Learning paths

Eight sequenced paths. Each lists who it is for, what you will be able to explain afterwards, and its lessons in order.

Radiogenomics Foundations

What it means to link an imaging phenotype to a genomic phenotype, and why public evidence is not clinical certainty.

For: Students, researchers, clinicians

You will be able to

Tell radiomics from radiogenomics
Describe semantic and quantitative imaging phenotypes
Explain why a published association is not a diagnostic test

Lessons

What radiogenomics studies introductory
Radiogenomics asks whether what we see on an image relates to what is happening in a tumour’s genes. The word also has a second, older meaning in radiation oncology.
Imaging phenotypes and genomic phenotypes introductory
An imaging phenotype is what a tumour looks like on a scan, described by a reader or measured by software. A genomic phenotype is its molecular make-up.
Radiomics versus radiogenomics introductory
Radiomics extracts many quantitative features from images. Radiogenomics relates imaging features, radiomic or semantic, to molecular data.
Example concepts: EGFR, KRAS, IDH, MGMT intermediate
A tour of genes that appear often in radiogenomics research, with the caution each one needs.
Public evidence versus clinical certainty intermediate
Why a significant p-value in one dataset is the start of a story, not the end, and how RadioGx shows that difference.

Nuclear Medicine Foundations

Radionuclides, radiopharmaceuticals, PET, SPECT and planar imaging, and the hybrid scanners that combine them.

For: Students, technologists, public

You will be able to

Explain the tracer principle
Describe positron, beta, alpha and gamma emissions
Compare PET, SPECT and planar imaging

Lessons

What nuclear medicine is introductory
Nuclear medicine uses small amounts of radioactive compounds to image how the body functions, and in some cases to treat disease from the inside.
Radionuclides: decay, half-life and production introductory
How radionuclides decay, what half-life means, and where medical radionuclides come from.
Diagnostic versus therapeutic use introductory
The physics that makes a radionuclide good for imaging is different from the physics that makes it good for therapy.
PET, SPECT and planar imaging intermediate
How PET uses coincidence detection, how SPECT uses collimators and rotation, and what planar imaging still does well.
Hybrid imaging and public safety context intermediate
PET/CT, SPECT/CT and PET/MR pair function with anatomy. Radiation protection principles keep exposures as low as reasonably achievable.

Radiopharmaceuticals & Theranostics

How a target, a ligand, a chelator and a radionuclide combine, and how theranostic pairs let the same target be imaged and treated.

For: Researchers, clinicians, regulatory

You will be able to

Name the parts of a radiopharmaceutical
Describe common theranostic pairs
Read a radiopharmaceutical record without treating it as advice

Lessons

Anatomy of a radiopharmaceutical introductory
Most targeted radiopharmaceuticals have four parts: a biological target, a vector that binds it, a chelator or prosthetic group, and a radionuclide.
Theranostic pairs: see it, then treat it introductory
Theranostics images a target with one radionuclide and treats it with another, using the same or a closely related molecule.
Targets in practice: SSTR, PSMA, NIS, FAP and more intermediate
A guided tour of the targets most often seen in nuclear medicine and theranostics research.
Alpha and beta therapy, and emerging radionuclides advanced
Why actinium-225, lead-212, terbium-161 and astatine-211 attract attention, and the practical questions they raise.
Reading a radiopharmaceutical record safely intermediate
Labels, trial records and papers name the same molecule in different ways. Here is how to read them without turning them into advice.

Imaging Biomarkers & Radiomics

What makes a measurement a biomarker, how SUV and reporting systems work, and the radiomics pipeline with its reproducibility caveats.

For: Researchers, physicists, data scientists

You will be able to

Define an imaging biomarker and its validation steps
List radiomic feature families
Recognise reproducibility and AI pitfalls

Lessons

What makes a measurement an imaging biomarker introductory
A biomarker is a defined characteristic measured as an indicator of a biological process, disease or response. Imaging biomarkers need technical and clinical validation.
SUV and PET quantification intermediate
The standardised uptake value normalises PET activity to the injected amount and body size. It is useful and easy to over-trust.
Structured reporting systems introductory
RECIST, Deauville, Krenning, PSMA-RADS, PI-RADS and others standardise how findings are described.
The radiomics pipeline intermediate
From image acquisition to a validated model: each step adds variability and each needs reporting.
Feature families: shape, first-order, texture, higher-order intermediate
What each family measures and why texture features need extra care.
Reproducibility caveats advanced
Test-retest scans, inter-scanner variation, segmentation variability and harmonisation, in plain language.
AI segmentation and deep learning: caution intermediate
AI models can segment and classify images, but performance depends on data, and research code is not a regulated medical device.

Dosimetry & Radiation Safety Context

Half-lives, absorbed and effective dose, the MIRD idea and radiation protection, explained as concepts with no patient calculation.

For: Physicists, technologists, students

You will be able to

Distinguish physical, biological and effective half-life
Explain absorbed versus effective dose
Know who regulates radiation safety and where to look

Lessons

Physical, biological and effective half-life introductory
Radioactivity in the body falls for two reasons at once: decay and elimination. Their combination is the effective half-life.
Absorbed dose, equivalent dose and effective dose intermediate
Three quantities with different jobs. Mixing them up is the most common error in public discussions of radiation.
The MIRD idea, without the calculator advanced
Internal dosimetry combines how much activity sits in each organ over time with how much of the emitted energy each target absorbs.
Organs at risk and administered activity as label concepts intermediate
Why kidneys, salivary glands and bone marrow come up in radioligand therapy research, and how to read activity statements in labels.
Radiation protection and who regulates it introductory
ALARA, time, distance and shielding, and the regulators and frameworks that set the rules.

Trials, Regulatory, Literature & Patents

How to read a trial record, a label, an approval or clearance, an adverse event count, a paper and a patent without overclaiming.

For: Regulatory, researchers, informed public

You will be able to

Read ClinicalTrials.gov records correctly
Separate approval from clearance
Treat adverse event reports and patents with the right caution

Lessons

Reading a ClinicalTrials.gov record introductory
NCT numbers, statuses, phases and eligibility: what a registry record tells you and what it does not.
Labels versus trial interventions intermediate
How the same product can look different in a label and in a trial registry, and why RadioGx shows the match confidence.
Approval versus clearance intermediate
Drugs are approved; most devices are cleared. The difference matters for radiopharmaceuticals, scanners and microspheres.
Adverse event reports are not causality intermediate
Why FAERS and MAUDE counts cannot tell you how often harm happens or what caused it.
Literature literacy: peer review, preprints and open access introductory
How to weigh a paper, a preprint, a review and a meta-analysis, and how to reach legal open-access copies.
Patent records versus legal strength intermediate
What a patent document shows, what it does not, and why freedom to operate needs a patent attorney.

Public Imaging Datasets & Genomics Resources

TCIA, NCI Imaging Data Commons, GDC and cBioPortal, how they link, and why paired imaging-genomics is never assumed.

For: Researchers, data scientists

You will be able to

Find public imaging collections
Understand TCGA linkage
Respect controlled access and de-identification limits

Lessons

TCIA and the NCI Imaging Data Commons introductory
Where public, de-identified cancer imaging lives, and the rules that come with it.
GDC, cBioPortal and TCGA linkage intermediate
How public genomic resources are organised and how some imaging collections link to them.
Paired imaging and genomics: never assume intermediate
Why "this dataset has imaging and genomics" needs checking case by case.
Controlled access, de-identification and ethics intermediate
Why some data need approval, why de-identification is hard for images, and why RadioGx accepts no uploads in the workspace.

Open-Source Tooling & Repository Literacy

How to judge a repository, read a licence, and use OHIF, MONAI, PyRadiomics and friends safely on public data.

For: Developers, researchers

You will be able to

Evaluate a repository beyond its stars
Tell permissive from copyleft licences
Keep PHI out of every tool

Lessons

How to read a repository introductory
Stars measure attention, not validity. Here is what to look at instead.
Licences for code and for data intermediate
Permissive, copyleft and no licence at all, plus the separate licences that govern datasets and models.
The open imaging toolchain intermediate
A map of widely used open tools: reading DICOM, processing images, training models, extracting features and viewing.
Safe practice with tools and data introductory
Public, de-identified data only; no PHI uploads; reproducible environments; and a clear line between research code and medical devices.