Why Nuclear Medicine

The short version

Nuclear medicine images the body's biology — metabolism, blood flow, receptor expression — not just its structure. And in one part of the field, theranostics, the same targeting biology used to make the picture is then used to treat: a ⁶⁸Ga-PSMA PET scan shows where prostate-cancer deposits are, and days later ¹⁷⁷Lu-PSMA delivers a radiation dose to those same deposits. Being able to image a target and then treat it with the same biology is a genuine advantage — the picture becomes the treatment plan — and it is a large part of why the field is growing so quickly right now.

Why this is the future, not a niche

For decades nuclear medicine was a quiet corner of imaging. That era is over. The field is in the middle of a therapeutic transformation: radiopharmaceutical therapy is projected to grow from roughly $750 million (2022) to ~$5.5 billion by 2028, with 80+ new radioligand therapies in clinical development and pharmaceutical companies investing billions. Prostate cancer (PSMA) and neuroendocrine tumors (somatostatin receptor) are only the beginning — the same "one target, two isotopes" logic is being aimed at a growing list of cancers.

This is not incremental. The Lancet Oncology Commission estimated that meeting global demand for PSMA radioligand therapy alone would require more than half a million treatment doses per year — and "hundreds of new centers and more experts to deliver these therapies." The bottleneck is no longer the science. It is the number of trained physicians who can run these programs.

Reading physiology and anatomy together

A nuclear medicine study is an overlay. The functional image (PET or SPECT) shows where a biological process is happening — where tissue is metabolizing glucose, expressing a receptor, or being perfused — and it is fused with a CT, which shows the anatomy. The two are read as one image. A CT alone tells you a lymph node is enlarged; the molecular layer tells you whether that node is actually disease. It is the combination — biology localized precisely onto structure — that lets nuclear medicine draw conclusions a structural scan alone cannot: an enlarged, PSMA-avid node is disease; the same node without uptake usually is not.

As oncology becomes precision oncology, imaging that reads the biology — and, in theranostics, treats it — is increasingly central to how patients are staged and managed.

The theranostic loop

The defining workflow is a loop: select patients by imaging the target (is the target there, and everywhere?), treat with the matched therapeutic isotope, quantify the delivered dose (dosimetry), and reassess to decide the next cycle. It is precision medicine made literal — the diagnostic and the therapeutic are the same molecule wearing a different isotope. Explore the Theranostics domain of this atlas to see the loop in practice, and the Landmark Trials chapter for the evidence that built it (VISION, NETTER-1/2, TheraP, ALSYMPCA).

The AI accelerant

If the therapy transformation is the field's engine, artificial intelligence is its accelerant — and nuclear medicine may be the imaging specialty AI changes most. The reason is structural: nuclear medicine data is quantitative and numeric (SUV, counts, kinetics), and the theranostic loop is a self-labeling data engine — every patient imaged, treated, dosed, and reassessed generates exactly the paired imaging-and-outcome data that machine learning is starved for elsewhere. The immediate payoffs are already here: deep-learning denoising that lets scans run at a fraction of the dose or time, automated whole-body tumor-volume measurement in seconds, and — the biggest unlock — automated dosimetry that turns fixed-activity, one-size-fits-all therapy into personalized, dose-optimized treatment. Far from threatening the specialty, AI is what lets a too-small workforce meet exploding demand. For a student, that is the tell: nuclear medicine is where molecular biology, quantitative computing, and targeted therapy converge — see the AI & Machine Learning page for the full case.

The opportunity

Because demand for radiopharmaceutical therapy is outrunning the trained workforce, the career market is unusually strong — academic centers are increasingly hiring nuclear medicine physicians directly out of residency for faculty roles, a pattern that is expanding as programs race to open therapy services. The shortage spans the whole team — physicians, technologists, radiochemists, medical physicists, and radiopharmacists — but the physician who can lead a theranostics program is among the scarcest and most sought-after. Entering the field now means entering it at the moment it is being built out.

For the student deciding

You do not have to choose blindly. The training pathways page lays out exactly how to become a nuclear medicine physician — including routes through diagnostic radiology, so exploring nuclear medicine does not mean abandoning radiology. Spend time in a reading room during a theranostics clinic, watch a PSMA PET drive a treatment decision, and see whether the see-and-treat loop lights the same bulb it has lit for a generation of physician-scientists now shaping the field. The atlas exists to make that exploration easy: every tracer, target, therapy, and trial, in one place.

Related pages

• Theranostics Overview and the Therapy & Theranostics domain — the see-and-treat loop in clinical practice.
• The Therapy Transformation — the field-level shift and the workforce it demands.
• Becoming a Nuclear Medicine Physician — the training pathways, including via radiology.
• AI & Machine Learning in Nuclear Medicine — why AI may transform this field most, and the dosimetry unlock.
• Landmark Trials — VISION, NETTER-1/2, TheraP, ALSYMPCA, PSMAfore: the evidence base.