Abstract artistic rendering of immunohistochemistry-stained tissue with rich brown DAB-positive cells on blue-purple hematoxylin background

IHC Biomarker Scoring

Quantify HER2, PD-L1, and Ki-67 in 90 Seconds

Synthia analyzes whole-slide biopsy images automatically — delivering reproducible IHC scores before the next case is staged.

synthia_report_v2.4 — WSI_00312
HER2 classification 2+ (Equivocal)
Membrane score (H-score) 94 / 300
IHC-positive cells 73.4%
PD-L1 TPS 38%
Assay (22C3 clone) CPS = 12
Ki-67 proliferation index 42.1%
Hotspot region 61.7% (max)
Confidence High
Analysis time 00:01:23
3 Biomarkers Scored HER2 · PD-L1 · Ki-67
<90s Per Whole-Slide Image vs. hours of manual annotation
±4% Inter-Reader Variance Internal validation, n=200 WSI

Manual IHC Scoring Is the Diagnostic Bottleneck

Immunohistochemistry scoring is essential for HER2, PD-L1, and Ki-67 treatment eligibility decisions — yet it remains one of the most subjective, time-intensive steps in the oncology diagnostic workflow. A single breast cancer biopsy batch can require 30–90 minutes of manual annotation per pathologist.

The consequences are real: inter-pathologist variability of 15–25% on HER2 equivocal cases, delayed treatment decisions, and burned pathologist time that should be spent on diagnostic interpretation — not counting brown cells.

Synthia was built for one purpose: automate the quantification so pathologists can focus on the judgment.

How Our Model Works

IHC DIAGNOSTIC WORKFLOW

Biopsy Sample Tissue Acquisition IHC Staining DAB chromogen + hematoxylin WSI Scanning NDPI / SVS / MRXS Manual IHC Scoring ← BOTTLENECK 30–90 min per slide set · High variability Inter-observer Kappa 0.60–0.75 Pathologist Report Clinical interpretation + sign-off Treatment Decision Oncology team review

How Synthia Works

Three steps from WSI upload to structured biomarker report.

01

Upload WSI

Drag-and-drop NDPI, SVS, or TIFF whole-slide images — or connect via your PACS or LIS. Synthia accepts the major scanner formats from Hamamatsu, Leica/Aperio, and 3DHISTECH.

02

AI Scores Tissue

Synthia's deep learning model localizes IHC-positive cells, quantifies DAB staining intensity, and calculates a standardized score per ASCO/CAP guidelines — automatically, without human annotation.

03

Pathologist Reviews Report

Receive a structured PDF and structured data report in under 90 seconds. Review the quantified scores, annotate any disagreements, and sign off in one click. You interpret. We quantify.

Concordance Data Pathologists Can Evaluate

Every AI tool for pathology should publish its validation methodology and concordance statistics — not just say "AI-powered." Synthia's internal validation studies were designed with 3-pathologist reference reads and blinded AI-vs.-expert concordance testing.

Full Validation Data →
0.94 HER2 ICC
intraclass correlation
0.88 PD-L1 κ
weighted kappa (TPS)

Internal validation studies. Data on file. For investigational use only.

Works With Your Existing Digital Pathology Stack

Accepts standard WSI formats from major scanner platforms. REST API for LIS/PACS integration.

.svs
Aperio / Leica
.ndpi
Hamamatsu NanoZoomer
.mrxs
3DHISTECH
BigTIFF
Universal WSI
DICOM-WSI
Standard export
REST API
LIS / PACS / EHR
View All Integrations →

See Synthia Score Your Slides

Send us a de-identified WSI set and receive a full validation report within 48 hours. No setup required.

Request Pilot Access