Dose Rate Dependency Correction for ArcCheck Type 3 Diodes

Documentation: https://kkaan.github.io/drdcorrectionforAC/

Overview

ArcCheck devices, particularly those utilizing Type 3 diodes, exhibit a significant dose rate dependency. This dependency can be problematic when measuring dose for Volumetric Modulated Arc Therapy (VMAT) delivered by Elekta linear accelerators. This repository provides tools to implement a correction methodology based on the findings published by Jäger et al. (2021) to mitigate these dose rate effects.

The primary goal is to correct the raw accumulated counts from the ArcCheck device to provide a more accurate representation of the delivered dose. This involves applying corrections for pulse rate (PR) and dose per pulse (DPP), followed by an optional intrinsic correction.

Dose Correction Methodology

The correction process implemented in src/corrections.py follows the methodology proposed by Jäger et al. It involves two main correction factors: one for pulse rate dependency and another for dose per pulse dependency.

1. Preliminary Calculations

Before applying the Jäger corrections, several steps are performed:

1. Calculate Instantaneous Counts: The instantaneous count (or count rate, proportional to dose rate) is determined by taking the difference between successive accumulated count readings from the counts_accumulated_df.

   count_df = counts_accumulated_df.diff()
   

   The first row of count_df will be NaN as a result of the diff() operation and is subsequently dropped.

2. Background Subtraction: Background radiation values, obtained from the bkrnd_and_calibration_df (typically from an .acm file), are subtracted from the count_df.

   background_values = bkrnd_and_calibration_df['Background'].values.astype(float)
   # ... (handling reference detector value)
   count_df = count_df.subtract(background_values_series, axis='columns')

3. Calibration Scaling: The background-subtracted counts are then scaled using calibration factors, also obtained from bkrnd_and_calibration_df.

   calibration_values = bkrnd_and_calibration_df['Calibration'].values.astype(float)
   # ... (handling reference detector value)
   count_df = count_df.multiply(calibration_values_series, axis='columns')

   This count_df now represents the calibrated dose rate per measurement interval for each detector.

2. Jäger Correction Factors (JCF)

The core of the correction lies in applying the Jäger Correction Factors. These factors are calculated separately for pulse rate and dose per pulse dependencies. The general form of the Jäger correction factor (JCF) is:

JCF = c - a * exp(-b * D)

Where:

• D is the calibrated dose rate (represented by count_df in the code after background subtraction and calibration).
• a, b, and c are coefficients specific to either the pulse rate or dose per pulse correction.

The corrected dose rate (D_corrected) is then calculated as:

D_corrected = D / JCF

a. Pulse Rate (PR) Correction

This correction addresses the dependency of the diode response on the pulse repetition frequency of the linear accelerator.

• Coefficients (as defined in apply_jager_corrections):

  • a_pr = 0.035
  • b_pr = 5.21 * 10^-5
  • c_pr = 1

• Formula Applied (pulse_rate_correction function): JCF_pr = c_pr - a_pr * exp(-b_pr * count_df) pr_corrected_count_df = count_df / JCF_pr

The function then returns the sum of pr_corrected_count_df for each detector, representing the total corrected dose after PR correction.

b. Dose Per Pulse (DPP) Correction

This correction addresses the dependency of the diode response on the dose delivered by each individual radiation pulse.

• Coefficients (as defined in apply_jager_corrections):

  • a_dpp = 0.0978
  • b_dpp = 3.33 * 10^-5
  • c_dpp = 1.011

• Formula Applied (dose_per_pulse_correction function): JCF_dpp = c_dpp - a_dpp * exp(-b_dpp * count_df) dpp_corrected_count_df = count_df / JCF_dpp

The function then returns the sum of dpp_corrected_count_df for each detector, representing the total corrected dose after DPP correction.

3. Combining Corrections

The apply_jager_corrections function calculates both pr_corrected_count_sum and dpp_corrected_count_sum. These represent two independently corrected versions of the total dose. The current implementation returns both in a DataFrame and subsequently an array, but typically one of these (or a combined/further processed version) would be chosen for final analysis.

4. Intrinsic Correction (Optional)

An additional correction, termed "Intrinsic Correction," can be applied. This correction is derived from the device's own calibration or reference data.

• Calculation (get_intrinsic_corrections function): The intrinsic correction factor for each detector is calculated as: Intrinsic_Correction_Factor = Corrected_Counts / Raw_Counts This is performed element-wise, excluding header/footer rows and specific columns containing positional data.

• Application (apply_jager_corrections function): If an intrinsic_corrections array (generated by get_intrinsic_corrections) is provided, it is applied after the Jäger corrections. The Jäger-corrected dose values (currently, the code applies it to an array formed from both PR and DPP corrected sums, which might need clarification based on intended use) are multiplied element-wise by the numeric_intrinsic factors.

  # Assuming corrected_count_array contains Jäger-corrected values
  corrected_count_array *= numeric_intrinsic

Implementation Details

The primary script for these corrections is src/corrections.py.

• apply_jager_corrections(): Orchestrates the application of PR and DPP corrections and optionally the intrinsic correction.
• pulse_rate_correction(): Calculates and applies the PR-specific Jäger correction.
• dose_per_pulse_correction(): Calculates and applies the DPP-specific Jäger correction.
• get_intrinsic_corrections(): Calculates the intrinsic correction factors from 'Corrected Counts' and 'Raw Counts' data.

Input data typically consists of:

• counts_accumulated_df: A Pandas DataFrame with accumulated counts over time for each detector.
• bkrnd_and_calibration_df: A Pandas DataFrame containing background and calibration values for each detector.
• array_data (for intrinsic corrections): A dictionary containing NumPy arrays for 'Corrected Counts' and 'Raw Counts'.

The output is typically a NumPy array (or DataFrame) of corrected dose values.

Reference

The correction methodology implemented is based on the work by:

Jäger, R., Kapsch, R. P., & Karger, C. P. (2021). A correction method for the dose rate dependent response of silicon diode arrays for measurements in pulsed photon beams with varying pulse frequency and pulse dose. Physics in Medicine & Biology, 66(20), 205015. (Available at: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8504598/ - Note: Direct programmatic access to this URL may be restricted)

This README aims to provide a comprehensive understanding of the dose correction formulas and their implementation within this project.