An Investigation of the Feasibility and Utility of a Low-dose Cone-beam Computed Tomography Scan Protocol for Head and Neck Cancer Patients.

Introduction: Routine use of cone-beam computed tomography (CBCT) scan protocols as part of the image guidance process (image-guided radiation therapy) has become an integral part of the practice of radiation therapists (RTs). Concerns regarding imaging dose as well as increased in-room time for patients led to reluctance among site group members to adopt CBCT for all radical head and neck cancer (HNC) patients at our institution. This investigation set out to assess the feasibility and utility of a revised CBCT scan protocol with the aim of supporting daily CBCT for HNC patients receiving radiation therapy.

Methods: The project was performed in three phases. Phase 1 involved the experimental adjustment of CBCT scan protocol parameters in clinical use for HNC patients at our institution. An Elekta Synergy linear accelerator with kilovoltage CBCT capability and a RANDO head phantom were used for scan acquisition procedures. Image registration using bony anatomy was performed on two image sets generated using the current clinical scan protocol (HNS20) and an experimental modified scan protocol (MHNS20). Image registration results were compared by two investigators. Measurements of scan doses using a metal-oxide-semiconductor field-effect transistor and a Unidose meter were performed. Catphan phantom images were acquired using HNS20 and MHNS20 protocols. In phase 2, ten volunteer RTs performed image registration and matching processes on two image sets performed using HNS20 and MHNS20 protocols. RTs were unaware of the scan protocols used for image acquisition. A threshold of 3 mm was set (the current maximum couch shift allowance in the clinical HNC IGRT protocol) to compare the image registration data from HNS20 and MHNS20. In phase 3, after research ethics board approval, 10 HNC patients consented to the study. Two pretreatment CBCT scans were performed: scan 1 was acquired using MHNS20 protocol, and scan 2 was acquired using the HNS20 protocol. A threshold of 2 mm was set to compare the differences in couch shift data resulting from the image registration of the two image sets. Comparison of HNS20 and MHNS20 based on image registration results was performed.

Results: In phase 1, radiation doses measured by the investigators on the left optical lens using a metal-oxide-semiconductor field-effect transistor and a Unidose meter indicated that the MHNS20 protocol would result in a lower dose to the left optical lens. In phase 2, shifts of the treatment table to achieve the planned isocentre, which were recorded after the image matching process, were within 3 mm in 80% of the RT procedures. In the y-axis (superior/inferior direction), 100% of the procedures were within 3 mm. In the z-axis (anterior/posterior) and x-axis (lateral), 90% of the procedures were within 3 mm. Qualitative data from a questionnaire completed by RTs after the image matching indicated that 50% of the RTs had no preference between the images sets in terms of visibility of structures. Forty percent of RTs had no preference regarding speed of matching or preference for registration between the image sets. When a preference was indicated, the HNS20 scan protocol was chosen by the RTs. In phase 3, couch shift data recorded after each CBCT scan were compared. All results in all three planes for 10 patients included in the study were within the 2-mm threshold.

Conclusions: The feasibility and clinical utility of a potential lower-dose CBCT scan protocol has been investigated. The modified protocol (MHNS20) produced image data acceptable within current practice using bony anatomy for registration purposes. The MHNS20 protocol also delivered lower doses to the left optical lens and therefore potentially to other pertinent structures. The actual delivered doses to patients during IGRT procedures using the MHNS20 may be different than those measured during this investigation.