Quantifying the interfractional displacement of the gastroesophageal junction during radiation therapy for esophageal cancer

Abstract

Purpose: Accounting for interfractional changes in tumor location improves the accuracy of radiation treatment delivery. The purpose of this study was to quantify the interfractional displacement of the gastroesophageal junction (GEJ) based on standard treatment setup in patients with esophageal cancer undergoing radiation therapy.

Methods and materials: Free-breathing four-dimensional computed tomography (4D-CT) datasets were acquired weekly from 22 patients during treatment for esophageal adenocarcinoma. Scans were registered to baseline (simulation) 4D-CT scans by using bony landmarks. The distance between the center of the GEJ contour on the simulation scan and the mean location of GEJ centers on subsequent scans was used to assess changes in GEJ location between fractions; displacement was also correlated with clinical and respiratory variables.

Results: The mean absolute random error was 1.69 mm (range, 0.11-4.11 mm) in the lateral direction, 1.87 mm (range, 0.51-4.09 mm) in the anterior-posterior (AP) direction, and 3.09 mm (range, 0.99-6.16 mm) in the superior-inferior (SI) direction. The mean absolute systemic GEJ displacement between fractions was 2.88 mm lateral (≥ 5 mm in 14%), mostly leftward; 2.90 mm (≥ 5 mm in 14%) AP, mostly anterior; and 6.77 mm (≥ 1 cm in 18%) SI, mostly inferior. Variations in tidal volume and diaphragmatic excursion during treatment correlated strongly with systematic SI GEJ displacement (r = 0.964, p < 0.0001; and r = 0.944, p < 0.0001, respectively) and moderately with systematic AP GEJ displacement (r = 0.678, p = 0.0005; r = 0.758, p < 0.0001, respectively). Systematic displacement in the inferior direction resulted in higher-than-intended doses (≥ 60 Gy) to the GEJ, with increased hot-spot to the adjacent stomach and lung base.

Conclusion: We found large (>1-cm) interfractional displacements in the GEJ in the SI (especially inferior) direction that was not accounted for when skeletal alignment alone was used for patient positioning. Because systematic displacement in the SI direction had dosimetric impact and correlated with tidal volume, better accounting for depth of breathing is needed to reduce interfractional variability.

Figure 1

Box-and-whisker plots showing interfractional systematic displacement for each patient in the lateral (A), anterior-posterior (AP) (B), and superior-inferior (SI) (C) directions. The band at the middle represents the median; the bottom and top of the box represent the 25^th and 75^th percentiles. The whiskers indicate the lowest datum still within 1.5 interquartile range (IQR) of the lower quartile and the highest datum still within 1.5 IQR of the upper quartile. Xs indicate outliers of maximum displacements. Positive numbers indicate left, anterior, and inferior displacement; negative numbers indicate right, posterior, and superior displacement.

Figure 2

Differences in weekly measurements of tidal volume (panels A-C) and vertical diaphragmatic excursion (panels C-F) plotted against interfractional displacements in three dimensions.

Figure 3

Dosimetric impact of systematic error for two patients with large superior (panels A-C) or inferior (panels D-F) displacements of the gastroesophageal junction (GEJ). A-C, original treatment plan (with 1-cm planning target volume [PTV] margins) (A), week 5 treatment plan (B) and normalized dose-volume histograms (C) for patient 1, who had large superior displacement of the GEJ. Dotted lines indicate dose distribution on the original treatment plan; solid lines, dose distribution on week 5. D-F, original treatment plan (with 1-cm PTV margins) (D), week 5 treatment plan (E), and normalized dose-volume histograms (F) for patient 12, who had large inferior GEJ displacement. Dotted lines indicate the original treatment plan; thick solid lines, the cumulative dose distribution; and thin solid lines, the individual weekly dose distributions for patient 12. G and H, portal films for patient 12, obtained at simulation (G) and at week 5 (H). Brown, gross tumor volume (GTV); cyan, clinical tumor volume (CTV); purple, heart; white, hot spots.