MLI™ Slide Scanning Service
What We Offer
illumiSonics MLI™ provides a non-destructive, label-free scan of unstained FFPE tissue. Because the section is preserved, it comes back to you ready for H&E, IHC, or any other downstream testing. This enables direct, one-to-one paired training of every cellular and stromal feature: each nucleus, organelle, vessel, and fibre in the label-free image is matched with its stained counterpart on the identical section. This yields a unique dataset primed for developing AI disease classifiers, novel and companion diagnostics, and virtual stains (H&E, IHC, and multiplexed).

Built for AI Classifier Development and Novel Diagnostics
Conventional staining uses up the tissue it labels, and each stain shows only part of the picture: a general overview like H&E, or a single target like IHC. Every new question needs another section, another stain, and another source of lab-to-lab variability.
MLI™ is a new kind of microscope that measures how FFPE tissue interacts with light, with no stains, dyes, or labels. Each MLI™ scan captures six 40x-equivalent channels of molecular and morphological contrast, generating a comprehensive map of biomolecular absorption signatures. The tissue is left intact for whatever comes next.

Rich Data for Novel Diagnostics
-
Data no stain can capture. MLI™ measures how the tissue itself absorbs light, not how a dye binds to it. Nuclei, cytoplasm, collagen, elastin, red blood cells, and melanin are all captured in a single acquisition. That is a fundamentally different kind of data from H&E or any single-target stain.
-
No stain variability. Contrast comes from the tissue, not from reagents, so there are no batch effects, stain fading, or protocol differences between labs. AI models train on the biology rather than on staining variability.
-
Slides returned unaltered. Your section comes back ready for H&E, IHC, downstream testing, or archiving. One specimen can support classifier development, biomarker work, and companion diagnostics.
-
Perfectly paired training data. Scan first, then stain the same section. Every cell in the MLI™ image matches its stained counterpart, so you can train and validate classifiers, or virtual H&E, IHC, and special stains, with no serial-section mismatch.
Our Scanning Process
-
Request a Quote - Tell us about your sample and scanning needs, and we will provide a quote
-
Send Samples - Ship or drop off unstained FFPE slides or blocks
-
Scanning & QA - We will scan, digitize, and ensure high quality images
-
Delivery - You will receive image files via secure link and slides/blocks will be returned to you
Technical Specifications
Sample Requirements
-
25x75mm microscope slides
-
Unstained FFPE tissue, without coverslip or deparaffinization
Output Specifications
-
High resolution whole slide images (40x)
-
6 channel multiplexed MLI™ contrast
-
Image delivery in OME-TIFF is standard
Interested in our services?
Slide scanning available today. Beyond scanning as a service, illumiSonics will be your partner in AI development and support rollout of MLI™ devices paired with custom algorithms into research and commercial placements.
Service Details:
-
Turnaround Time: typically 5-10 business days
-
Pricing: available upon request - contact us for a quote
Interested in our products?
Our microscope platform MLI 1 is now available for purchase or lease. Realize the potential of label-free MLI™ in your own lab, with six channels of molecular and morphological contrast from a single scan of unstained FFPE, fresh frozen, or bulk tissues. Contact us today to learn more about configurations and pricing.


Why choose illumiSonics?
-
A new class of microscope. MLI™ is a patented platform developed by our team and protected by 40+ patents in 12 families. No other slide scanner captures this kind of contrast.
-
From study design to validated model. Our team of optical engineers, AI researchers, and clinicians help plan your study, scan your samples, and train and validate custom models to meet your needs.
-
Global reach. We accept samples from anywhere in the world.
Peer-Reviewed Evidence
-
Breast cancer (needle core biopsies). Seven breast pathologists found 98% agreement on cancer vs. benign between MLI™ virtual H&E and chemical H&E, and could not tell the two apart beyond chance. (Current Oncology, 2023)
-
Skin cancer (SCC and BCC excisions). Seven dermatopathologists found 95.5% primary diagnostic agreement. Diagnostic confidence was statistically equivalent to H&E. (Diagnostics, 2026)
-
Virtual multi-staining. H&E, Masson's trichrome, PAS, and Jones' silver were generated from a single label-free scan. Pathologists rated diagnostic quality as equal to chemical stains and could not reliably tell real from virtual. (NPJ Imaging, 2026)







