ABSTRACT
Pathology, which plays a vital role in clinical diagnosis, faces numerous challenges that impact its efficacy. For
example, resected specimens often require preparing and analysis of as many as 30-40 slide blocks under a
microscope until the disease is confirmed; selection of slices for slide preparation uses subjective methods such
as palpation, which depend greatly on the skill of the individual performing the assessment and introduces
inconsistency in the clinical process; and for each slide block examined, analysis and annotation requires manual
observation of every microscopic region of the tissue. As a result, most pathology evaluations often take 1-3
weeks to analyze samples and reach a conclusion regarding potential cancers. An added challenge is that
insurance reimbursements are capped per case regardless of the number of slide blocks processed, with any
additional costs being absorbed by the hospital. Consequently, hospitals must balance the trade-off between
minimizing the number of slices (for economic viability) and not compromising diagnostic care. These challenges
affect not only clinical pathology but also research involving pathology specimens and tissue selection for
biobanking. There is a critical need to eliminate subjectivity, reduce pathologists’ workload, and increase
throughput in histological analysis. We propose to meet this need by developing a new technology called X-ray
diffraction imaging (XRDI), which can scan any number of surgically resected, sliced pathology specimens and
automatically indicate the likelihood and location of disease in each slice within minutes. In collaboration with
Duke University, we previously built a research prototype XRDI system and demonstrated its utility by scanning
and evaluating 300 breast cancer slices with high accuracy and resolution. In this direct-to-phase-II SBIR
application, we will now construct a new clinical version of the XRDI scanner that is affordable, reliable, and
accurate, and can be directly integrated into the clinical pathology workflow. We will build the scanner, test and
evaluate its performance, and demonstrate its utility through field-testing in collaboration with clinical pathology
laboratories in the US. This project will provide a first-of-its-kind, commercially feasible XRDI scanner for rapid,
non-destructive imaging of pathology specimens with the ability to inform pathologists about the presence and
location of cancer within the different tissue slices. The proposed clinical scanner will enable: 1) analysis of the
whole slice volume of the specimen rather than a few microns at the surface of a subset of the slices, which is
the current standard of care process using microscopy, 2) quantitative identification of disease based on XRD
information obtained directly from the tissue, and 3) slice selection based on quantitative, reproducible
measurements, thereby eliminating user-related subjectivity. Importantly, it would significantly speed up
pathology workflow, providing decisions within hours instead of days, and improve the productivity and
profitability of pathology labs by reducing the number of slide blocks analyzed per case.
