Mohammad Imran Hossain

← Back to research overview

RESEARCH FOCUS

Personalizing breast cancer screening.

Current doctoral research and a selected master’s thesis in computational pathology.

CURRENT DOCTORAL RESEARCH

Can we make breast cancer screening more personal?

At the Department of Medical Epidemiology and Biostatistics, Karolinska Institutet , I develop and evaluate mammography-based AI models for short-term breast cancer risk prediction and study how a negative breast MRI informs subsequent risk.

As Doctoral Candidate DC09 in the EU-funded HER-CARE MSCA Doctoral Network , I investigate who may benefit from supplemental imaging and when longer screening intervals may be appropriate. The planned work includes external validation across Swedish and US screening populations.

Explore HER-CARE ↗

THE RESEARCH QUESTION

Mammography Breast MRI Clinical context
↓
AI-driven risk assessment
↓
More personalized screening

Research direction · clinical benefit to be evaluated

Doctoral workflow, imaging examples & credits

THE RESEARCH PIPELINE

From breast imaging to future risk.

Illustrative imaging examples and a conceptual modelling workflow.

01 / IMAGING & CONTEXT

Published four-panel figure comparing standard and contrast-enhanced mammography; panels A and B show standard mammograms.
Mammography Standard & contrast-enhanced views
Published four-panel breast MRI example with contrast enhancement, diffusion-weighted images, and an ADC map.
Breast MRI Complementary imaging information
Age & family history Breast density Clinical & genetic risk factors

Single examinations and longitudinal mammograms, where available.

02 / LEARN & INTEGRATE

Deep learning & multimodal modelling

Extract image representations and combine them with clinical risk factors.

Image-derived features Clinical feature integration Time-to-event risk modelling

Evaluate discrimination, calibration, and performance across populations.

03 / PREDICT & EVALUATE

Short-term breast cancer risk

Baseline 1 year 3 years 5 years
Higher risk Who may benefit from supplemental imaging?
Lower risk When might longer screening intervals be appropriate?
After a negative MRI How does subsequent risk change over time?

Research aims, not individual predictions or clinical recommendations.

Conceptual overview of planned research; no model outputs or study results are shown. MRI and mammography examples depict published cancer cases, not negative screening examinations or participants in my PhD cohorts. Modalities illustrate complementary research streams; MRI is not assumed to be available for every participant.
Image sources & credits +

Mammography: Moustafa AFI et al. (2018), reproduced as Figure 3 in Novel Technologies in Breast Imaging: A Scoping Review . CC BY-NC-ND 4.0 ; figure reproduced unchanged.

Breast MRI: Camps-Herrero J. (2019), Diffusion-weighted imaging of the breast , reproduced as Figure 4 in the same review. CC BY 4.0 ; figure reproduced unchanged. Neural network and risk timeline are explanatory schematics created for this site.