PapSmearAI
Explainable AI for Cervical Cytology Research and Quality Improvement

PapSmearAI is an early-stage research platform developing AI-assisted cervical cytology image classification, human-in-the-loop annotation, and explainable decision support for safer screening workflows, quality improvement, and future NHS validation.

Contact / Collaborate

Research-use concept only. Not intended for clinical diagnosis or patient management.

The Challenge

Cervical screening services face increasing workload pressures, diagnostic variability, and growing demand for safe quality assurance pathways.Manual cytology review remains highly dependent on expert interpretation, while workforce shortages and rising screening demands increase operational pressure across laboratories.There is a strong need for explainable AI tools that support—not replace—clinical expertise, helping improve consistency, triage efficiency, and future validation pathways to avoid false negative results and provide early diagnosis.

Our Solution

PapSmearAI combines AI-assisted cytology image classification, explainable nucleus-focused (metric)analysis, and human-in-the-loop review to support safer and more consistent screening workflows.In addition to image-based evaluation, the platform integrates relevant patient background information, epidemiological risk factors, HPV status, and cancer genomics biomarkers to provide stronger clinical context for cytologists and support more accurate, precise, and explainable decision-making.Rather than replacing expert interpretation, PapSmearAI is designed to strengthen quality assurance, improve triage efficiency, enhance annotation consistency, and support future NHS validation pathways.

✓ Explainable AI + nucleus-focused cytology analysis✓ Multimodal integration of clinical, epidemiologic, HPV, and genomics data✓ Human-in-the-loop validation for future NHS adoption

Evidence Base

PapSmearAI is built on standardized cytology annotation guidelines designed to ensure consistency, reproducibility, and explainable AI training.Each image is classified using a single-label approach with one dominant cell per image, prioritizing nuclear morphology as the primary diagnostic feature. When reliable classification is not possible, an uncertainty class (artifact_uncertain) is used to protect dataset quality and reduce annotation bias.This structured framework supports robust model training, safer validation pathways, and stronger clinical trust for future translational use.

✓ Single dominant cell classification✓ Nuclear morphology-driven decisions✓ artifact_uncertain class for safer AI training✓ Human-reviewed annotation consistency

Funding Goal

PapSmearAI is seeking grant funding and strategic collaboration to support the next stage of research validation and translational development.The immediate goal is to strengthen clinical evidence, expand annotated datasets, improve multimodal validation, and prepare the platform for future NHS pilot studies and regulatory pathways.This project is positioned as a research-first innovation focused on quality improvement, precision diagnostics, and safer cervical screening support.

Funding priorities include:✓ Retrospective clinical validation✓ Expansion of annotated cytology datasets✓ Integration of HPV, biomarkers, and genomics data✓ Cloud-based prototype development✓ NHS pilot preparation and regulatory roadmap

Founders

Dr. Mehrdad Salehmanesh

Cytologist with expertise in molecular medicine, cancer genomics, and translational healthcare innovation.Focused on diagnostic pathology, biomarker interpretation, cancer research, and AI-driven medical workflow improvement, with a strong interest in building safe, explainable, and clinically relevant tools for precision diagnostics and cytology quality improvement.The clinical vision of PapSmearAI is centered on supporting—not replacing—expert cytology decision-making through responsible healthcare AI development.

Connect on LinkedIn

Mr.Houman Mansour

Data Engineer and transformational technical program and product manager with a strong track record of leading complex programs, enterprise system development, integrations, business process re-engineering, and large-scale migrations across the financial services industry.With extensive experience supporting capital markets, brokerage firms, mutual funds, wealth managers, portfolio managers, and asset managers, he brings deep expertise in data architecture, technical delivery, product strategy, and operational transformation.Within PapSmearAI, he supports platform development strategy, data infrastructure planning, scalable system design, and the translational pathway required to move early-stage healthcare innovation toward robust, real-world implementation and long-term growth.