My name is Stathis Kamperis, and I live in Greece. I am a radiation oncologist and physicist with a master degree in computational physics and another one in medical research methodology. I currently work on my Ph.D. thesis, which involves applying machine learning in radiomics in head neck cancer. You can check my detailed bio in LinkedIn.

Hands-on experience

⇝ Machine-learning model for virtual QA/verification

Keras logo Tensorflow logo DICOM logo SQLite logo

I developed an end-to-end machine-learning model for predicting quality assurance and verification outcomes in radiation oncology departments. It uses as input a diverse set of clinical and dosimetric features, such as complexity metrics, dose distributions, clinical and dosimetric traits. The extraction and calculation of many of these features is done by rteval, an evaluation framework I wrote in Python.

Machine learning model


⇝ Feature-engineering framework for radiation oncology treatment plans

Python logo Linux logo

I developed a specialized feature engineering framework for radiation oncology departments. Also, I wrote a web application with R Shiny for the statistical analysis and presentation of the extracted data.


⇝ DICOM viewer

Qt logo Opengl logo DICOM logo

I developed a cross-platform, OpenGL-accelerated, multi-threaded DICOM viewer (C++, OpenGL, Qt).

For more details, check this link.

Google Summer of Code

⇝ Profiling CERN’s Geant4 high energy physics simulation framework

Google logo CERN logo Geant4 logo Solaris logo Linux logo

For more details, check this link.


⇝ Auditing and extending NetBSD’s math library

Google logo NetBSD logo AMD64 logo

I audited and extended NetBSD’s operating system math library. Specifically:

For more details, check this link.


⇝ Auditing DragonFlyBSD’s POSIX/C99 conformance

Google logo DragonFlyBSD logo NetBSD logo Linux logo PostgreSQL logo

I audited DragonflyBSD operating system against the latest POSIX and C99 standards. Specifically:

Academic

⇝ Monte Carlo simulations in human phantoms (Master thesis in Computational Physics)

Geant4 logo Monte Carlo logo

You can read my first master thesis, where I used Geant4 to run Monte Carlo simulations of external photon beams in human phantoms. Regrettably, the text is in Greek, but you might find the images and the code excerpts useful.


⇝ Complexity analysis of VMAT prostate plans (Master thesis in Medical Research Methodology)

R logo Python logo Scikit logo Jupyter logo Mathematica logo

You may also read my latest master thesis, where I analyzed the complexity of Volumetric Modulated Arc Therapy prostate plans. In the first part, I examined various complexity metrics with Principal Component Analysis and Mutual Information Analysis. In the second part, I developed both a linear and a LASSO logistic regression model to predict complexity by clinical and dosimetric plan features.

This is the list of my publications on medical journals.