From research algorithms to routine AI biomarker scoring
AI biomarker scoring is the use of trained algorithms to detect, quantify, and interpret disease-related signals in medical images, such as cancer markers on pathology slides or patterns in lung scans, so that clinicians can make more consistent, data-driven diagnostic and treatment decisions at scale. For years, this field sat mostly in research labs, limited by fragmented software, experimental data, and regulatory uncertainty. That is changing as imaging hardware vendors and cancer pathology AI specialists build integrated, clinical-grade workflows. Whole-slide scanners, browser-based viewers, and clinical imaging software are now being packaged with validated AI biomarker tools, transforming digital pathology from an optional innovation into an operational decision layer. At the same time, lung imaging platforms are acquiring AI startups to fold advanced detection and characterization directly into commercial offerings, signaling a shift from pilots to deployed, reimbursable services.
Leica, Indica Labs and Lunit tighten the PD-L1 workflow
Leica Biosystems, Indica Labs, and Lunit have formed a strategic alliance that puts AI biomarker scoring directly inside a cleared clinical digital pathology workflow. Their first release, Lunit SCOPE PD-L1 CAL10 NSCLC, is a PD-L1 detection automation algorithm tuned for Leica’s CAL10 antibody and deployed through the Aperio AI Store. The algorithm slots into a pipeline that joins Leica’s Aperio GT 450 DX scanner with Indica Labs’ browser-based Aperio HALO AP DX image-management software, giving pathology labs a single, end-to-end environment. Rather than manually estimating tumor cell percentages under a microscope, pathologists receive a quantitative PD-L1 score to confirm, reducing inter-observer variability that can alter treatment selection thresholds at 1% and 50%. According to Leica Biosystems, the aim is to “read the slide once and read it the same way everywhere,” standardizing PD-L1 assessment across sites and study cohorts.
An open clinical imaging software store for cancer pathology AI
Leica’s strategy is to own the full digital pathology stack while still attracting best-of-breed cancer pathology AI. The company holds a significant equity position in Indica Labs and distributes a clinically validated version of HALO AP for primary workflows, from tissue sectioning and staining through whole-slide visualization. On top of this, the Aperio AI Store acts as an application layer where third-party algorithms such as Lunit SCOPE, MindPeak’s tools, or bespoke pharma assays can plug into the same standardized pipeline through SDKs and APIs. This model turns clinical imaging software into an extensible platform rather than a fixed product. Developers gain access to Leica’s installed base via a revenue-share model, while labs gain a menu of validated AI biomarker scoring tools that can be deployed without rebuilding infrastructure, speeding the move from research to regulated clinical use.
4DMedical and contextflow scale AI lung imaging in clinical practice
In thoracic imaging, 4DMedical’s agreement to acquire Vienna-based contextflow marks a similar shift toward integrated AI in routine care. 4DMedical is known for its CT:VQ platform and software-based functional lung imaging; contextflow contributes AI that helps radiologists identify and characterize lung diseases and cancer on CT scans. The deal gives 4DMedical CE-marked products, existing clinical customers, and a local team, creating an immediate commercial and clinical platform and expanding its portfolio into AI-assisted lung cancer screening. It also brings access to established reimbursement pathways, shortening the timeline for clinical deployment of AI lung imaging across hospitals and screening programs. With contextflow CEO Markus Holzer stepping in as General Manager of Europe, 4DMedical gains regional leadership and relationships that can anchor wider adoption of AI-supported respiratory imaging in day-to-day radiology workflows.
Consolidation and the road to reimbursed computational diagnostics
These moves sit within a broader wave of consolidation in clinical imaging software and cancer pathology AI, as vendors race to secure scale, geographic reach, and regulatory-ready portfolios. Digital pathology adoption is still early; in a 2024 survey cited by Leica, only 28% of practice leaders reported digitizing slides with whole-slide imaging, and 10% said pathologists used remote sign-out for primary diagnosis. Yet interest is rising as computational companion diagnostics inch toward reimbursement, promising to connect AI biomarker scoring directly to drug selection. Strategic partnerships such as Leica–Indica–Lunit, and acquisitions like 4DMedical–contextflow, are designed to be ready for that moment: integrated, regulated workflows where AI performs the first read, and clinicians arbitrate. The emerging pattern suggests that the path to clinical scale will be paved more by platform ownership and alliances than by stand-alone algorithms.
