How to Document Radiology Reports and Diagnostic Imaging Findings

A radiology report is not a summary of images. It is a clinical document that drives diagnosis, guides treatment, determines follow-up intervals, and in some cases, triggers urgent interventions within the hour. The quality of that document depends on how well the radiologist translates visual and quantitative findings into language that is precise, standardized, and actionable.

This guide addresses the specific documentation challenges that come up most often in practice: how to organize a report that is both comprehensive and readable, how to apply standardized reporting systems without losing clinical nuance, how to document critical findings and prove they were communicated, how to write addenda correctly, and how to create a documentation trail that holds up in peer review and legal scrutiny.

Why Radiology Report Structure Matters

A poorly organized report forces the ordering clinician to read four paragraphs before learning whether the scan is concerning. A report with a vague impression shifts decision burden back to someone who was not trained in imaging interpretation. A report that uses inconsistent terminology makes longitudinal comparison impossible.

The American College of Radiology (ACR) practice parameters define the minimum components a radiology report must contain: patient demographics, clinical indication, technique, body of findings, and impression. Most practices add a few more elements depending on subspecialty and payer requirements. What the ACR does not fully govern is how well those sections are written, and that is where most reports fall short.

The sections below walk through each component in the order they should appear, with specific guidance on what belongs in each.

Report Organization: Section by Section

Patient Demographics and Exam Identification

This section is often auto-populated by the RIS or PACS system, but the radiologist bears responsibility for verifying its accuracy before signing.

Required elements:

• Patient name and date of birth (or medical record number)
• Exam date and time
• Ordering provider name
• Clinical indication (pulled from the order, but the radiologist should flag missing or inadequate indications before interpreting)
• Exam type and laterality (e.g., MRI right knee without contrast)
• Accession number

A report signed with the wrong patient name or exam type is a patient safety event, not just an administrative error. Verification before dictation is a two-second step that prevents significant downstream harm.

Clinical History and Indication

This section should reflect the clinical context the radiologist used to frame interpretation. If the ordering clinician provided "leg pain," that is insufficient to guide interpretation. Good practice is to document what was available and, when relevant, note that the indication was incomplete.

Example: "Clinical indication per order: 'right knee pain.' No additional clinical history available at time of interpretation."

For more complex cases: "Clinical indication: known stage IIIB non-small cell lung cancer, presenting for staging CT prior to initiation of immunotherapy. Prior chest CT from March 2026 available for comparison."

Documenting the available clinical context also provides legal protection. If the interpretation was guided by limited information, the record reflects that.

Technique

The technique section documents how the imaging was acquired. This section matters for reproducibility, for explaining why certain findings may be limited, and for radiation dose accountability.

Include:

• Modality (CT, MRI, US, X-ray, nuclear medicine)
• Body region and field of view
• Contrast administration: agent, dose, route, and timing (arterial/venous/delayed phase for CT); if no contrast, document why (allergy, renal function, clinical decision)
• For MRI: field strength, sequences performed, planes acquired
• For CT: kVp, mAs (or dose length product where applicable), reconstruction kernel
• For fluoroscopy and nuclear medicine: specific protocol, radiopharmaceutical agent, and administered dose in millicuries or megabecquerels
• Comparison studies reviewed (modality, date, accession if available)
• Technical limitations (patient motion, body habitus, incomplete breath-hold, limited field of view)

A statement like "CT abdomen and pelvis with contrast" is the minimum. A statement like "CT abdomen and pelvis performed with 100 mL of intravenous iohexol 350 mg I/mL administered at 3.5 mL/sec with portal venous phase acquisition. Axial 1.25 mm slices reconstructed in coronal and sagittal planes. Compared to prior CT from December 2025 (Accession 2025-048271)" is defensible documentation.

Findings

This is the body of the report. Findings should be organized by anatomical region or organ system, presented in a consistent order, and written in plain professional language.

Organize systematically, not chronologically

Do not dictate findings in the order you noticed them on the screen. Organize by anatomy: lungs, mediastinum, heart, pleura, bones, soft tissue, for a chest CT. Clinicians scanning for a specific system should be able to locate that information without reading the entire report.

Be specific and measurable

Avoid vague qualifiers like "slightly," "mildly prominent," or "possibly." Every finding that will drive a decision should have dimensions, location, density or signal characteristics, and morphology.

Weak example: "There is a lung nodule in the right upper lobe."

Stronger example: "Solitary solid pulmonary nodule in the posterior segment of the right upper lobe, measuring 8 mm in greatest axial dimension (series 3, image 47), with well-defined margins and no calcification. No satellite nodules. No mediastinal or hilar adenopathy."

For mass lesions: document size in at least two dimensions (three dimensions for treatment response), enhancement characteristics, relationship to adjacent structures, and evidence of involvement of critical structures (vascular encasement, airway compression, perineural spread).

Document relevant negatives

If the clinical indication raises a specific question, answer it explicitly. "No pulmonary embolism identified to the segmental level" is more useful than "pulmonary vasculature appears normal." "No acute intracranial hemorrhage, mass lesion, or midline shift" addresses the emergency room clinician's actual question.

Relevant negatives are particularly important in time-sensitive contexts. Failing to explicitly exclude a finding that the ordering clinician was worried about may result in that finding being chased with additional imaging unnecessarily.

Incidental findings

Every report should address incidental findings that carry clinical significance. If you find an adrenal nodule on a CT ordered for renal colic, document it. Then provide guidance in the impression or a dedicated follow-up recommendation section.

For incidental findings, include:

• Description and measurement
• Characterization where possible (e.g., Hounsfield unit (HU) measurement to differentiate adenoma from lipid-poor adenoma or metastasis)
• ACR Incidental Findings Committee guideline reference where applicable
• Recommended follow-up

Standardized Reporting Systems

Several subspecialty areas have adopted lexicon-based categorical scoring systems that replace descriptive uncertainty with a standardized risk tier. Using these correctly is both a clinical and documentation obligation.

BI-RADS (Breast Imaging)

The Breast Imaging Reporting and Data System (BI-RADS), developed by the ACR, assigns findings to categories 0 through 6 with defined management implications:

Documentation standard: Every mammography and breast MRI report must include a final BI-RADS category and the corresponding management recommendation. Subspecialty language like "indeterminate" or "recommend clinical correlation" without a BI-RADS category is non-compliant with ACR practice parameters.

Lung-RADS

Lung-RADS (ACR, currently version 1.1) applies to low-dose CT lung cancer screening. Categories range from 1 (no nodules or definitely benign nodules) through 4X (additional features increasing suspicion). Each category links to a defined follow-up interval.

Lung-RADS documentation must include:

• Category assigned (1, 2, 3, 4A, 4B, 4X)
• Nodule characteristics that drove the categorization: type (solid, part-solid, non-solid), size (longest diameter for solid; longest diameter of solid component for part-solid), location
• If multiple nodules are present, categorize each and report the overall highest category
• Management recommendation per the assigned category

A Lung-RADS 3 nodule (6-8 mm solid, or 6 mm new solid) carries a 1-2% malignancy risk and requires 6-month LDCT follow-up. That management recommendation must appear explicitly in the report, not just in a follow-up worklist.

TI-RADS (Thyroid)

The ACR Thyroid Imaging, Reporting and Data System (TI-RADS) assigns points across five feature categories: composition, echogenicity, shape, margin, and echogenic foci. Point totals determine the TR category (TR1 through TR5) and whether FNA is warranted based on the nodule's maximum diameter.

Documentation must include:

• Each of the five feature categories with the selected descriptor and assigned point value
• Total point score and resulting TR category
• Whether FNA is recommended based on the TR category and nodule size threshold
• Size measurement (maximum dimension)

A common documentation error is assigning a TR category without documenting the scoring rationale. If a reviewer or clinician questions the recommendation, the scoring trail must be visible in the report itself.

LI-RADS (Liver)

LI-RADS is used for hepatocellular carcinoma (HCC) surveillance in at-risk patients (cirrhosis, chronic hepatitis B, prior HCC). The system applies to CT and MRI with contrast and uses a combination of observation size and imaging features to assign categories from LR-1 (definitely benign) through LR-5 (definitely HCC), with additional categories for treatment response and malignancy not specific to HCC.

Key LI-RADS documentation elements:

• Confirm the patient population is LI-RADS-eligible (documented at-risk status)
• Observation size and location (segment)
• Major features: arterial phase hyperenhancement (APHE), washout appearance, enhancing capsule, size threshold, and threshold growth
• Ancillary features that increase or decrease LR category
• Assigned LR category with management implication
• LI-RADS version used (2018 is current; note if using CT LI-RADS vs. MRI LI-RADS)

C-RADS and Other Systems

C-RADS (colon CT colonography), PI-RADS (prostate MRI), O-RADS (ovarian/adnexal US and MRI), and NI-RADS (head and neck post-treatment surveillance) follow the same documentation logic: assign the category, document the features that drove it, and state the management recommendation.

If your institution uses a system not covered here, the principle is the same: the category alone is insufficient. The report must show the reader why that category was assigned.

Critical Findings: Documentation and Communication

Critical findings (also called urgent or significant unexpected findings) require immediate communication to the ordering clinician, not just reporting into the electronic medical record. The documentation of that communication is a separate and essential step.

What qualifies as a critical finding

Each institution should maintain an approved critical findings list, but common examples across radiology subspecialties include:

• New intracranial hemorrhage or herniation
• Tension pneumothorax
• Pulmonary embolism with right heart strain
• Acute aortic dissection or aortic aneurysm with rupture signs
• Malpositioned endotracheal tube or vascular catheter
• Bowel perforation or ischemia
• New mass highly suspicious for malignancy in a previously undiagnosed patient (often categorized as "significant unexpected finding" requiring timely but not emergent communication)

What the communication documentation must include

The ACR Practice Parameter on Communicating Findings requires that the report or an addendum document:

1. Who was notified (name, role, and if possible, direct callback number or beeper)
2. When the notification occurred (date and time)
3. How the notification was made (direct phone call, pager, secure message)
4. What was communicated (the specific finding, not just "I called about the scan")
5. Acknowledgment received (that the clinician confirmed receipt)

Example documentation: "Critical finding communicated by direct telephone call to Dr. James Okonkwo (attending emergency physician, ED) at 14:32 on April 27, 2026. Finding communicated: acute right-sided epidural hematoma with 7 mm midline shift and early uncal herniation on CT head. Verbal acknowledgment received."

If the ordering clinician cannot be reached, document every attempt: "Attempted contact with Dr. Okonkwo (ED attending) at 14:28 via pager, no response. Called ED charge nurse (nurse acknowledged and confirmed neurosurgery was already at bedside) at 14:35."

The documentation of failed attempts protects the radiologist legally. An unreturned page with no follow-up attempt does not.

Significant unexpected findings (non-critical urgency)

Some findings are clinically important but do not require notification within minutes. An incidental renal mass suspicious for clear cell carcinoma in a patient scanned for appendicitis is significant but not a life-threatening emergency. Document the finding clearly in the report, include a follow-up recommendation, and note in the impression that the ordering clinician should be notified of this incidental finding.

Impression Section

The impression is the section ordering clinicians read first, often exclusively. It must stand alone as a clinically useful summary.

Principles for a strong impression:

• Answer the clinical question directly. If the indication was "rule out PE," the first sentence of the impression should address PE.
• List findings in order of clinical significance, not anatomical order.
• Assign categories when applicable (BI-RADS, Lung-RADS, etc.) in the impression, not only in the findings section.
• Provide actionable follow-up recommendations with time frames, not vague language like "may warrant further evaluation."
• Avoid hedging that obscures actionability. "Cannot exclude" a finding when the imaging is essentially negative may generate unnecessary follow-up. Be specific about what level of confidence you have and why.
• Limit length. Three to five numbered items cover most studies. A twelve-item impression is a findings section masquerading as an impression.

Example impression for a complex study:

1. 2.1 cm solid enhancing renal mass in the mid-pole of the right kidney (series 4, image 62), with imaging features consistent with clear cell renal cell carcinoma (RCC). No evidence of renal vein or inferior vena cava involvement. No regional lymphadenopathy. Recommend urology referral and contrast-enhanced CT urogram or renal protocol MRI within 4 weeks for surgical planning.
2. Incidental 1.2 cm right adrenal nodule with Hounsfield unit measurement of 12 on unenhanced phase, consistent with benign adrenal adenoma. No follow-up required per ACR Incidental Findings guidelines.
3. Scattered diverticulosis of the sigmoid colon without evidence of diverticulitis.
4. No acute osseous abnormality.

Dictation Best Practices

Most radiologists dictate reports rather than type them. Voice recognition software introduces specific documentation vulnerabilities that structured habits can prevent.

• Dictate section headers explicitly. Many errors come from findings that bleed across sections when headers are assumed rather than spoken.
• Dictate measurements with units. "Eight" is not a measurement. "Eight millimeters" is.
• Proofread before signing. Voice recognition errors in radiology reports have caused patient harm, including incorrect laterality and missed critical findings that were dictated but misrecognized.
• Avoid filler language. Phrases like "there appears to be" and "one might consider" reduce the clinical utility of the report. If you saw it, describe it. If you are uncertain, qualify the uncertainty specifically ("low-grade T2 signal abnormality in the posterior limb of the internal capsule of uncertain significance; differential includes artifact, early ischemic change, or demyelination").
• Use standardized terminology. Within your institution, agree on terminology for common findings. "Hypodensity" and "low attenuation" mean the same thing on CT; using both interchangeably makes computer-assisted extraction and longitudinal comparison harder.
• Final review against the images. Proofread the signed report against the original images at least for complex cases. A finding dictated in the wrong laterality or incorrect anatomical location is a reportable error.

Addendum and Amended Report Protocols

Radiology reports are medical records. Altering them after signature requires following a specific protocol to maintain an accurate audit trail.

Addendum

An addendum is appropriate when you need to add information to a previously signed report without changing the original findings or impression. Common reasons include:

• Comparison studies became available after the original report was signed
• Additional clinical history was provided that changes the clinical context but not the findings
• A follow-up recommendation needs to be added after a clinical discussion

An addendum must:

• Be clearly labeled as an addendum (not a correction)
• Reference the original report by accession number and date
• State why the addendum is being made
• Be electronically signed with a new date and timestamp

Amended report

An amended report (sometimes called a corrected report) is necessary when a substantive error in findings or impression must be corrected. This includes errors in laterality, organ involvement, measurement, or clinical impression.

An amended report must:

• Preserve the original report (never delete or overwrite)
• Clearly label the amended version as an amendment
• Identify specifically what was changed and why
• Be signed with a new date and timestamp
• Trigger communication to the ordering clinician if the change is clinically significant

The decision between addendum and amendment is not always obvious. A rule of thumb: if the change could alter clinical management, it is an amendment and requires clinician notification.

What never to do

• Never sign a report you did not read and are not personally vouching for
• Never delete a report from the system, even if it was signed in error
• Never alter a report without leaving a documentation trail
• Never back-date a correction to match the original sign time

Quality Assurance Documentation

Radiology departments are required to maintain QA programs. As a radiologist, your individual documentation habits feed directly into practice-level QA metrics.

Peer review

Peer review programs (many using the RADPEER system or institution-specific tools) require that a portion of signed reports be reviewed by peers for diagnostic accuracy and documentation quality. When your report is reviewed, the reviewer is evaluating whether your findings and impression are supported by the images.

QA documentation best practices at the individual level:

• Maintain a personal log of cases you flag for self-review (complex cases, cases where you had significant diagnostic uncertainty, cases where you changed a preliminary read)
• Document discrepancies with preliminary reads in the final report when they are clinically significant
• When you review a prior report and find a missed or miscategorized finding, follow your institution's discrepancy reporting protocol rather than simply noting it in your own interpretation

Incidental finding tracking

Many institutions use incidental finding registries or follow-up tracking systems to ensure that recommended follow-up actually occurs. Radiologists who recommend a 6-month CT follow-up have a professional (and in some states, a legal) responsibility to ensure that recommendation is visible and tracked. Documentation in the report is necessary but not sufficient in high-volume practices.

If your institution does not have an automated follow-up tracking system, document your recommendations in a format that the ordering clinician can act on: specific time frame, specific modality, and specific clinical contact if the study should be ordered by a different provider.

Well-Structured vs. Poorly-Structured Report: A Comparison

The following fictional examples illustrate the difference in documentation quality for the same CT study.

Poorly structured report

Dr. Maria Vasquez dictates the following for a CT abdomen/pelvis with contrast on a 52-year-old woman presenting with right upper quadrant pain:

"The liver, spleen, pancreas, and kidneys are normal. Gallbladder shows some stones. The bowel appears unremarkable. There is a small nodule in the lung base. The bones look okay. Impression: Cholelithiasis. Recommend clinical correlation."

Problems: No technique documented. No comparison study referenced. "Some stones" has no count, size, or evidence of wall thickening or pericholecystic fluid. "A small nodule in the lung base" has no measurement, no characterization, no Lung-RADS category, and no follow-up recommendation. "The bones look okay" is not a professional finding statement. The impression is non-actionable.

Well-structured report

Dr. Marcus Reyes documents the same study:

Technique: CT abdomen and pelvis performed with 90 mL of IV iohexol 350 mg I/mL administered at 4 mL/sec, portal venous phase acquisition. Axial 1.25 mm slices reconstructed in coronal and sagittal planes. No prior CT available for comparison.

Findings:

Liver: Normal size and attenuation. No focal lesion. No biliary ductal dilation.

Gallbladder: Multiple gallstones, largest measuring 1.4 cm in greatest dimension. No gallbladder wall thickening or pericholecystic fluid. No sonographic Murphy sign correlate available; clinical findings should guide management.

Pancreas: Normal size, contour, and parenchymal attenuation. No ductal dilation or peripancreatic stranding.

Spleen: Normal size and attenuation.

Kidneys: Bilateral kidneys normal in size and enhancement. No hydronephrosis, calculus, or solid mass.

Adrenal glands: Normal bilaterally.

Bowel: No bowel wall thickening, obstruction, or free air. Normal appendix.

Retroperitoneum/Lymph nodes: No pathologic adenopathy.

Bones: No acute osseous abnormality. Degenerative changes of the lumbar spine, most pronounced at L4-L5.

Lung bases (imaged): 6 mm solid pulmonary nodule in the medial right lower lobe (series 2, image 14), with smooth margins and no satellite nodules. This nodule is not optimally characterized on abdominal CT; recommend dedicated chest CT for further evaluation. See impression.

Impression:

1. Cholelithiasis without imaging evidence of acute cholecystitis.
2. Incidental 6 mm solid right lower lobe pulmonary nodule, incompletely characterized on abdominal CT. Recommend low-dose chest CT for characterization and Lung-RADS categorization. Per ACR guidelines for 6 mm solid nodules in patients with average risk and no prior CT available for comparison, CT follow-up in 12 months is recommended if dedicated chest CT confirms a sub-centimeter nodule without high-risk features.
3. No acute intra-abdominal or pelvic pathology.

The difference is not length. It is specificity, actionability, and the presence of a documented reasoning trail.

Documentation Checklist

Use this checklist before signing any radiology report.

Demographics and exam verification

• Patient name and date of birth verified against the order
• Exam type, date, and laterality confirmed correct
• Ordering provider identified

Technique section

• Modality, body region, and contrast documented
• Comparison studies identified and reviewed (or absence noted)
• Technical limitations noted if applicable

Findings section

• Organized by anatomical region, not observation order
• All target organ findings addressed
• Relevant negatives documented for findings implicated by the clinical indication
• Measurements with units for all significant findings
• Incidental findings described, characterized, and followed up

Standardized reporting systems

• Applicable system (BI-RADS, Lung-RADS, TI-RADS, LI-RADS, etc.) assigned
• Category justified by documented features
• Management recommendation linked to the assigned category

Impression

• Clinical question answered directly in the first or second item
• Findings listed in order of clinical significance
• Follow-up recommendations with specific time frames and modality
• Category assigned in impression if using a standardized system

Critical findings

• Critical finding identified if present
• Direct communication documented: who, when, how, what, acknowledgment
• Attempts documented if initial contact was unsuccessful

Addendum or amendment (if applicable)

• Labeled correctly (addendum vs. amendment)
• Original report preserved
• Reason for change documented
• Clinician notified if change affects management

---

Consistent radiology report documentation protects patients, protects the radiologist, and makes every downstream clinical decision more efficient. The habits described above take longer to build than to maintain. Once they are part of your dictation workflow, the time cost is minimal and the quality difference is significant.

If you work in a setting where post-session documentation follows a notes-from-summary workflow, tools like NotuDocs support template-first structured note generation from brief written summaries. Note that NotuDocs is not HIPAA-compliant and is not designed for radiology report generation within a clinical EHR or RIS system.

---

Related guides:

• How to Document Psychiatric Medication Management Visits and Prescriber Notes
• How to Document Oncology Patient Visits and Cancer Treatment Plans
• How to Document Home Health Nursing Visits and Patient Assessments