Evaluation of image-guided and surface-guided radiotherapy for breast cancer patients treated in deep inspiration breath-hold: A single institution experience

Abstract

Introduction: Nowadays, deep inspiratory breath-hold is a common technique to reduce heart dose in left-sided breast radiotherapy. This study evaluates the evolution of the breath-hold technique in our institute, from portal imaging during dose delivery to continuous monitoring with surface-guided radiotherapy (SGRT).

Materials and methods: Setup data and portal imaging results were analyzed for 98 patients treated before 2014, and SGRT data for 228 patients treated between 2018 and 2020. For the pre-SGRT group, systematic and random setup errors were calculated for different correction protocols. Residual errors and reproducibility of breath-holds were evaluated for both groups. The benefit of using SGRT for initial positioning was evaluated for another cohort of 47 patients.

Results: Online correction reduced the population mean error from 3.9 mm (no corrections) to 1.4 mm. Despite online setup correction, deviations greater than 3 mm were observed in about 10% and 20% of the treatment beams in ventral-dorsal and cranial-caudal directions, respectively. However, these percentages were much smaller than with offline protocols or no corrections. Mean absolute differences between breath-holds within a fraction were smaller in the SGRT-group (1.69 mm) than in the pre-SGRT-group (2.10 mm), and further improved with addition of visual feedback (1.30 mm). SGRT for positioning did not improve setup accuracy, but slightly reduced the time for imaging and setup correction, allowing completion within 3.5 min for 95% of fractions.

Conclusion: For accurate radiotherapy breast treatments using deep inspiration breath-hold, daily imaging and correction is required. SGRT provides accurate information on patient positioning during treatment and improves patient compliance with visual feedback.

Keywords: (U, V), ventral-dorsal and cranial-caudal direction in the tangential beam, respectively; Breast; Breath-hold; CBCT, cone-beam CT; CT, computer tomography; DIBH; DIBH, Deep inspiratory breath-hold; DRRs, digitally reconstructed radiographs; LAT, medio-lateral direction; LNG, cranial-caudal direction; NAL, no-action-level setup correction protocol; OTM, online treatment monitor; SGRT, surface-guided radiotherapy; Surface-guided radiotherapy; VRT, anterior-posterior direction; eNAL, extended NAL setup correction protocol.

Fig. 1

Surface guided radiotherapy (SGRT) workflow for deep-inspiration breath-hold since May 2018. SGRT signal for the movement in ventral-dorsal direction is shown in time for one fraction. During the imaging procedure (t = 50–235 sec), reference surfaces are captured (blue arrow) at the end of CBCT acquisition (indicated in green) to link the external surface to the internal anatomy on the CBCT, and after couch movement for setup correction (red). Afterwards, dose was delivered with four treatment beams (magenta). Prior to each DIBH, the patient was asked to deeply breathe in and out once. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 2

A. Histograms of the applied online cone-beam CT corrections in the pre-SGRT group, in anterior-posterior (left), cranial-caudal (middle) and medio-lateral (right) directions. B. Online treatment monitor, showing the projection of the breast contour (green) and the lung contour ± 4 mm margin (red and green) from the DRRs on top of individual, real-time acquired portal imaging frames. In this case, image registration revealed a residual offset after online CBCT correction of −4 mm in cranial-caudal direction (V). c. Cumulative histogram of the residual 2D systematic setup error projected on the UV-plane of the medio-lateral beam: without correction protocol (black triangles), with online corrections (blue circles), with a no-action-level protocol (NAL N = 3, red stars), and with the extended NAL protocol (cyan line). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 3

Frequency (%) of detected deviations larger than 1, 2, 3, and 5 mm in ventral-dorsal (U) and cranial-caudal direction (V) in 98 OTM-monitored patients, and in anterior-posterior (VRT), cranial-caudal (LNG) and medio-lateral (LAT) directions in 228 SGRT-monitored patients. For a study-group of 19 patients during the transition period, both data is shown.

Fig. 4

Deep inspiration breath-hold reproducibility in ventral direction improves with surface guidance. Maximum variation in deep inspiration breath-hold over all beams per fraction in ventral direction from portal imaging (OTM cohort, left) and surface guidance (SGRT cohort, right).

Fig. 5

Time needed to complete the imaging procedure (CBCT acquisition, matching and on-line setup correction) for patients positioned on tattoos only (red, 25 patients) or using SGRT (blue, 47 patients). All patients were treated for left-sided breast cancer in deep inspiration breath-hold. * denotes a significant difference based on a Wilcoxon rank sum test (p-value < 0.01). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)