Shevs Connect | Full-Width Navigation
Mon – Fri: 08:00am – 05:00pm Magen Plaza, Juja – Along Thika Road, next to Juja Flyover. info@shevsconnectinstitute.com +254 780 253160

Pathology AI

SCILabel | Medical Imaging & Radiology AI

Pathology AI

Training AI to read the cellular landscape of disease

Industry Challenge | SCILabel

Industry Challenge

Digital pathology is undergoing a transformation. Whole-slide imaging (WSI) technology now generates gigapixel images of tissue biopsies, enabling AI to assist with tumour grading, cell counting, margin assessment, and biomarker quantification. The challenge is annotation: pathology AI models require thousands of annotated whole-slide images labeled by qualified pathologists or trained pathology technologists — a workforce that is scarce and expensive in most markets.

Radiology AI Challenge
How SCILabel Serves This Industry | Radiology AI

How SCILabel Serves This Industry

Data Collection

We source whole-slide image datasets from pathology laboratory partners and oncology research centres globally. Data types include H&E-stained slides, immunohistochemistry (IHC), fluorescence microscopy, and cytology specimens. All samples are de-identified with full chain-of-custody documentation.

Data Annotation & Labeling

Our specialist pathology annotators — medical laboratory scientists, biomedical scientists, and clinical officers with pathology training — annotate whole-slide images at the region-of-interest, tissue, and cell level. Annotation types include tissue classification, tumour boundary delineation, mitosis detection, cell-level instance segmentation, and grading (e.g., Gleason grading for prostate cancer). We use specialised WSI viewers and support all major formats (SVS, NDPI, MRXS, TIFF).

Data & Model Evaluation

Expert evaluators assess model performance on tumour grading accuracy, mitotic figure detection precision and recall, and false-positive rates on benign tissue. We conduct comparative evaluations against pathologist ground truth and published grading standards.

Annotation Types & Formats

  • Region-of-interest annotation on whole-slide images
  • Tumour boundary polygon segmentation
  • Cell-level instance segmentation (mitotic figures, lymphocytes, plasma cells)
  • Tissue classification (tumour, stroma, necrosis, normal)
  • Immunohistochemistry scoring (e.g., HER2, PD-L1 expression)

Specialist Workforce Tracks

Track 2 (Medical Imaging): with pathology specialisation: Medical Laboratory Scientists, Biomedical Scientists, Histology Technologists.

Example Deliverable | SCILabel

Example Deliverable

Client Output Example
An annotated prostate biopsy dataset of 800 whole-slide images, with Gleason grade annotations (primary + secondary pattern), tumour percentage estimation, and perineural invasion flagging — delivered with QA certification and inter-annotator Kappa of 0.87.