Patient-Specific QA and Dosimetric Verification

The implementation of advanced radiotherapy techniques—specifically Intensity-Modulated Radiation Therapy (IMRT) and Volumetric Modulated Arc Therapy (VMAT)—allows radiation doses to conform tightly around complex tumor shapes. However, these techniques require rapid, dynamic movements of the Multileaf Collimators (MLCs) and continuous variations in gantry speed and dose rate. Because these plans are highly complex, standard manual checks are insufficient. The medical physicist must perform pre-treatment Patient-Specific Quality Assurance (PSQA) for every plan before delivery.

This article reviews measurement-based verification systems, the mathematics of the Gamma Index algorithm, and secondary calculation verification methods.

1. Measurement-Based Verification Systems

In patient-specific QA, the patient's plan is transferred onto a phantom containing an array of detectors. The LINAC delivers the plan exactly as it would to the patient, and the measured dose is compared to the treatment planning system (TPS) calculations:

2D Diode/Ion Chamber Arrays * **Design**: A flat panel containing hundreds of silicon diodes or ionization chambers spaced $5\text{ to }10\text{ mm}$ apart (e.g., MapCHECK). * **Use Case**: Placed at a specific depth inside a plastic water phantom, it records the 2D dose distribution perpendicular to the beam axis.

3D Cylindrical Arrays * **Design**: A cylindrical phantom containing detectors arranged in a spiral or grid pattern (e.g., ArcCHECK). * **Use Case**: Ideal for VMAT plans, as the cylindrical geometry accounts for gantry rotation and measures the dose delivered from all $360^\circ$ angles.

Electronic Portal Imaging Device (EPID) * **Design**: The amorphous silicon (a-Si) MV flat-panel imager integrated opposite the LINAC head. * **Use Case**: Known as **Portal Dosimetry**, it measures the raw radiation fluence emerging from the collimator. This provides high spatial resolution ($< 0.5\text{ mm}$ pixel pitch) and eliminates the need for bulky phantoms.

2. The Gamma Index Formalism

Comparing two dose distributions (measured vs. calculated) point-by-point is challenging in regions of steep dose gradients. A minor spatial displacement (e.g., 1 mm) can cause a massive dose difference (e.g., 30\%), resulting in false-positive failures.

To solve this, Low et al. introduced the Gamma Index (\gamma) Formalism. The Gamma index evaluates both Dose Difference (DD) and Distance-to-Agreement (DTA) simultaneously:

Where: * r_m: The position of the measurement point. * r_c: The position of the calculation point. * \Delta D = D_m(r_m) - D_c(r_c): The dose difference between the measurement point and the calculation point. * r = |r_m - r_c|: The physical distance between the measurement point and the calculation point. * \Delta D_{tol}: The dose difference tolerance (e.g., 3\% of the prescription dose). * d_{tol}: The distance-to-agreement tolerance (e.g., 2 mm).

Evaluation Criteria: * If **$\gamma(r_m) \le 1.0$**: The measurement point passes the criteria. * If **$\gamma(r_m) > 1.0$**: The measurement point fails.

A plan is clinically acceptable if a target percentage of points—the Gamma Pass Rate—exceeds a threshold. According to AAPM TG-218 guidelines: * Standard IMRT/VMAT: Target pass rate of > 95\% using 3\% / 2 mm global criteria. * Stereotactic Treatments (SRS/SBRT): Often requires tighter criteria, such as 2\% / 2 mm or 3\% / 1 mm.

3. Secondary Dose Calculations (Software-Based QA)

In addition to physical measurements, physicists use independent dose calculation software (e.g., RadCalc, Mobius3D) as a secondary check. These programs: 1. Import the patient's plan DICOM files (leaf sequences, gantry angles, Monitor Units). 2. Recalculate the dose in the patient's CT voxel geometry using an independent algorithm (such as a Collapsed Cone or Monte Carlo code). 3. Compare the results to the TPS dose. This helps catch plan corruption during transfer, TPS modeling bugs, or software calculation errors.

4. Troubleshooting Plan Failures

When a plan fails its patient-specific QA (Gamma pass rate < 95\%), the physicist must investigate the root cause: * MLC Leaf Speed and Position Errors: Mechanical wear on the MLC motors can cause leaves to lag behind their planned positions during dynamic sweeps, creating dose discrepancies. * MLC Transmission Calibration: The TPS must model the fraction of radiation leaking through the tungsten leaves (1 to 2\%) and between adjacent leaves (intraleaf leakage). If the transmission factor in the TPS is miscalibrated, calculated doses will deviate from measurements. * Gantry Angle Offsets: Gantry sag due to gravity during rotation can introduce small alignment shifts during VMAT delivery.

Conclusion

Patient-Specific Quality Assurance is the final safety net in the treatment planning chain. By measuring dose distributions using advanced diode arrays or portal imagers, and comparing them to planning calculations using the Gamma Index formalism under TG-218 guidelines, medical physicists verify that the LINAC delivers the planned dose distribution with high fidelity. This provides absolute assurance of safety and accuracy before the patient's first treatment.