ideasIdea
Structured, coded radiology reports for cancer response instead of free text
Radiologists would record tumour measurements and response in tick-box, coded form rather than prose, so progression is machine-readable across every scan.
Response assessment (RECIST, PERCIST) is documented in free text that machines cannot reliably parse. Structured reporting templates with coded lesion measurements (RSNA RadReport, DICOM SR) and a response field would make progression a first-class data element for registries and RWE. The proposal mandates structured oncology follow-up reporting in publicly funded imaging and funds template integration in reporting software.
Hypothesis
Structured response reporting will raise machine-derivable progression dates from under 30 percent to over 90 percent of follow-up scans and make real-world PFS as reliable as trial PFS.
Rationale
Structured reporting has improved completeness in prostate MRI (PI-RADS) and lung nodules (Lung-RADS); response assessment is the missing oncology template.
What would test it
Deploy a structured response template at three centres; compare agreement of derived progression dates with trial-style central review in 500 patients against free-text NLP extraction.
Maturity
early clinical
Who has to act
clinic
Cost to try
Small (under $1M)
Years to first evidence
2
Bottlenecks it attacks
- Data silos · Records, scans, genomes and outcomes sit in separate systems that cannot talk. Every patient's experience is lost to the next.
- Weak real-world evidence and registries · We do not reliably know what happens to patients after approval, so we cannot tell which drugs deliver in practice.