Radiology reporting skills improve fastest when a resident treats the report as its own discipline, with its own framework, its own practice method and its own error log, rather than as a by-product of knowing radiology. The reliable path is a repeatable loop: review the study systematically, organise findings before writing, write a focused findings section and a committed impression, compare your report against the faculty version, and log the differences. This article works through each step and ends with a four-week plan.
Nothing here is patient-specific clinical advice. The examples are fictional teaching constructs, and reporting in training always happens under faculty supervision.
Why reporting is a separate skill
Recognition is perceptual; reporting is compositional. Between seeing a finding and producing a useful report sit several decisions the eye does not make: what else must be checked, what deserves mention, in what order, in what words, with what degree of confidence, and what the referring clinician actually needs from the conclusion. Residents who read extensively but report little develop the first skill and not the second, which is why reporting volume under correction, not reading volume, predicts reporting quality.
The qualities of a useful report
A useful report is accurate, complete for the clinical question, organised so findings can be located quickly, written in precise descriptive language, honest about uncertainty, and finished with an impression that answers the question asked. Brevity serves all of these; a report is not improved by length. Radiological societies’ guidance on communication, including the reporting guidance published by bodies such as the American College of Radiology, converges on these same qualities, and they are worth internalising early because they are also what examiners reward.
A repeatable image-review framework
Reporting errors usually begin before writing, in an incomplete search. Fix the search first:
- Confirm patient, study, technique and adequacy, and read the clinical history
- Review prior imaging before forming an opinion on the current study
- Apply a fixed search pattern for the modality and region, the same order every time, including review areas commonly missed (study edges, bones on soft-tissue studies, the regions away from the obvious finding)
- Re-window and review additional planes or sequences deliberately, not incidentally
- Only then decide what the findings mean
A search pattern is personal, but it must be fixed. The pattern that changes with each study is not a pattern.
The five-step reporting improvement loop
Figure: Reporting improves through comparison and correction. Suggested placement: after the review-framework section.
Organise findings before writing
Before typing, sort what you found into three mental bins: findings relevant to the clinical question, significant incidental findings, and normal or irrelevant observations. The first bin leads the report, the second follows clearly labelled, and the third mostly stays out. Thirty seconds of sorting prevents the commonest structural fault in resident reports: findings listed in the order noticed, forcing the reader to do the organising.
Writing a focused findings section
- Describe before you interpret: site, size, morphology, density or signal characteristics, enhancement and relationships, in consistent terminology
- Group by organ or system, not by the order of discovery
- Include the relevant negatives that address the clinical question
- Measure what will be followed; a lesion without a size cannot be tracked
- Omit normal-structure inventories that add length without information, unless departmental format requires them
Constructing a useful impression
The impression is the report. Most clinicians read it first and some read only it. A dependable structure: answer the clinical question in the first line; give a short, prioritised differential only where genuine uncertainty exists; state significant incidentals; and recommend the next step where one is clearly indicated. Number the points if there are more than two. An impression that restates the findings section verbatim, or lists six differentials without preference, has declined to do its job.
Prioritising what matters clinically
Lead with what changes management. A finding that requires urgent communication is also a communication task, not only a line in a report; follow your department’s protocol for direct notification of critical findings and document that it happened. Ordering the impression by clinical importance, not anatomical sequence, is one of the fastest visible upgrades a resident can make.
Managing uncertainty appropriately
Uncertainty is legitimate; vagueness is not. Calibrate language to confidence: state confident findings plainly, use “likely” or “consistent with” where imaging supports but does not prove, and reserve hedged phrasing for genuine ambiguity, paired with the step that would resolve it. Two faults sit at the extremes: hedging everything, which transfers the entire diagnostic burden to the clinician, and overcalling certainty, which is worse. When supervision is available, uncertainty is also a prompt to show the case, not merely to soften the sentence.
Structured reporting: where it helps and where it restricts
Structured templates improve completeness, consistency and clarity, particularly for standardised examinations and follow-up studies. Initiatives such as the RSNA’s structured reporting effort exist because free prose is demonstrably variable. The limitation is real, though: complex, multi-system or unexpected findings can be flattened by rigid fields, and template-filling can dull the active search if boxes replace looking. A sensible residency habit is to learn the structured formats your department uses, and to treat them as a floor for completeness rather than a ceiling on thought.
Compare your report with the faculty report
This is the engine of improvement. For every study you report provisionally, place your version beside the final faculty version and ask four questions: What did I miss? What did I include that was removed, and why? How did the wording change? How did the impression change? The last question teaches the most, because impressions are where experience shows. Where your platform or department archives both versions, review them weekly rather than case by case, so patterns become visible.
Keep a reporting error log
Log every meaningful difference in four columns: date and study type, the error (miss, overcall, wording, structure, impression), the corrected version, and the lesson in one line. Review it weekly. Within a month the log will show your pattern, and it is usually narrower than feared: most residents cycle through a small set of repeated faults, and a named fault is a fixable one.
A four-week reporting improvement plan
- Week 1 – Search: fix your search pattern per modality; report your usual volume but audit only completeness against faculty versions
- Week 2 – Findings: audit organisation and language; rewrite two of your own reports each day into grouped, precise findings sections
- Week 3 – Impression: draft the impression first for each study, then write findings to support it; compare every impression against the faculty version
- Week 4 – Integration: full-loop reporting with the error log driving a re-review of the month’s misses; repeat the cycle at higher study complexity
Weak versus improved reporting language
Fictional, non-patient-specific examples:
| Weak | Improved | Why |
| “There appears to be a possible lesion in the liver which could be significant.” | “2.1 cm hypodense lesion in segment VI, with peripheral enhancement.” | Describes instead of hedging; measurable and locatable |
| “No significant abnormality seen as such.” | “No acute intracranial haemorrhage, infarct or mass effect.” | Relevant negatives tied to the clinical question |
| “Impression: findings as described above.” | “Impression: 1. Acute uncomplicated appendicitis. 2. Incidental 4 mm left renal calculus.” | The impression answers the question and ranks the findings |
| “Cannot rule out infection or malignancy or other pathology.” | “Findings favour an infective process; malignancy is not excluded. Suggest follow-up imaging in 6 weeks or tissue sampling if clinically indicated.” | Calibrated uncertainty with a resolving step |
How DICOM-based practice supports reporting
Reporting practice requires whole studies, because the skill includes the search, the windowing decisions and the triage of incidentals, none of which exist in a pre-selected image. Practising on scrollable DICOM cases, drafting a full report and then comparing it against a model report reproduces the faculty-comparison loop outside duty hours. It supplements supervised reporting; it does not replace it. Residents who want a ready-made version of this loop can use the long and short cases in RadNexa’s DICOM Atlas, each paired with clinical context, as report-writing material for the four-week plan above.
