Investigation of feasibility and diagnostic capability of ultrasound-guided photoacoustic tomography for breast surgical margins.

We explored the feasibility and preliminary diagnostic capability of ultrasound-guided photoacoustic (US-PA) tomography for breast cancer margin assessment by analysing PA derived lipid, collagen, and hemoglobin distribution patterns. Fifty-three ex-vivo specimens from breast-conserving surgeries were imaged using a handheld US-PA imaging probe illuminated between 700 and 1100 nm across radial margins. Photoacoustic patterns were analysed and assessed for margin involvement using a proprietary algorithm and compared with standard of care against histopathology as the gold standard. US-PA imaging demonstrated 95.45% sensitivity and 76.33% specificity for margin assessment with a positive predictive value (PPV) of 34.43.% and negative predictive value (NPV) of 99.23%. Clinical standard of care showed 40.91% sensitivity, 88.17% specificity, 31.03% PPV and 91.98% NPV. Analysis of false positives revealed that 75% occurred within 5 mm of the margin, with 35% within 2 mm. Overall, US-PA imaging demonstrated superior sensitivity, specificity, PPV, and NPV compared to standard approaches, supporting its potential to reduce missed positive margins. The high false positive rate, primarily due to close margins and tissue compression effects, reflects a conservative classification strategy. Detection of biochemical markers, particularly collagen highlights potential of US-PA in enhancing margin assessment accuracy, although further refinement is needed to improve specificity.