2975 - Quantitative and Qualitative Evaluation of Advanced Cone-Beam CT Imaging In Lung, Breast, and Head-and-Neck Cancer Patients

12:30pm - 01:45pm ET

Poster Hall - Exhibit Hall A

Screen: 15

POSTER

Presenter(s)

Wouter van Elmpt, PhD - MAASTRO Clinic, Maastricht, Drenthe

N. Coorens¹, C. Hazelaar², V. Trier Taasti¹, S. Davidson³, and W. van Elmpt⁴; ¹Maastro, Maastricht, Netherlands, ²Maatro, Maastricht, Netherlands, ³Varian Medical Systems, Palo Alto, CA, ⁴Department of Radiation Oncology, GROW – School for Oncology and Developmental Biology, Maastricht University Medical Center, Maastricht, Netherlands

Purpose/Objective(s):

To evaluate the quality of cone-beam computed tomography (CBCT) scans acquired with an advanced imaging system in lung, breast, and head-and-neck cancer patients.

Materials/Methods:

For six lung stage I, seven lung stage II-IV, seven breast, and seven head and neck cancer patients, a planning CT (pCT), a conventional CBCT (C-CBCT) and a CBCT scan using an advanced imaging system (H-CBCT) were acquired. All lung stage I and breast H-CBCTs and breast C-CBCTs were acquired in breath-hold, and the other CBCT scans in free-breathing. H-CBCT scans were acquired in 6 seconds, and for lung stage II-IV, an additional 60 seconds H-CBCT was acquired. All H-CBCT scans were reconstructed twice, using a filtered back-projection algorithm (FDK) and iterative reconstruction (IR). C-CBCT scans were acquired in 60 seconds and reconstructed using FDK.

For quantitative assessment of the image quality, mean CT number and standard deviation (SD) were calculated from homogeneous volume-of-interest (0.1-4.0 cm³) in tumor and muscle tissue. Image noise in the CBCTs was compared to the pCT by calculating the ratio of the SD of the CT numbers.

Qualitatively, experienced observers (two radiation oncologists for lung cancer patients and seven radiotherapy technicians for head-and-neck cancer patients) rated their confidence in contouring organs-at-risk (OARs) using a 5-point Likert scale (1 = lowest, 5 = highest confidence). Confidence scores were recorded for the pCT, IR-H-CBCT and C-CBCT (only for lung), and observer ratings were averaged per patient.

Results:

Quantitative analysis showed comparable image noise for IR- and FDK-H-CBCT 6s scans. Across the four patient groups, the CBCT/pCT SD ratios for muscle tissue ranged from 1.1 to 2.1 for IR and from 1.0 to 1.8 for FDK-H-CBCT. For tumor tissue, the SD ratios ranged from 1.4 to 3.2 for IR and from 0.9 to 2.2 for FDK-H-CBCT. The image noise increased in C-CBCT scans (muscle: 2.9–3.9; tumor: 1.8–4.4) and in the 60s H-CBCT scans of lung stage II-IV patients (IR: 3.5; FDK: 3.3 for muscle, and 3.5; 2.7 for tumor).

The qualitative assessment revealed that OAR contouring confidence was highest on pCT, with median scores of 5.0 (lung stage I), 4.5 (lung stage II–IV), and 3.7 (head-and-neck). Confidence was rated slightly lower for the IR-H-CBCT (4.0; 3.0; 3.1) and lowest for the C-CBCT (2.0; 1.5, only lung).

Conclusion:

CBCT images acquired with the advanced system and a 6-second protocol led to lower image noise and higher contouring confidence compared to longer acquisition protocols and conventional CBCT scans.