true value of AI ultimately depends on what each radiologist needs from it. For some, AI can streamline workflows and improve efficiency; for others, it becomes a reliable “second reader” that enhances diagnostic confidence.
What AI consistently brings to the table is more structured, quantitative information. And with richer data—whether for prostate MRI, lesion characterization, or follow-up studies—radiologists are better equipped to deliver precise, actionable insights to clinical partners in Urology and Pathology.
In the end, AI isn’t here to replace radiologists; it’s here to expand what’s possible in imaging, improve collaboration, and strengthen the quality of patient care.
Prostate Cancer Screening & bpMRI
Screening for prostate cancer is a hot topic. How far are we from making it a reality, and how will AI tools and reflex tests drive its success?
Prostate cancer screening with imaging can mean different things depending on who you ask. Unlike mammography—performed broadly across an entire population of asymptomatic women—prostate cancer screening relies on a preliminary test such as PSA. This initial step is essential, as it helps determine who truly benefits from further imaging, including prostate MRI.
The growing adoption of biparametric MRI is an exciting development, especially as the field moves toward more scalable, population-based prostate cancer screening for patients with elevated PSA. Equally promising is the role of AI in prostate MRI interpretation. Modern AI models can analyze biparametric MRI data to support radiologists with faster, more consistent image assessment.
This combination—quick acquisition with biparametric MRI and advanced interpretation powered by AI—creates a powerful synergy with the potential to improve prostate cancer screening on a global scale.
AI for Clinically Significant vs. Indolent Disease
How can AI improve the accuracy and efficiency of prostate MRI interpretation, especially when distinguishing between clinically significant and indolent diseases?
Artificial intelligence is an incredibly powerful tool in medical imaging, but its impact often depends on how each radiologist chooses to use it. Experienced readers may leverage AI to streamline workflows, speed up interpretation, and focus on complex cases, while less experienced readers might rely on AI as a “second reader.” In this role, AI not only validates findings they have identified but can also highlight subtle lesions or patterns they might have missed, increasing diagnostic confidence.
One of AI’s most significant advantages is its ability to analyze features beyond traditional quantitative metrics, providing insights that go beyond what the human eye can detect. As AI becomes more integrated into prostate MRI workflows, it promises to enhance accuracy, efficiency, and overall diagnostic quality—helping radiologists deliver more precise, actionable insights for patient care.
Fusion Biopsy Today: Strengths, Limitations, and Technological Progress
Fusion biopsy has become key to the specific diagnosis of prostate cancer. From a technical standpoint, what are the current limitations and advantages?
Fusion biopsy has become an essential tool in prostate cancer diagnosis, as it combines high-quality MRI data with real-time B-mode ultrasound used during the biopsy procedure. This integration significantly improves the detection of clinically significant prostate cancers while reducing the overdetection of indolent or clinically insignificant tumors, making diagnoses more precise and targeted.
From a technical perspective, there are still some limitations—particularly in accurately registering MRI images with ultrasound data. However, ongoing advancements in segmentation and registration algorithms are steadily improving the process. As these technologies continue to evolve, fusion biopsy is expected to become even more accurate and reliable, enhancing both the detection and characterization of prostate cancer for patients worldwide.
What’s Next for Prostate MRI? AI Integration and Global Standardization
Looking ahead, what advances in prostate MRI research do you think will transform patient care?
We are at an incredible point with prostate MRI today. Over the past three decades, we have made enormous strides in imaging equipment, MRI sequences, and reporting systems. Now, we are at a major inflection point in the field, as we work to determine how best to use AI to optimize image acquisition—both in terms of speed and image quality—and how to integrate AI effectively into image interpretation workflows.
My sincere hope is that AI will help standardize prostate MRI imaging and interpretation, so that patients around the world can benefit from consistent, high-quality diagnostic imaging. Variability in interpretation continues to be a challenge worldwide, and leveraging AI could be a key step toward more uniform and reliable prostate cancer diagnosis.