Breast Imaging: Less May Be More

Breast tomosynthesis has progressively displaced digital mammography in the U.S. and other regions.  As a company that supplies an AI-enabled cloud solution for medical imaging reading and reporting, we work with many medical practices and see the average screening breast tomosynthesis study ranging between roughly 260 and 750 images per exam.  Let’s do the math.  Given 3600 seconds in an hour, lets assume a radiologist can have eyes glued to the images for four hours per day, understanding time also must be spent reading the history, prior reports, current report, coding the exam, as well as consulting with technologists, schedulers, referring doctors, and patients (some radiologist are even allowed to blink and eat). That leaves 14,400 seconds for viewing images.

100 breast tomosynthesis exams might include 26,000 to 75,000 images. One can therefore see that leaves far less that a second per image. While some may argue that the tomo frames can be played at 20 frames per second as a cine, or perhaps faster, the fundamental reason that tomo frames are acquired is to find the cancer that may be visible on only one or a few frames.

Clearly, determining the “right” number of images per exam must assess conflicting factors, such as human fatigue, reading speed, time taken away from other valuable consulting tasks, and other items that all contribute to the overall accuracy and value of the exam.  As far back as 2020, Purjara et al (Radiology,, Vol. 297, No 3) reported that, “A digital breast tomosynthesis reconstruction protocol that uses 6-mm slabs with 3-mm overlap, without 1-mm slices, had similar diagnostic performance compared with the standard protocol and led to a reduced interpretation time for three of four readers.” Chumsaengsri et al replicated these findings as published in QIMS Vol 15, No 10, Oct 2025, as have a few other publications.

Multiple breast tomosynthesis vendors and a few PACS vendors now offer options for breast tomosynthesis slabs.  Perhaps someday AI will further improve efficiency by triaging out the negative exams or negative slices, but in the interim, it seems valid to conclude that reducing the number of images per exam is an effective way to improve efficiency without sacrificing accuracy.

If less images translate into more efficient imaging at equivalent accuracy in breast tomosynthesis, where else might this apply?

Murray A. Reicher, MD FACR

CEO Synthesis Health, Inc.