Dosimetric impact of MLC positional errors on dose distribution in IMRT

Area of Science:

• Medical Physics
• Radiation Oncology
• Radiotherapy Technology

Background:

• Accurate radiotherapy delivery relies on precise positioning of multileaf collimators (MLC).
• MLC positional errors can significantly impact dose accuracy and increase radiation exposure to healthy tissues.
• Understanding dose sensitivity and complexity metrics is vital for optimizing MLC performance.

Purpose of the Study:

• To investigate dose sensitivity variations and complexity metrics associated with MLC positional errors in volumetric modulated arc therapy (VMAT).
• To determine acceptable ranges for MLC positional accuracy across different clinical scenarios.
• To establish correlations between MLC positional error sensitivity and treatment plan complexity.

Main Methods:

• Treatment plans for prostate, lung, spinal, and brain metastases were generated using different treatment planning systems (TPS) and fraction sizes.
• Systematic or random MLC leaf bank errors (0.25-2.0 mm) were introduced to assess dose sensitivity (gEUD) and complexity metrics (MU/Gy, plan irregularity).
• Correlation coefficients between gEUD sensitivity and complexity metrics were calculated to determine required positional accuracy tolerances.

Main Results:

• Dose sensitivity to systematic MLC positional error varied significantly by treatment site, TPS, and fraction size.
• Strong correlations (r = 0.88-0.93) were observed between gEUD sensitivity and plan complexity metrics.
• Minimum allowable MLC positional errors ranged from 0.28 mm (spinal metastasis) to 1.02 mm (brain metastasis).

Conclusions:

• The acceptable range for MLC positional accuracy is treatment site-dependent.
• Complexity metrics, such as MU/Gy, can guide the setting of appropriate MLC positional accuracy tolerances for each treatment site.
• Implementing complexity-based quality control can lead to more refined MLC positional accuracy management and reduced dose errors in radiotherapy.