Breast MRI is effective for measuring breast tumors before a woman undergoes surgery, according to a study published June 16 in the Annals of Surgical Oncology.
"[We found that] MRI demonstrated the highest concordance with tumor size and T stage," wrote a team led by Abigail Daly, MD, of Massachusetts General Hospital in Boston.
Accurate preoperative imaging of breast tumor size is essential, as small measurement differences can influence the treatment strategy, the researchers noted. They investigated the accuracy of tumor size estimation by mammography, ultrasound, and MRI compared with pathology results.
Their study included 460 patients with breast cancer who were treated between 2019 and 2024. The team considered imaging measurements concordant if they fell within ± 20% of the pathological size of the tumor. Average tumor size on pathology was 17.3 mm.
Overall, the group found that breast MRI had the highest concordance with pathology results for measuring tumor size.
Tumor size estimation by imaging modality | |
Modality | Concordance with pathology results |
| MRI | 62% |
| Mammography | 57% |
| Ultrasound | 53% |
The researchers also found that MRI slightly overestimated tumor size (18.4 mm), while mammography and ultrasound underestimated (14.8 mm and 14.3 mm, respectively). They reported the following:
- Of all three imaging modalities, MRI showed the highest accuracy in T-stage classification (89%).
- MRI's concordance with pathology was highest for masses without nonmass enhancement.
- The modality's accuracy improved in tumors greater than 15 mm (odds ratio [OR], 2.47, with 1 as reference) and high-grade tumors (OR, 1.75), but declined in extremely dense breasts (OR, 0.42).
Despite MRI's performance, using other modalities with it to assess tumor size before surgery could be a good idea, according to the authors.
"A combined imaging approach using MRI and ultrasound may enhance preoperative size estimation," they concluded.
The complete study can be found here.

![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)









