Doctors don't always agree on breast biopsy results, and women with atypical hyperplasia (atypia) or ductal carcinoma in situ (DCIS) should seek second opinions, according to a new study published online in the Journal of the American Medical Association.
Researchers from Dartmouth-Hitchcock Medical Center found that the overall concordance rate among pathologists was only 75.3%.
"About 1.6 million breast biopsies are done every year in the U.S., yet in nearly half of the cases labeled atypia, our study indicates that there simply isn't 'concordance,' meaning there is not agreement with a standard reference diagnosis," said lead author Anna Tosteson, ScD, in a statement released by Dartmouth-Hitchcock.
For the study, the researchers asked a panel of three expert pathologists to determine the reference diagnosis for each of 240 cases. These consisted of 23 cases of invasive breast cancer, 73 cases of DCIS, 72 individuals with atypia, and 72 benign cases without atypia. Tosteson and colleagues then asked 115 pathologists from eight states to provide breast pathology interpretations of the cases (JAMA, March 17, 2015).
The 115 pathologists produced a total of 6,900 interpretations. The overall concordance rate was 75.3%, with a rate of 96% for biopsies with invasive carcinoma. The concordance rate for atypia was 48%, and for DCIS it was 84%, the researchers found.
"The take-home message for women is that it's not overly cautious to seek a second opinion with one of these diagnoses," said co-author Tracy Onega, PhD, in Dartmouth-Hitchcock's statement. "In fact, it's likely to be a good idea."
![A normal mammogram confirmed by three-year radiologic follow-up illustrates reader-marked regions of interest (ROIs) during (A) unaided (round 1) and (B) artificial intelligence (AI)–assisted (round 2) reading. Each colored dot represents an ROI for recall by a human reader. Readers could mark more than one ROI per case, represented by multiple dots of the same color. During AI-assisted reading, the AI system displayed three visible prompts: two with suspicion of malignancy scores of 35% (left mediolateral oblique [L MLO] and craniocaudal [L CC]) and one with a suspicion of malignancy score of 10% (right craniocaudal [R CC]), shown as polygonal overlays. Without AI, six of 10 readers (60%) marked a false-positive ROI. With AI assistance, this fell to two of 10 (20%). R MLO = right mediolateral oblique.](https://img.auntminnie.com/mindful/smg/workspaces/default/uploads/2026/07/2026-07-14-radiology-mammogram-ai-auto-bias.H0bYO8QlWs.jpg?auto=format%2Ccompress&fit=crop&h=100&q=70&w=100)







![A normal mammogram confirmed by three-year radiologic follow-up illustrates reader-marked regions of interest (ROIs) during (A) unaided (round 1) and (B) artificial intelligence (AI)–assisted (round 2) reading. Each colored dot represents an ROI for recall by a human reader. Readers could mark more than one ROI per case, represented by multiple dots of the same color. During AI-assisted reading, the AI system displayed three visible prompts: two with suspicion of malignancy scores of 35% (left mediolateral oblique [L MLO] and craniocaudal [L CC]) and one with a suspicion of malignancy score of 10% (right craniocaudal [R CC]), shown as polygonal overlays. Without AI, six of 10 readers (60%) marked a false-positive ROI. With AI assistance, this fell to two of 10 (20%). R MLO = right mediolateral oblique.](https://img.auntminnie.com/mindful/smg/workspaces/default/uploads/2026/07/2026-07-14-radiology-mammogram-ai-auto-bias.H0bYO8QlWs.jpg?auto=format%2Ccompress&fit=crop&h=112&q=70&w=112)









