Diagnostics roadmap: stains → gene panels → blood tests that decide treatment

Histopathology means looking at cancer cells under a microscope, and immunohistochemistry stains them for specific proteins. Together they are still the foundation of every diagnosis.

Looking at the leukaemia's chromosomes under a microscope, or lighting up specific gene breaks with fluorescent probes, to classify risk.

The tests that grade a breast or stomach cancer's HER2 level, from the original trastuzumab test in 1998 to the new 'HER2-low' and 'ultralow' cut-offs.

The test that decides whether a specific drug is right for you, approved together with the drug.

Taking a sliver of tissue (core) or a few cells (fine-needle aspiration) through a needle guided by ultrasound, CT or MRI, to diagnose the cancer and test its markers without surgery.

The lung cancer gene test that became the first blood-based companion diagnostic the FDA ever approved.

A family of quick single-gene tests that decide who can have several bowel, lung, breast and bladder cancer drugs.

A 21-gene test that tells most women with early hormone-positive breast cancer whether they can safely skip chemotherapy.

Showed that most women with the commonest breast cancer can safely skip chemotherapy if a gene test says their risk is low or intermediate.

Postmenopausal women with a few positive lymph nodes and a low gene-test score can skip chemotherapy; premenopausal women still benefit from it.

A 70-gene test that tells whether an early breast cancer is genomically low or high risk, used to decide who can skip chemotherapy.

A 50-gene test run in local hospital laboratories that estimates the ten-year risk of a hormone-positive breast cancer coming back.

Sequencing hundreds of cancer genes at once from a biopsy to find the mutations a drug can target.

The FDA-approved tissue (324 genes) and blood genomic tests that serve as companion diagnostics for dozens of drugs.

A large gene panel hospitals can run themselves, approved by the FDA in 2024 as a companion diagnostic for the tumour-agnostic drug larotrectinib.

Tempus's tumour-and-normal gene panel, FDA-approved in 2023 as a companion test for EGFR antibodies in bowel cancer.

A blood test that reads a tumour's mutations without a tissue biopsy and is the FDA-approved gateway to several targeted drugs.

Tests that show whether a tumour has lost its DNA spell-checker; if so, immunotherapy works unusually well and an inherited syndrome may be present.

Counting how many mutations a tumour carries per stretch of DNA; heavily mutated tumours are more likely to respond to immunotherapy.

Tests that reveal whether a tumour has a broken DNA repair system, which predicts response to PARP inhibitors and platinum.

Curated databases that say what each mutation means for treatment, and the expert meetings that use them to decide on therapy.

Regular meetings where surgeons, oncologists, radiologists, pathologists and others review each patient's case together and agree a plan; mandatory in many countries and associated with more guideline-concordant care.

A tumour-agnostic approval lets a drug be used for any cancer carrying a specific molecular feature, regardless of where it started.

An ultra-sensitive blood test after surgery that detects leftover cancer months before a scan would.

A blood test that reads fragments of DNA shed by the tumour, so you can genotype or monitor cancer without a needle in the tumour.

Showed that a blood test can safely halve the number of colon cancer patients given chemotherapy after surgery.

Japan's national programme tests whether a blood test after surgery should decide who gets chemotherapy, and whether treating a positive test early helps.

The first trial to change treatment because of a blood test rather than a scan: switching to camizestrant when an ESR1 mutation appeared in the blood delayed progression by seven months.

The first trial to use a blood test for leftover cancer to decide who gets immunotherapy, and it worked.

The most widely used blood test for detecting leftover cancer after surgery, personalised to each patient's tumour mutations.

NeoGenomics' personalised blood test for tiny amounts of leftover cancer, tracking up to 48 mutations from the patient's own tumour.

A blood test for leftover cancer after surgery that needs no tumour sample, so results come faster than tumour-informed tests.

A blood-test-selected trial testing whether matching the drug to the specific resistance mutation beats the standard second-line drug.

Scanning microscope slides and letting software measure things a pathologist cannot see, including predictions of who will benefit from a treatment.

Very large AI models trained on millions of slides or scans that can be adapted to almost any diagnostic question.

The scanners that turn glass slides into gigapixel images, and the software that stores and serves them, without which pathology AI cannot run.

A pathology foundation model trained on millions of slides that can detect cancer and predict biomarkers from an ordinary H&E slide.

An open pathology model trained on 1.3 billion image tiles from a US health system, modelling whole slides at gigapixel scale.

The first AI tool cleared by the FDA to predict both prognosis and treatment benefit from a routine biopsy slide, in prostate cancer.

An FDA-cleared AI test (May 2026) that reads breast cancer slides to estimate recurrence risk in early hormone-positive disease.

The first randomised trial of AI in breast screening found more cancers and cut radiologists' reading work almost in half without more false alarms.

A model that reads slides and clinical text together to predict who will respond to immunotherapy.

Reading the genes of each individual cell, and mapping where each cell sits in the tumour.

Choosing treatment from a map of where each cell type sits in the tumour, not just from a list of its mutations.

Spatial biology instruments are machines that map which genes and proteins are active in each part of a tumour slice.

Measuring the proteins in a tumour, which is what drugs actually hit, rather than the genes that encode them.

Reading DNA in very long stretches, which reveals rearrangements and methylation that short-read machines miss.

Instead of testing blood every three months, sampling constantly, so a relapse is caught the week it starts.

Fragmentomics reads the sizes and positions of DNA fragments in blood, not the mutations. Cancer cells die messily and leave a recognisable fragmentation pattern.

Reading chemical tags on DNA that reveal a cell's identity, used to classify brain tumours and to detect cancer in blood.

Tests that decide who gets a drug are often not validated prospectively and are measured differently in every lab.

After a 'successful' treatment, cells can sleep for years then relapse. We can barely detect them and cannot target them.

Records, scans, genomes and outcomes sit in separate systems that cannot talk. Every patient's experience is lost to the next.

A tumour is many tumours. Treatments that kill most cells leave the rest to grow back, changed.

Drug trials must be registered before they start so results cannot be hidden or reshaped. Studies that claim a biomarker predicts outcome should be registered too.

Blood tests can now find leftover cancer months before scans, but most patients who test positive have nothing to enrol in. One standing trial per country would fix that.

Leftover-cancer blood tests are being sold faster than evidence that acting on them helps. Paying for them only when the result is recorded would generate the missing evidence.

Sequence every cancer at diagnosis, along with the patient's inherited genes, and pool the results with treatments and outcomes so every patient teaches the system how to treat the next.

Most cancer tests in the US, including Galleri, Signatera and Oncotype DX, are 'lab-developed tests' overseen through lab standards rather than FDA approval; the FDA's 2024 attempt to change that was struck down in court in 2025.

The big labs that run most biomarker tests, and the reagent makers whose stains decide who gets a drug.