Loading
The numbers pulled from scans that decide whether a cancer is shrinking, growing or dead: tumour diameters for RECIST, sugar uptake on PET, water movement on MRI; they run every trial and most clinic decisions, and they are only as good as the way the scan was taken.
What they measure. A scan is a picture, but a decision needs a number. Response criteria turn measurements into categories: RECIST 1.1 sums the longest diameters of up to five target lesions on CT or MRI and calls response, stable disease or progression from the percentage change; PERCIST does the same with the peak standardised uptake value of FDG on PET; the Deauville five-point scale reads FDG uptake in lymphoma against liver and blood pool; PI-RADS, LI-RADS and BI-RADS grade the probability of prostate, liver and breast cancer on multiparametric imaging; the apparent diffusion coefficient on diffusion MRI measures how freely water moves, which falls in dense tumour and rises when cells die; and dynamic contrast measures such as Ktrans track blood vessel leakiness. Radiomics adds hundreds of texture and shape features on top of these.
Who uses them and what changes. Every oncology trial defines its endpoints with these criteria, so a RECIST progression call ends a treatment on trial and usually in clinic too. Deauville scores after two cycles decide escalation and de-escalation in Hodgkin lymphoma; PI-RADS decides who is biopsied; the Lugano criteria decide remission in lymphoma; PSMA-RADS and PERCIST are being written into radioligand and immunotherapy trials. Immunotherapy needed its own variant, iRECIST, because tumours can swell before they shrink.
What limits them. Diameters ignore necrosis and cavitation; standardised uptake values shift with scanner, reconstruction, glucose level and time after injection; and readers disagree, which is why pivotal trials use blinded central review. The RSNA Quantitative Imaging Biomarkers Alliance publishes profiles setting the acquisition and analysis standards under which a measurement can be trusted to a stated precision. None of these measures needs new equipment; the cost is in standardising protocols and training readers.
Standardised measurement of lesion size, tracer uptake, diffusion or perfusion on routinely acquired images, converted by validated criteria into response categories or risk scores with known reproducibility.
Query for this technology: (TITLE:"Quantitative imaging biomarkers" OR ABSTRACT:"Quantitative imaging biomarkers" OR TITLE:"RECIST, PERCIST, SUV, ADC" OR ABSTRACT:"RECIST, PERCIST, SUV, ADC" OR TITLE:"imaging biomarkers" OR ABSTRACT:"imaging biomarkers" OR TITLE:"response criteria" OR ABSTRACT:"response criteria") AND (cancer OR tumor OR tumour OR oncology OR carcinoma OR lymphoma OR leukemia OR leukaemia OR myeloma OR sarcoma OR melanoma OR glioma). Results are unfiltered search hits about Quantitative imaging biomarkers (RECIST, PERCIST, SUV, ADC), not a curated reading list.
Shares PET-adapted (response-adapted) therapy, Deauville score and PET-adapted therapy, Hodgkin lymphoma, Non-Hodgkin lymphoma (all types).
Shares Radiomics, AI in radiology, Non-small-cell lung cancer.
Shares RECIST, CT (computed tomography), MRI.
Shares PET-adapted (response-adapted) therapy, Deauville score and PET-adapted therapy, Hodgkin lymphoma, Non-Hodgkin lymphoma (all types).
Shares FDG PET, PSMA PET, Metastatic cancer (cancer that has spread), Prostate cancer.
Shares AI in radiology, CT (computed tomography), Hepatocellular carcinoma, Non-small-cell lung cancer.
Shares RECIST, CT (computed tomography), MRI.
Open-source projects that implement or serve this technology, from OnCo's own catalogue: licence and last activity as the repository reported them on the day of the fetch. Listing is not endorsement; check the licence before reuse and the validation before clinical use.
The reference open implementation of radiomic feature extraction, standardised against the Image Biomarker Standardisation Initiative.
Converts segmentations and quantitative results to and from standard DICOM objects, from the Quantitative Image Informatics for Cancer Research project.
A self-supervised CT foundation model for lesion characterisation and prognosis, with code and weights.
Commercial and regulated products that serve this technology. Each card says what is behind it: a regulator's database, the literature, a public body's list, or only the company's own words. Listing is not endorsement, and a clearance is a regulatory fact, not a clinical one.
Software that segments the prostate on MR and highlights areas suspicious for clinically significant cancer.
Radiomic feature extraction from PET, CT and MR images, free to use and widely cited in imaging biomarker work.
The picture archiving and communication system a radiology department reads on, and in many hospitals the same viewer now used for digital pathology slides.