The future of nuclear medicine: Artificial intelligence, theranostics and beyond

While technological advances such as full-body imaging are revolutionizing the diagnostic capabilities of positron emission tomography (PET) imaging systems, theranostics (the ability to find and attack a tumor) demonstrates that the field is indeed nuclear medicine — not merely diagnostics. It’s also become an essential component of precision medicine — especially precision psychiatry.

And while as-yet-unrealized, artificial intelligence in nuclear medicine is poised to take on the more tedious routine tasks, allowing nuclear medicine physicians to focus on more complex activities. Moreover, research continues in all these areas, promising even more innovation over the next few years.

Nuclear medicine: a brief primer

Nuclear medicine involves using radiopharmaceuticals (aka radioactive tracers or radioisotopes) to diagnose and treat disease. In diagnostic studies, a clinician injects the radiopharmaceutical into the patient’s bloodstream. An imaging tool — often a PET scan — tracks its path.

The extent to which the tissue or organ absorbs the radiopharmaceutical can indicate the level of function of the organ or tissue being studied. This allows the physician to better assess and diagnose different conditions.¹

But nuclear medicine isn’t limited to just diagnostics.

Theranostics: Combining treatment with diagnostics

In nuclear medicine, theranostics — a portmanteau of therapy and diagnostics — refers to deploying radiopharmaceuticals that can find a patient’s tumor and attack the tumor site. This pairing can help customize disease management and predict treatment response and toxicity.²

The field is advancing rapidly, but it’s not new. In fact, nuclear theranostics is almost 80 years old. It began with using radioactive iodine (Iodine-131) for the diagnosis and treatment of thyroid cancer.³ It’s come a long way since then.

In particular, the establishment of theranostics in neuroendocrine tumors and prostate cancer represents one of the most significant emerging trends in nuclear medicine, says Johannes Czernin, MD, professor and vice chair of the Department of Molecular and Medical Pharmacology, chief of the Ahmanson Translational Theranostics Division and director of the JCCC Cancer Molecular Imaging, Nanotechnology and Theranostics Program at the University of California in Los Angeles.

The National Comprehensive Cancer Network (NCCN) recently included prostate membrane specific antigen (PSMA) PET imaging in its guidelines for managing prostate cancer.⁴ Czernin calls this “a major game changer” that could support improved patient outcomes.

In addition to allowing physicians to provide more precise therapies, improved cancer staging also gives them insight into when a patient requires palliative care.⁵

Czernin also points to the recent VISION trial. This study showed that targeting PSMA therapeutically prolongs the lives of late-stage prostate cancer patients.⁶ He expects many patients to benefit from treatment, noting that “many studies are now underway to determine whether earlier PSMA-targeted treatments can lead to even better results.”

Overall, advances in theranostics have significantly increased the number of diseases that nuclear medicine can treat, Czernin explains. In addition to thyroid and prostate cancer, therapies are available for paraganglioma/pheochromocytoma and neuroendocrine tumors. Innovation continues, he adds, with researchers exploring several new targets.

New diagnostic directions: Precision psychiatry & neurodegeneration

Nuclear theranostics has become an essential component of precision medicine.²

However, until recently, advances in precision medicine largely left psychiatry behind. That’s changing with the emergence of precision psychiatry.⁷ Recent studies have explored PET as a tool for identifying subtypes of psychosis.⁸ This could guide subtype-specific therapeutic interventions — in other words, finding the appropriate treatment for the specific condition. This may, in turn, support better patient outcomes.⁹

Nuclear medicine may also lead to earlier diagnoses of neurodegenerative disorders such as Parkinson’s and Alzheimer’s disease. Many experts believe brain PET tracers will eventually visualize most, if not all, of these conditions. This could lead to diagnosis before any symptoms appear.¹⁰ Earlier diagnosis allows for earlier treatment, when available. It also gives the patient more opportunities to participate in clinical trials, many of which seek pre-symptomatic subjects. This gives them access to experimental medications not yet on the market.

Full-body PET imaging: A major breakthrough

PET technology innovation is transforming the diagnostic capabilities of nuclear medicine. According to Czernin, the most important emerging technologies are in PET instrumentation–full-body PET imaging in particular. “These systems have a 40-fold improved sensitivity, allowing physicians to detect cancer earlier and stage it more accurately.”

Full-body PET imaging provides a molecular and anatomical view, giving clinicians a broader perspective on the patient’s condition. Among the other benefits, this approach:^9,11

• Allows clinicians to see the whole body at once
• Provides higher-quality diagnostic PET scans than ever before
• Scans up to 40 times faster than current PET with radiation doses up to 40 times lower — this is especially beneficial for repeat scans and pediatric use
• Performs a whole-body scan in 20 to 30 seconds
• Improves image quality, allowing physicians to see smaller tumors and other diseases earlier
• Offers the ability to evaluate activity in all organs and tissues simultaneously

Czernin and co-authors of a 2020 Journal of Nuclear Medicine article on the future of medicine predicted: “Bringing devices with a longer axial field of view to research and the clinic will be an important component of future work.”

AI in nuclear medicine: Yet to be realized

Despite the buzz around artificial intelligence and machine learning, their true clinical value to nuclear medicine has yet to be realized, Czernin says. AI for diagnostic nuclear medicine is still in its infancy and has yet to deliver on its promises for improvements in disease diagnosis. So broad clinical application isn’t in the immediate future.

Listening to the way some people talk about the future of artificial intelligence in medicine, you’d almost think it was magic. The irony, of course, is that AI’s greatest potential lies in managing the quotidian, not the sophisticated. Soon, AI will allow for the automation of labor-intensive but cognitively undemanding tasks.¹²

These cumbersome and time-consuming tasks will likely be fully automated thanks to “trained” AI algorithms (aka machine learning).¹³ Machine-learning applications evaluate data in real-time and change their behavior accordingly. “I can foresee improvements in workflows including image acquisition and image reconstruction,” Czernin says.

This means humans will spend more time doing what humans do best: activities that require insight, intellect and imagination. The time they gain would allow nuclear physicians to provide comprehensive and integrative expertise for overall patient management.

If AI does become as competent as humans in interpreting images, imaging experts will become more like consultants, Czernin says. Eventually, AI could detect disease and make predictions and classifications based on molecular imaging features.¹³ For example, according to Czernin, AI may become helpful in identifying patients who would be the best candidates for theranostic approaches such as PSMA-targeted therapy.

Don’t expect AI to take over nuclear medicine; its potential lies in supporting physicians in doing their best work.

The future of nuclear medicine: The profession & the science

Nuclear medicine trends demonstrate it is becoming an integral part of patient care. It will, according to Czernin, “stand on its own as a highly specialized discipline that is well integrated in overall patient care.”

That’s a broad prediction and for good reason. In that 2020 future of nuclear medicine article, he and his co-authors warn that it has been “notoriously difficult to predict.”⁹

Not that unpredictability is a bad thing. They write: “The development of nuclear medicine has been unpredictable because the field has been so innovative and able to quickly adapt to new scientific developments and clinical needs. If this innovation continues, there is no question that the future of nuclear medicine is bright.”⁹ 

Be advised that information contained herein is intended to serve as a useful reference for informational purposes only and is not complete clinical information. This information is intended for use only by competent healthcare professionals exercising judgment in providing care. McKesson cannot be held responsible for the continued currency of or for any errors or omissions in the information.