Photoacoustic Computed Tomography of Breast Cancer in Response to Neoadjuvant Chemotherapy

Abstract

Neoadjuvant chemotherapy (NAC) has contributed to improving breast cancer outcomes, and it would ideally reduce the need for definitive breast surgery in patients who have no residual cancer after NAC treatment. However, there is no reliable noninvasive imaging modality accepted as the routine method to assess response to NAC. Because of the inability to detect complete response, post-NAC surgery remains the standard of care. To overcome this limitation, a single-breath-hold photoacoustic computed tomography (SBH-PACT) system is developed to provide contrast similar to that of contrast-enhanced magnetic resonance imaging, but with much higher spatial and temporal resolution and without injection of contrast chemicals. SBH-PACT images breast cancer patients at three time points: before, during, and after NAC. The analysis of tumor size, blood vascular density, and irregularity in the distribution and morphology of the blood vessels on SBH-PACT accurately identifies response to NAC as confirmed by the histopathological diagnosis. SBH-PACT shows its near-term potential as a diagnostic tool for assessing breast cancer response to systemic treatment by noninvasively measuring the changes in cancer-associated angiogenesis. Further development of SBH-PACT may also enable serial imaging, rather than the use of current invasive biopsies, to diagnose and follow indeterminate breast lesions.

Keywords: breast cancer; neoadjuvant chemotherapy; photoacoustic computed tomography; response to treatment; tumor‐associated microvasculature.

Figure 1

Schematic of the SBH‐PACT and breast phantom imaging. a) Perspective cut‐away view of the SBH‐PACT breast imaging system. DAQ, data acquisition module; Pre‐amp, preamplifiers. b) Sketch of the breast‐mimicking phantom. c) Maximum‐amplitude‐projection (MAP) of the breast phantom image acquired by SBH‐PACT, which revealed the smallest tumor phantom (1‐mm diameter). The bright dots in the image background were caused by air bubbles embedded in the phantom.

Figure 2

Breast images of a patient treated with NAC. a) Depth‐encoded angiograms of the unaffected (right) breast acquired by SBH‐PACT at three time points. b) Vessel density maps overlaid on the maximum‐amplitude‐projections (MAPs) of the breast images in (a). c) Depth‐encoded angiograms of the affected (left) breast of the patient. d) Vessel density maps overlaid on the MAPs of the images in (c). Regions with higher vascular densities are highlighted.

Figure 3

SBH‐PACT images of the other two patients treated with NAC. a) Angiograms of Patient 2 who had a partial response to NAC. Close‐up views of the cancer‐affected area are shown from T1 to T3 in the region outlined by yellow dashed boxes (bottom right corner of figure). Breast cancer is identified by yellow arrows. Prominent common blood vessels in T1 and T2 images are marked by magenta arrow lines. b) Angiograms of Patient 3 who had a complete clinical response to NAC. Angiogenesis associated with the cancer was not detected at T3.

Figure 4

Breast cancer segmentation based on blood vascular irregularity. a) Entropy maps computed from the breast angiograms. b) Anisotropy‐weighted entropy maps, which suppress the structures in healthy tissues. c) Segmented maps of the cancer‐affected area extracted from (b).

Figure 5

Comparison of the images acquired by SBH‐PACT and MRI. a) SBH‐PACT angiograms weighted by Figure 4b. b) MRI images of the same breast with dynamic post contrast sequence acquired at T1 and T3 of the NAC treatment. Correlated structures are marked by magenta arrow lines. SBH‐PACT reveals more angiographic details within a single breath hold of 15 s.

Figure 6

Quantitative measurements of the tumors’ transformation during the NAC. Statistical data are presented as mean ± standard error; p‐values are calculated using one‐tailed Welch's (unequal variances) t‐test. a) Measurements of tumor dimensions and volume. LA, long axis; SA, short axis. b) Quantification of the relative blood vascular density in the ROI and healthy tissues. c) Quantification of the blood vascular entropy in the ROI and healthy tissues. d) Quantification of the blood vascular anisotropy in the ROI and healthy tissues.