HEDOS : Hematological Dose

Installation

Checkout the python source code

$ git clone https://github.com/mghro/hedos.git

Install dependent packages

$ pip3 install -r requirements.txt

Background

Blood flow is stochastically modeled and superimposed with a - potentially time varying - dose rate to calculate the blood dose distribution.

This code is a revised version of HEDOS. The old code can be found in the branch named "hedos_old". The main differences are:

• The new code is orders of magnitude faster thanks to vectorization.
• We have corrected a mathematical inaccuracy in the way transition probabilities between compartments were computed.

For blood dose calculations, the configuration parameters (patient, treatment and simulation parameters) are set up in BloodDose. It then calls one of two possible workflows:

• BloodDoseFromFields calculates the blood dose given the patient's dose distribution and the segmentations of the organs which are contributing to the overall blood dose. For the example here we have used a mesh reference phantom (ICRP Publication 145) and create an artificial sample dose. For patient-specific blood dose calculations, this should be replaced with patient data.
• BloodDoseFromDVHs does the same thing, but uses DVHs of each of the organs contributing to the overall blood dose.

The calculation of blood dose follows these steps in succession:

• FlowModel: Set up a graph that reflects the connectivity and magnitude of blood flow between blood compartments. Convert this into a matrix of transition probabilities.
• TemporalDistribution: Simulate the blood flow over time. Blood particles flow through the model by a stochastic jumping process based on survival analysis.
• CompartmentDose: Accumulate dose in the blood particles over time, based on a time-dependent dose rate.

Their relationship looks like following image:

Example

Below example output of the fractional blood dose distribution for an imaginary thorax patient who received lung RT with a fraction dose of 2 Gy. Top panel: Treatment given in 4 beams of 20s each, temporally separated by 10s to allow for gantry motion. Bottom panel: Treatment given in a single beam of 10s. It is clear that changing the dose rate quite dramatically affects the blood dose distribution (but not the mean blood dose). Non-zero contributions to the blood dose from various anatomical structures are also indicated.

Team

• Chris Beekman
• Jungwook Shin
• Stella Xing
• Lucas McCullum
• Clemens Grassberger
• Harald Paganetti

Acknowledgements

This work was supported by:

• R21 CA248118 : A Computational Method to Calculate the Radiation Dose to Circulating Lymphocytes
• R01 CA248901 : Developing whole-body computational phantoms for blood dosimetry to model the impact of radiation on the immune system

Publications

• Beekman C, et al. A stochastic model of blood flow to calculate blood dose during radiotherapy. Phys Med Biol. 2023;68(22):10.1088/1361-6560/ad02d6.
• Shin J, Xing S, McCullum L, et al. HEDOS-a computational tool to assess radiation dose to circulating blood cells during external beam radiotherapy based on whole-body blood flow simulations. Phys Med Biol. 2021;66(16):10.1088/1361-6560/ac16ea. Published 2021 Aug 3. doi:10.1088/1361-6560/ac16ea