Great post today at Linked In by Gary Gustavsen, Partner, Health Advances. Find it here.
The Linked In post is concise and primarily directs the reader to the publication, Aggarwal et al.
###
Are we at risk of oncologists viewing DigitalPathology as a black box? Check out our new publication with AstraZeneca, “Clinician Perspectives on Digital and Computational Pathology: Clinical Benefits, Concerns, and Willingness to Adopt”, published in Diagnostics MDPI. Download the PDF here: https://lnkd.in/eA3rBpje or drop a comment to chat further!
###
Find the open access article here.
https://www.mdpi.com/2075-4418/15/19/2527
###
Here is the authors' abstract:
Background/Objectives: Precision medicine has transformed how we manage cancer patients. As treatments and drug targets become more complex, the associated diagnostic technologies must also evolve to actualize the benefit of these therapeutic innovations. Digital and computational pathology (DP/CP) play a pivotal role in this evolution, offering enhanced analytical techniques and addressing workflow constraints in pathology labs. This study aims to understand clinicians’ awareness, utilization, and willingness to adopt DP/CP-based tools, as well as the role they perceive themselves playing in the adoption of CP-based tests.
Methods: A double-blinded, online quantitative survey was conducted among 101 U.S.-based medical oncologists.
Results: Awareness of DP/CP varied among clinicians, with only 17% identifying as very aware. Subsequently, the current utilization of CP-based tests is also low. Despite this, clinicians are optimistic about the potential benefits of DP/CP, including reduced turnaround times, improved therapy selection, and more consistent slide review. To achieve full adoption, clinicians recognize that barriers must be addressed, including cost, regulatory guidance and, to a lesser extent, concerns with the “black box” nature of CP algorithms. While the focus for the adoption of DP has centered on pathologists, clinicians anticipate playing a more significant role in the adoption of CP-based tests. Finally, clinicians demonstrated clear willingness to utilize a CP-based CDx, with 90% of respondents identifying as potential adopters.
Conclusions: This study highlights a positive outlook for the adoption of DP/CP among clinicians, despite varied awareness and low current utilization. Clinicians recognize the potential benefits of DP/CP but also acknowledge barriers to adoption. Addressing these barriers through education, regulatory approval, and collaboration with pathologists and biopharma is essential for successfully integrating DP/CP technologies into clinical practice.
###
See in context of a recent posting on dig path by Braxton:
https://www.discoveriesinhealthpolicy.com/2025/10/braxton-writes-digital-pathology-will.html
####
AI CORNER
####
Chat GPT 5 reads and discusses the Aggarwal paper.
###
This paper reports a double-blinded, online survey of 101 U.S. medical oncologists that probes awareness, current use, perceived value, and adoption thresholds for digital and computational pathology (DP/CP). Despite DP/CP’s growing technical maturity—framed here with recent examples such as quantitative continuous scoring (QCS) for HER2 and TROP2 NMR—the clinician side of the ordering equation remains early: only 17% of respondents described themselves as “very aware,” 25% had no awareness, and just 13% knowingly order a CP-based test today. Awareness skews higher in academic settings and where pathology is in-house; comfort is strongest for DP/CP as decision support (region finding and quantification ~3.6/5) and weaker for fully automated interpretation (~3.2/5), signaling that validation and education will matter most at the “automation” frontier. Clinicians nevertheless endorsed clear clinical utility: reduced turnaround time, improved therapy selection, and more consistent slide review topped the value stack.
Barriers map to the usual adoption economics and guardrails rather than to AI mystique: out-of-pocket cost, payer coverage, FDA clearance, and NCCN guideline inclusion dominate concerns; “black box” worries were comparatively modest, though a nontrivial minority (≈12% of non-users) expressed high concern. Consistent with those barriers, respondents judged FDA approval and guideline placement as the most decisive levers for scale. Importantly, the survey shifts the locus of adoption from the lab to the clinic for CP-based tests: a majority of current users viewed themselves as primary decision-makers, and ~70% of non-users expect to play a role, implying that evidence packages must speak directly to oncologist-centric endpoints (accuracy with scant tissue, workflow gains, and access to targeted therapy were the top adoption drivers).
Asked to consider a hypothetical FDA-approved CP-based CDx in lung cancer, 90% were at least somewhat willing to order it, but preferences were nuanced: many clinicians would accept either a traditional or CP-based assay, and >30% of current CP users would still prefer a traditional test if both existed—hinting at pain points in real-world ordering or reporting that future products must surface and fix. The authors conclude that accelerating DP/CP in oncology will require coordinated play: biopharma to generate peer-reviewed, clinically anchored evidence; pathologists to vet analytical rigor and operational fit; and regulators/guideline bodies to provide clear, confidence-building pathways. The study’s scope (medical oncologists only; modest N; self-report) and funding (AstraZeneca; Health Advances involvement) are transparent, but its take-home for the field is crisp: the clinician demand signal is present, and movement from “promise” to “default” hinges less on demystifying algorithms than on meeting coverage, labeling, and workflow realities with level-of-evidence commensurate to a CDx.
