Radiology report format: structure, sections and what a normal CT or MRI report contains

Section by section

The structure below is what the ACR Practice Parameter for Communication of Diagnostic Imaging Findings describes and what most Indian departments follow¹. The order is deliberate: identification first, so a report can never be attached to the wrong patient; findings before impression, so the reasoning is visible; signature last, so what is signed is unambiguous.

What must appear on a signed report in India

No single Indian statute lists the contents of a radiology report. The requirement is assembled from three places. First, professional regulation: the MCI (Professional Conduct, Etiquette and Ethics) Regulations 2002 require a registered practitioner to display their registration number on prescriptions, certificates and receipts given to patients, and to keep indoor records for three years³. Second, the NMC’s position on who may sign: in a January 2025 RTI reply the NMC stated that the Telemedicine Practice Guidelines do not permit a practitioner to sign pathology and radiology lab reports via remote authentication without physical presence and supervision, and that the guidelines do not address the question⁴. The practical distinction is between ghost-signing another person’s work and a registered radiologist personally interpreting the images and signing their own report; the signature must belong to the radiologist who read the study. Third, the law on signatures: the Information Technology Act 2000 gives electronic signatures legal equivalence under section 5, with Aadhaar eSign recognised under section 3A; a typed name or drawn image is enforceable as a contract term but has no statutory equivalence⁵.

The rules on who may report remotely, and the NMC’s distinction between ghost-signing and a registered radiologist personally interpreting a study, are covered in the guide on teleradiology regulations in India.

Structured reporting versus free text

Free-text reporting is prose dictated in whatever order the radiologist thinks. Structured reporting fixes the sections and, at its fullest, the fields and vocabulary inside them. Three standards define the pieces:

• IHE MRRT (Management of Radiology Report Templates) defines templates as HTML5 documents with named sections and fields so they can be exchanged between systems. RSNA’s RadReport library uses this format, though RSNA has not reviewed or published new submitted templates since December 2022 and is working with the ACR on element-based content instead⁶.
• ACR-RSNA Common Data Elements (radelement.org) give standard definitions and permitted values for imaging findings, so that “lesion size” or “degree of stenosis” means the same thing in every report that uses the element⁷.
• RadLex is RSNA’s radiology lexicon, twenty years old in 2025 and available as HL7 FHIR terminology⁸. It supplies the words; the CDEs supply the structure.

The benefits are practical. Structured reports are more complete because the template asks about every organ; more consistent between radiologists, which matters in a network where the next report on the same patient may come from someone else; faster to produce when combined with dictation and template fill; and searchable, so a department can audit its own work. In India there is a further reason: the NRCES FHIR implementation guide for ABDM defines a DiagnosticReportImaging profile⁹, and a structured report maps onto it far more cleanly than prose. The MoHFW EHR Standards 2016, which are voluntary, point the same way¹¹.

The case for measurements as data

A measurement typed into prose (“a 12 mm nodule in the right upper lobe”) is lost to every system except a human reader. The same measurement stored as a field, with the series and image it was made on, can be compared automatically with the next study, plotted over time and checked for transcription errors before signing. Viewers can export measurements as DICOM Structured Reports; a reporting platform that keeps them as fields rather than words makes follow-up honest.

Example structure: normal CT brain and MRI brain

Example structure only, for education. Not a template for clinical use. Every report must reflect what the radiologist actually saw in the study they signed.

Normal non-contrast CT brain

• Clinical information: the presenting complaint and the question asked.
• Technique: non-contrast axial CT of the brain from skull base to vertex, with coronal and sagittal reformats.
• Comparison: none available, or the prior with its date.
• Findings: No acute intracranial haemorrhage, mass effect or midline shift. Grey-white matter differentiation is preserved. Ventricles and sulci are normal for age. No extra-axial collection. Basal cisterns are patent. Visualised paranasal sinuses and mastoid air cells are clear. No calvarial fracture.
• Impression: No acute intracranial abnormality.
• Signature block: name, qualification, registration number and council, date and time, status.

Normal MRI brain

• Clinical information: the presenting complaint and the question asked.
• Technique: the sequences acquired (for example T1, T2, FLAIR, diffusion-weighted imaging, susceptibility-weighted imaging), planes, and whether contrast was given.
• Comparison: none available, or the prior with its date.
• Findings: No restricted diffusion to suggest acute infarction. No intracranial haemorrhage or susceptibility artefact. No mass lesion or abnormal signal in the supratentorial or infratentorial brain parenchyma. Ventricles and sulci are normal for age. No extra-axial collection. Major intracranial flow voids are preserved. Pituitary, orbits and visualised paranasal sinuses are unremarkable.
• Impression: Normal MRI of the brain.
• Signature block: as above.

Addenda, versions and immutability

A report has a status. A preliminary report is issued when a result is needed before the final read is complete, typically in emergency work; it must be labelled as such and followed by a final. A final report is the definitive documentation of the examination¹. An addendum is a separately signed note that corrects or adds to a final report, referencing it, stating what changed and why, and carrying its own timestamp. The ACR teleradiology white paper expects a teleradiology service to have a process for addenda and for documenting discrepancies between preliminary and final reads¹⁰.

The rule that follows is that a signed report is never edited in place. If the record can be silently changed, it is not a record. A well-designed system enforces this technically: once signed, the row cannot be updated, only superseded, and both versions remain visible with their signatures. How discrepancies between reads are measured is the subject of the guide on radiology peer review and quality.

The critical findings statement

When a report contains a finding that needs action before the referrer would normally read it, the report should say so and say what was done. The ACR communication parameter expects non-routine communication of urgent or significant unexpected findings, documented in the report¹. In practice the statement records four things: the finding, who was told, by what channel, and when. For example: “Critical finding discussed by telephone with the on-call physician at the referring hospital at 02:14 on the date of the study; the finding and the time were acknowledged.” The name of the person reached belongs in the record; whether it is printed on the patient-facing report is a policy choice for the hospital.

Bilingual and delivery considerations

Indian reports are written in English almost without exception, because the referring doctor, the next hospital, the insurer and the courts expect it. What patients often need is a plain-language summary in Hindi, Punjabi or another language, produced from the signed report and marked as a summary, never as a substitute. The delivered artefact is usually a PDF: it should carry the hospital’s header, the full identifiers, every page numbered, and a way to verify that the document is genuine, such as a verification code or QR link, because a PDF forwarded by WhatsApp can be edited by anyone with a laptop. Delivery back into a hospital system by HL7 ORU or FHIR is the ideal where the hospital has such a system; most tier-2 and tier-3 hospitals do not, and PDF by portal, email, WhatsApp or print at the centre is the working reality.

How Raydiac structures reports

Every Raydiac report follows the same structure: identifiers, clinical history and question (mandatory at intake, at least 50 characters), technique, comparison with priors from the same site, findings, impression, recommendations and measurements as fields, then the signature block. Templates for each modality and body part are curated by Dr. Tejinder Singh, the clinical co-founder, and radiologists edit within them, with dictation and template fill to reduce typing. A drafting tool can propose an impression from the radiologist’s own findings text and a pre-sign check reads the text for inconsistencies such as laterality; it runs on Raydiac’s servers in India, sees only the radiologist’s text and never the images, and the radiologist signs every word.

The radiologist reads against a pseudonym and identity is restored at signing, so the signed PDF carries the full patient identifiers, the radiologist’s name, qualification, registration number and council, and the timestamp. Signing is by Aadhaar eSign with a platform counter-signature and a QR verification page, never a scanned image. Signed reports are immutable and superseded only by addenda; critical findings open a closed-loop communication task whose outcome is written into the record. The platform page shows the report editor and the signed output.