What the study found
Radiologists Olgun Şahin and Gul Deniz published the analysis in Academic Radiology in July 2026, titled "Beyond the Primary Complaint: Analysis of Missed and Incidentally Detected Breast Cancers on Emergency Chest CT Scans." They identified 64 patients who had presented to the emergency department for non-breast complaints, then were later diagnosed with breast cancer, and went back through the 73 chest CT scans those patients had on file. Every one of those scans had a BI-RADS 4 or 5 breast lesion visible in the field of view — later confirmed malignant on histopathology.
The finding is stark: about 70% of those scans had the breast lesion missed or unreported at the time of the original read. Roughly 30% were caught and documented. These weren't subtle, borderline findings — they were lesions serious enough, by definition, to later prove malignant. The question isn't whether they were visible. It's why they weren't seen, as reported by Radiology Business.
Why the misses happen — and what made a lesion catchable
An emergency chest CT is ordered for a reason — suspected pulmonary embolism, chest trauma, unexplained pain — and it's read against that reason. Breast tissue sits inside the scan's field of view on almost every chest CT, but it isn't what the read is searching for. That gap between what's imaged and what's actively searched is exactly where an incidental finding on chest CT can slip through.
The study's comparison of missed versus detected lesions shows what tipped the balance:
| Characteristic | Missed lesions | Detected lesions |
|---|---|---|
| Mean lesion size | 13.6 mm | 22.2 mm |
| More common location | Right breast | Inner zone |
| Associated cues | Fewer supporting findings | Axillary lymphadenopathy, fatty (Type A) parenchyma |
| Resulting diagnostic delay | 4–33 months | — |
In other words: the lesions that got caught tended to come with extra visual cues — a suspicious lymph node, a fatty background that made a mass stand out — while the ones that got missed were smaller and less conspicuous, sitting quietly in a part of the image the read wasn't built to interrogate. When they were missed, the delay to eventual diagnosis ran from four months to nearly three years.
A specific version of a familiar problem
This isn't a story about diagnostic accuracy on the primary indication — chest CTs read for pulmonary embolism or trauma are typically read well for exactly that question. It's a story about everything else in the frame that a focused, time-pressured read doesn't have bandwidth to also cover. Radiologists reading emergency studies are, reasonably, optimizing attention toward the clinical question that got the patient scanned in the first place; a similar narrowing of attention under pressure shows up in other radiology detection research as well.
Three practical implications follow for emergency and general radiology reading:
The primary indication isn't the whole scan
Every chest CT captures anatomy well beyond the reason it was ordered — breast tissue, thyroid, adrenal glands, bones. A reporting workflow built only around the ordering complaint leaves that anatomy structurally under-reviewed.
Smaller, less-cued findings need a deliberate second look
The lesions most likely to be missed here were smaller and lacked supporting features. A checklist or secondary-review step focused specifically on frequently overlooked regions — breast parenchyma included — catches what a single indication-driven read tends to skip.
A missed incidental finding has a real timeline attached
A 4-to-33-month gap between a visible finding and its eventual diagnosis isn't an abstraction — it's the practical cost of a scan that already contained the answer going unreported.
Where structured, whole-scan review fits
The practical fix that follows from findings like these is systematic, whole-scan review — giving every anatomic region a look regardless of the ordering complaint, rather than treating anything outside the primary indication as out of scope. That's the same case for a foundation-model approach to CT reporting: a system that examines the whole volume and drafts a structured, comprehensive report — not just the organ tied to the ordering indication — is built to surface incidental findings on chest CT alongside the primary answer, in a ready-to-sign draft the radiologist reviews before it becomes the final report.
Frequently asked questions
What did the new study find about missed breast lesions on chest CT?
A 2026 study in Academic Radiology reviewed 73 emergency-department chest CT scans from 64 patients who later had a biopsy-confirmed, BI-RADS 4 or 5 breast cancer visible in the field of view. About 70% of those scans had the breast finding missed or not reported at the time of the original read; roughly 30% were caught and reported.
Why are visible breast lesions missed on emergency chest CT?
Emergency chest CTs are ordered and read for a specific clinical question — chest pain, trauma, suspected pulmonary embolism — not to screen the breast tissue that happens to sit in the scan's field of view. The study found missed lesions were smaller on average (13.6 mm vs. 22.2 mm for detected ones), more often in the right breast, and less often accompanied by cues like axillary lymphadenopathy that helped detected lesions stand out.
What is the clinical cost of a missed incidental breast finding on CT?
In the study, a missed incidental breast lesion translated into a diagnostic delay ranging from 4 to 33 months before the cancer was eventually caught by other means. That gap represents real time lost between a scan that already showed the disease and the point of diagnosis.
How can radiology teams catch more incidental findings on chest CT?
The practical fix is structured, whole-scan review — evaluating every anatomic region in the field of view, including breast parenchyma, rather than treating anything outside the ordering complaint as out of scope. AI-assisted image analysis that reviews the whole volume, not just the organ tied to the primary indication, is one way to make that systematic check practical without adding to a radiologist's read time.
Source: O. Şahin, G. Deniz, "Beyond the Primary Complaint: Analysis of Missed and Incidentally Detected Breast Cancers on Emergency Chest CT Scans," Academic Radiology (2026), https://doi.org/10.1016/j.acra.2026.06.052, as reported by Radiology Business. Figures are rounded as reported.