In mammographic screening, the breast is compressed between a compression paddle and the breast platform during x-ray imaging, to minimize patient motion, spread out overlapping structures within the breast and reduces the path of the x-rays through tissue, increasing image contrast and reduces the required x-ray dose. However, excessive compression can lead to significant patient pain, which can affect participation in screening. Despite this, mammography technicians are often trained to apply the maximum tolerable compression, because insufficient compression is well established to reduce clinical performance of mammography. Furthermore, applied force alone may not accurately reflect how compressed the breast is, because it does not account for variable breast size; compression pressure (the ratio of force to the contact area between the breast and the compression paddle) is thus a better measurement.
This study sought to investigate the effect of compression pressure on mammography performance in the Dutch screening program. In addition, it accounted for the confounding factors of volumetric breast density (which is knowns to also affect performance) and breast volume. 132 776 exams from 57 179 women were binned into quintiles, based on compression pressure as measured by Volpara (with thresholds at 7.7, 9.3, 10.8 and 13.0 kPa). Volpara was also used to measure the total breast volume (BV) and volumetric breast density (VBD). Compression pressure was then related to measures of mammographic performance (Recall rate, false positive rate, screen detected cancer rate, specificity, positive predictive value (PPV), interval cancer rate, sensitivity (24 months; 12 months)) and generalized estimating equations were used to adjust for BV and VBD.
It was found that exams in the lowest quintile of compression pressure had significantly lower specificity, compared to higher compression pressure. This arises due to the tendency of low compression to be associated with more recalls and false positives.
Sensitivity tended to decrease with increasing pressure—when estimating sensitivity for 12-month intervals, the highest compression pressure had significantly lower sensitivity (compared to intermediate compression). This finding is in line with recent results that have been published from the Norwegian Breast Screening Program.
The low specificity seen with low compression likely arises because insufficient compression is applied to separate overlapping structures within the breast, which may trigger a recall. But why does high compression reduce sensitivity? At this point, there is no clear answer, though several possibilities exist. One is that soft tumours spread out and lose contrast upon excessive compression, making them harder to detect. Pressure could also affect how much architectural distortion a lesion creates—a hallmarks sought by radiologists. Finally, it’s hypothesized that very high compression could reduce the blood flow within the breast, leading to a reduced lesion contrast, and hence making lesions appear less suspicious. This would be an area for further investigation.
There are few official guidelines for mammographic compression, but this study highlights the fact that compression pressure is an important aspect in optimizing performance. This necessitates accurate measurement of compression pressure in large numbers of women to allow technologists to train to compress to an optimal pressure. Volpara automatically generates this output after each breast exam. Furthermore, VolparaEnterprise keeps track of all compressions performed at a breast centre. This can allow managers to assess patterns of technologists’ compression behaviours and identify training opportunities.
