Detect Cancer Years Earlier: The Science of Proactive Full-Body MRI
Medical imaging stands at a turning point comparable to the invention of the telescope: AI-powered longitudinal full-body MRI can now detect cancer and other diseases years before symptoms emerge, while slashing false-positive rates from 64% to under 10% by adding context. Dr. Daniel Sodickson argues that everyone should establish a baseline scan in their 20s and monitor biology continuously — just as we already track blood biomarkers over time — rather than waiting for symptoms to force a diagnosis. Combined with genetics, proteomics, and AI, this approach forms what Sodickson and Dr. Hyman call 'medical intelligence': a GPS-like system guiding individuals toward 100 healthy years.
Overview
Dr. Mark Hyman hosts Dr. Daniel K. Sodickson — physicist, radiologist, and Chief Medical Scientist at Function Health — to explore why medical imaging has historically been reserved for symptomatic patients and why that approach is now obsolete. Sodickson explains that the core problem with proactive imaging is not the technology itself but the lack of context: interpreting a single snapshot produces high false-positive rates, whereas AI models fed sequential scans, blood tests, and clinical history reduced prostate-cancer false positives from 64% to under 10% in his NYU lab research.
He introduces the principle of 'visualization first rather than last,' calling for baseline full-body MRI scans starting in a person's 20s, then repeated every one to two years to track change rather than hunt for disease. A critical insight is that more imaging paradoxically enables cheaper imaging: once a baseline exists, AI neural networks can reconstruct high-quality follow-up images from 20–30 times less raw data, paving the way for low-cost 'everywhere scanners' embedded in chairs, beds, and wearables.
Sodickson distinguishes imaging from blood tests as 'spatial context' versus 'biochemical context,' arguing that combining both provides a complete structure-and-function picture that neither modality can deliver alone. Real-world scans through Function Health's imaging partner have already caught prostate, brain, and kidney cancers at pre-symptomatic stages, enabling swift treatment with far better outcomes. The hosts also discuss liquid biopsies and proteomics — cancer-associated proteins detectable in blood years before clinical diagnosis — as a complementary layer that AI can interpret at a population scale no individual physician could match.
Sodickson frames the ultimate goal as a 'GPS for health': a platform integrating all of a person's biological data and projecting trajectories forward, triggering a quiet alert only when a course correction is needed rather than generating constant alarm. The conversation closes with a shared call for a fundamental paradigm shift away from episodic, symptom-driven care toward continuous, proactive, data-informed health management.
Key quotes
5We already have a cure for cancer. It's early detection.
In a medical sense, more information helps you make better decisions.
I really think of it almost like building ourselves a new augmented artificial sensory system.
Your body is not designed to be sick. It's not a design flaw.
The future of seeing involves actually looking.
Key ideas
9From reactive to proactive imaging
Traditional medicine deploys imaging only after symptoms appear, which means disease is often found at an advanced, harder-to-treat stage. Sodickson argues for deploying imaging as a first-line preventive tool rather than an end-stage diagnostic one.
Context is the cure for false positives
A single MRI snapshot carries a high false-positive rate because incidental findings have no reference point. Knowing whether a lesion is new or long-standing — via prior scans, blood tests, and clinical history — transforms ambiguous signals into actionable intelligence.
AI slashes prostate-cancer false positives from 64% to under 10%
Sodickson's NYU lab trained an AI model on sequential MRIs plus blood and clinical data. Adding longitudinal context reduced the false-positive rate by an order of magnitude, proving that multi-modal, time-series data dramatically improves diagnostic precision.
More imaging makes future scans 20–30× cheaper and faster
Once a baseline exists, AI neural networks can reconstruct high-quality follow-up images from 20 to 30 times less raw data than a traditional scan requires. This means follow-up scans become far cheaper and feasible with low-power, inexpensive hardware.
The Everywhere Scanner: imaging moves into chairs, beds, and wearables
With a baseline established, simple low-power scanners embedded in everyday furniture or worn on the body could detect biological changes continuously. This vision reframes health monitoring as ambient and ongoing rather than episodic and institutional.
Imaging provides spatial context; blood tests provide biochemical context
Blood biomarkers reveal what the body is producing chemically; imaging shows where things are spatially. Combined, they offer a complete structure-and-function picture that neither modality can provide alone — the gold standard for medical intelligence.
Early detection is effectively a cure for most cancers
Prostate, brain, and kidney cancers caught before symptoms allow swift intervention with dramatically better outcomes. Real-world scans have already saved lives by identifying tumors years before they would have become symptomatic.
Baseline scans in your 20s, repeated every one to two years
Sodickson recommends establishing a reference scan as early as the 20s because the goal is measuring change over time, not just finding existing disease. Frequency should increase with age — every two years when young, annually when older.
Medical intelligence as a GPS for health
Function Health's Medical Intelligence Lab aims to integrate imaging, blood biomarkers, genetics, wearables, and AI into a personalized roadmap — projecting biological trajectories forward and alerting users when a course correction is needed before disease takes hold.
Practical takeaways
7- 1
Get a baseline full-body MRI in your 20s 42:42
Even without symptoms, a reference scan establishes your personal normal — the essential foundation for all future comparisons and the starting point for detecting meaningful biological change.
- 2
Scan every two years when young, annually when older 44:56
Increasing scan frequency with age matches the rising likelihood of biological change. Even a second scan after the baseline dramatically reduces false-positive risk by establishing your personal trajectory.
- 3
Pair imaging with comprehensive blood biomarkers 34:09
Imaging and blood testing are complementary: one reveals spatial structure, the other biochemical function. Together they enable earlier, more accurate identification of problems that either modality would miss alone.
- 4
More longitudinal data reduces — not increases — medical anxiety 24:14
Counter-intuitively, accumulating more health data over time shrinks false-positive rates and gives clinicians and patients the context they need to avoid unnecessary follow-up procedures and needless worry.
- 5
Disease is slow and often reversible — act on early signals 58:30
Chronic conditions develop over decades, not overnight. Early shifts in fasting glucose, PSA, or brain structure are detectable long before diagnosis and are frequently reversible with targeted lifestyle and medical intervention.
- 6
Think of health monitoring as a background safety net, not a source of alarm 51:01
The goal of continuous biological sensing is to operate quietly in the background — like the body's own nervous system — and signal only when a meaningful course correction is required.
- 7
Put your biology online now 1:01:00
The combination of full-body MRI, comprehensive blood panels, genetics, and AI is already available. Getting started with a baseline scan and regular testing is the most actionable step toward 100 healthy years.
Topics & chapters
15Opening: how should we really think about imaging today?
Host and guest frame the central question: medical imaging is changing rapidly, and the key shift is from symptom-driven diagnosis to proactive, preventive monitoring.
The Future of Seeing: imaging has an image problem
Sodickson explains what prompted his book — a paradox in which we live more imaged lives than ever yet understand imaging less than ever, disconnecting people from the devices reshaping their health.
A brief history: from X-rays to MRI and tomography
Hyman traces imaging's evolution from early (and sometimes harmful) X-ray use to CAT scans, MRI, PET, and ultrasound, each extending human vision deeper into the body without a single cut.
Visualization first, not last: the paradigm shift
Sodickson describes his personal awakening: imaging was being used to chase symptoms, always arriving too late. He began asking whether these tools could be deployed proactively, before disease becomes advanced.
The false-positive obstacle and why it is not fixed
The main argument against proactive imaging — that it produces too many false positives — is real but not inevitable. False-positive rates are a function of how imaging is used, not a built-in property of the machines.
Longitudinal imaging surveillance: tracking change over time
A quiet paradigm of surveillance imaging already exists in oncology (e.g., annual prostate MRI in moderate-risk patients). Sodickson argues this should be extended to people of unknown risk, using context to interpret findings.
AI and big data: the computational solution
In the era of big data and AI, collating diverse longitudinal health data and detecting subtle patterns is computationally tractable in a way it never was before, changing the risk-benefit calculation for proactive imaging.
Cost deflation: full-body MRI drops from thousands to hundreds of dollars
Hyman notes that a knee MRI cost him $2,500, while full-body MRI is now available for $499–$999. Sodickson explains why the cost trajectory will continue downward, especially once baselines are established.
The Everywhere Scanner: ambient imaging in everyday life
With a baseline on file, AI can reconstruct quality images from 20–30× less data, enabling low-power scanners in chairs, beds, CVS pharmacies, or wearables — making imaging as continuous as the body's own nervous system.
Blood tests vs. imaging: biochemical and spatial context
Blood tests reveal biological chemistry; imaging reveals spatial anatomy. Sodickson describes their synergy as 'cooking with gas' — together they provide structure-and-function data that transforms clinical decision-making.
Real findings: early cancer, brain changes, and body composition
Prostate, brain, and kidney cancers have been caught pre-symptomatically through Function Health's imaging service. The hosts also discuss coronary artery calcium scoring, fatty liver, and brain structural changes predictive of neurodegeneration.
Who should scan, when to start, and how often
Sodickson recommends a baseline in the 20s, with follow-ups every two years when young and annually with age. Even the second scan substantially reduces false positives by establishing a personal trajectory.
An augmented artificial nervous system for early warning
The body's native sensory system warns of immediate harm but cannot detect internal disease early. The Everywhere Scanner vision builds a complementary artificial nervous system that provides exactly that missing early-warning layer.
Medical Intelligence Lab: a GPS for personal health
Function Health's Medical Intelligence Lab aims to integrate all biological data — imaging, labs, genetics, wearables — into a personalized GPS that projects health trajectories and recommends course corrections before disease becomes irreversible.
Closing: the future of seeing means actually looking
Both doctors close with an appeal to embrace proactive biology monitoring. Sodickson's final message: now that we have the capability to see deep into human biology and connect it with AI-powered knowledge, the future of seeing simply involves using that capability.