Prediction and monitoring of adaptive radiation therapy timing using two-dimensional X-ray image-based water equivalent thickness

doi:10.21203/rs.3.rs-4614591/v1

Download PDF

Research Article

Prediction and monitoring of adaptive radiation therapy timing using two-dimensional X-ray image-based water equivalent thickness

https://doi.org/10.21203/rs.3.rs-4614591/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Background

This study aimed to predict and monitor the optimal timing for implementing adaptive radiation therapy (ART) using two-dimensional X-ray image-based water equivalent thickness (2DWET).

Methods

The study included 40 patients with oropharyngeal and hypopharyngeal cancer who underwent CT rescanning during the treatment period. An adaptive score (AS) was proposed as a quantitative indicator to facilitate the decision regarding when to implement ART. The AS was derived from changes in four key dose indices: target coverage, spinal cord dose, parotid gland dose, and over-dose volume. Delivered dose distributions were reviewed by two oncologists specializing in head and neck radiation therapy, and the need for ART was evaluated using a four-point score. Logistic regression analysis was used to determine the AS cutoff value, and receiver operating characteristic analysis was used to assess 2DWET as a predictor of ART timing.

Results

The AS strongly correlated with the decisions made by the radiation oncologists, with Pearson correlation coefficients of 0.74 and 0.64. An AS cutoff value of 7.5 was identified as an indicator of the optimal time to implement ART, predicting two oncologists' decisions with sensitivities of 79.2% and 89.5% and specificities of 87.5% and 81.0%, respectively. The 2DWET method detected AS = 7.5 with a sensitivity of 63.2% and a specificity of 81.0%.

Conclusions

An adaptive score of 7.5 strongly correlated with the radiation oncologists' decision to implement ART and could therefore be used as a surrogate marker. Two-dimensional WET detected AS = 7.5 with high sensitivity and specificity and could potentially be used as a highly efficient and low-exposure tool for predicting and monitoring the optimal timing of ART implementation.

Adaptive radiation therapy

2D X-ray image

Water equivalent thickness

Optimal timing of ART

Head and neck cancer

Monitoring anatomical changes

Intensity-modulated radiation therapy (IMRT) has become widespread in cancer treatment and is implemented in many centers [1]. In head and neck radiotherapy, side effects such as stomatitis, taste disorders, trismus, dysphagia, and esophagitis occur, reducing patient quality of life [2–6]. IMRT is effective in the treatment of complex anatomical structures in the head and neck region, resulting in a lower dose to organs at risk (OARs) and fewer side effects compared with conventional radiotherapy [7–9]. However, decreases in body thickness and tumor size during the treatment period may decrease the therapeutic dose to the target and increase the dose to OARs [10–12]. Therefore, adaptive radiation therapy (ART) is used to detect changes in patient anatomy during the treatment period and adjust the dose distribution accordingly [13, 14]. Numerous studies have reported the effectiveness of ART in response to anatomical changes in the head and neck during the course of IMRT [15–19]. The decision to implement ART is often made based on clinical assessments, such as tumor reduction and changes in body thickness observed on cone beam computed tomography (CBCT) images, exacerbation of side effects, and challenges related to patient immobilization [20]. However, the decision-making process for implementing ART is qualitative, and accurately predicting its benefits can be difficult. Determining the optimal timing for ART implementation also remains challenging. Castelli et al. reported that the benefits related to dose distribution and overdose correction in the parotid gland increase with reductions in tumor size and neck thickness [21]. Kumar et al. reported that changes in neck thickness correlate more strongly with changes in parotid gland dose than with body weight loss [22]. Therefore, by observing changes in neck thickness and tumor size in the course of daily treatment, the best time to implement ART may be quantitatively determined. Because the locations of the clinical target volume (CTV) and OARs in head and neck radiotherapy depend on bone alignment, two-dimensional (2D) X-ray images are often used in daily image-guided radiation therapy (IGRT). In this workflow, weekly CBCT is performed, but more frequent monitoring may be required to identify the optimal time to implement ART. CBCT can be used for daily IGRT, but doing so for every treatment results in a high radiation dose and prolonged treatment time [23–25]. In response, we developed the 2D image-based water equivalent thickness (2DWET) method to quantitatively measure changes in anatomical dimensions using 2D X-ray images acquired during daily IGRT [26]. Although previous studies of ART timing have investigated the relationship between various predictors and dose indices for OARs and target tissues, quantitatively predicting ART timing and benefits based on changes in anatomical structure remains a challenge [27]. The tumor shrinkage pattern during treatment and the benefits of ART implementation vary widely from patient to patient, and standardization using scheduled ART with a fixed replanning date is limited [28, 29]. To quantitatively determine ART timing, an index that can objectively determine when to implement ART is required. In response, we devised the adaptive score (AS) as a new index for quantitatively assessing the benefits of ART. The AS was calculated from dose indices that oncologists consider important in planning ART, including the mean parotid gland dose, maximum spinal cord dose, the volume of the overdose region, and the planning target volume (PTV) dose coverage. If the 2DWET method can predict the changes in AS during treatment period, the optimal timing of ART implementation could be quantitatively determined without additional radiation exposure or effort to detect anatomical changes.

This study had two goals. The first was to show that the AS can be used with appropriate thresholds as surrogate for an oncologist's decision to implement ART. The second was to evaluate the feasibility of quantitative ART decision-making based on 2DWET prediction of the AS.

Patients and imaging datasets

Forty patients with head and neck cancer were enrolled, and images from 40 simulation CT images and 40 CT rescans of each patient were retrospectively analyzed (Table 1). The patients were diagnosed with oropharyngeal or hypopharyngeal cancer and treated using IMRT. All patients were immobilized using a patient-specific pillow and a four- or five-point thermoplastic mask covering the shoulder. CT datasets were acquired using an Aquilion ONE scanner (Canon Medical Systems, Tochigi, Japan). The decision to perform a CT rescan was based on clinical judgment of worsening of side effects, loss of ≥ 10% of the patient’s weight at the time of initial treatment, and an ill-fitting mask. Patient information was anonymized. Although informed consent was not required, the homepage of the National Cancer Center Hospital East published details of this study and allowed patients to refuse to participate. The study methods, including the investigation procedure and handling of patient information, were approved by the institutional review board of the National Cancer Center Hospital East (IRB No. 2020 − 282).

Table 1

Patient characteristics.
Patient	Age, years	Sex	Dose (Gy/fr)	TNM*	Site
1	72	Female	70/35	T4N1M0	Oropharynx
2	57	Male	70/35	T4N2M0	Oropharynx
3	69	Male	70/33	T2N2M0	Oropharynx
4	42	Male	70/33	TXN3M0	Oropharynx
5	80	Male	70/35	T4N2M0	Oropharynx
6	55	Male	70/35	T4N2M1	Oropharynx
7	54	Male	70/35	T4N2M0	Oropharynx
8	74	Male	70/35	T1N2M0	Oropharynx
9	63	Male	70/33	T2N0M0	Oropharynx
10	45	Male	70/35	T4N1M0	Oropharynx
11	68	Male	70/35	T4N1M0	Oropharynx
12	65	Female	70/35	T4N2M1	Oropharynx
13	79	Male	70/35	T2N1M0	Oropharynx
14	60	Female	70/35	T4N2M0	Oropharynx
15	80	Male	70/33	T4N1M0	Oropharynx
16	42	Male	70/33	T1N3M0	Oropharynx
17	52	Male	70/35	T2N1M0	Oropharynx
18	51	Male	70/35	T2N3M0	Oropharynx
19	75	Male	70/35	T3N3M1	Oropharynx
20	81	Male	70/35	T2N2M0	Oropharynx
21	71	Male	70/35	T4N3M0	Oropharynx
22	67	Male	70/35	TXN3M0	Oropharynx
23	79	Male	70/35	T4N1M0	Oropharynx
24	71	Male	70/35	T1N2M0	Hypopharynx
25	77	Male	70/33	T2N0M0	Hypopharynx
26	74	Male	70/35	T2N2M0	Hypopharynx
27	69	Male	70/35	T2N3M0	Hypopharynx
28	71	Male	70/35	T2N3M0	Hypopharynx
29	61	Male	70/33	T4N3M0	Hypopharynx
30	63	Male	70/35	T2N1M0	Hypopharynx
31	68	Male	70/33	T2N0M0	Hypopharynx
32	73	Male	70/35	T4N0M0	Hypopharynx
33	71	Male	70/33	T1N3M0	Hypopharynx
34	41	Male	70/35	T4N2M0	Hypopharynx
35	66	Male	70/33	T2N2M0	Hypopharynx
36	80	Male	70/35	T2N3M0	Oropharynx
37	76	Male	70/35	T2N1M0	Oropharynx
38	74	Male	70/35	T4N1M0	Oropharynx
39	69	Male	70/35	T2N2M0	Oropharynx
40	78	Male	70/35	T4N3M0	Oropharynx
* 8th edition of UICC (44)

Treatment planning

Forty VMAT plans based on the initial CT images and 40 VMAT plans based on the CT rescan images were created. Most plans were created using two arcs and collimator angles of 350° and 10°. Three arcs and arbitrary collimator angles were used in some cases with very large targets. All treatments were planned using a RayStation treatment planning system (RaySearch Laboratories AB, Stockholm, Sweden) by a well-trained radiation therapist and medical physicist. The CTV and OARs were contoured on the simulation CT images by a radiation oncologist, and the CTV and OARs on the CT rescan were defined by the medical physicist using a deformed region of interest (ROI) based on the contouring by the radiation oncologist. Treatment plans were generated with 35 or 33 fractions of 2 Gy to give a total dose of 70 or 66 Gy for cases intended for curative treatment. All plans were designed in accordance with the clinical protocol: PTV (high-risk), > 98% of the volume to receive > 95% of 70 Gy; PTV (low-risk), > 98% of the volume to receive > 95% of 54 or 56 Gy; PTV, ≤ 2% of the volume to receive < 110% of 70 Gy; maximum dose to spinal cord, < 45 Gy; maximum dose to brain stem, < 54 Gy; mean dose to parotid gland, < 26 Gy; dose to cochlea, < 40 Gy; mean dose to oral cavity, < 35 Gy; ≤2% of lower jawbone to receive < 105% of 70 Gy. However, if all clinical goals were difficult to achieve, plans were created with the highest priority given to achieving the spinal cord and PTV (high-risk) dose constraints.

Dosimetric evaluation

The IMRT plan based on the initial CT images (initial plan) was replaced with the recalculated plan based on the rescanned CT images. The dose distribution in the replacement plan was determined by importing the initial CT and rescanned CT images into the treatment planning system, aligning both images using the automatic rigid image registration function in RayStation, and then recalculating them using the beam geometry information in the initial plan. The recalculated plan (delivered plan) was compared with the adaptive plan based on the rescanned CT images to evaluate the effects of the ART plan on the following dose indices: parotid gland D50, spinal cord Dmax, and oral cavity D50 for OAR evaluation; conformity index, homogeneity index, low-risk PTV D98, and high-risk PTV D98, D50, and D2 for target evaluation; body Dmax, V105, and V107 for over-dose region evaluation. The ROIs used for dosimetric evaluation of the parotid gland, spinal cord, oral cavity, and PTV were transformed from the initial CT image to a rescanned CT image using the deformable image registration (DIR) function in RayStation. The ROIs for the deformed target and OARs were verified and corrected by a radiation oncologist. The change in each dose index was obtained by calculating the difference in its values between the replacement and adaptive plans.

The 2D image-based WET method

The 2DWET method described in our previous study (26) was changed from a Java to MATLAB environment and improved as outlined below. First, 2D X-ray images (sagittal, 70 kV and 200 mA) obtained on the initial treatment day and the day of the CT rescan were imported into the 2DWET software. The acquired 2D X-ray images were automatically registered using the bone structure as a marker, and then image subtraction was performed. Tables for converting the X-ray intensity into WET were determined for each linear accelerator and used to transform the difference X-ray images into images of WET change. A color map was applied such that an increase or decrease in WET was indicated by a hot or cold color, respectively. On the output WET image, the improved 2DWET method allows an ROI to be created and the mean WET within that ROI to be quantified. The processing time of the improved method was approximately 1 min, depending on the PC environment. Figure 1 shows representative images obtained using the 2DWET method, demonstrating the ability to locate and quantify changes in body and tumor shape. In this study, an ROI was centered on the isocenter of the sagittal image and encompassed the base of the skull and the C1 vertebra, the hyoid bone, the posterior cervical muscles, and the boundary between the shoulder and neck.

Two-dimensional WET images of the head and neck region. Hot and cold colors in the color map indicate an increase and decrease, respectively, in WET. (a) No anatomical changes or positioning errors. (b) Tumor shrinkage and a decreased body thickness. (c) Positioning errors of the mandible and vertebral body. (d) Tumor growth and an increased body thickness. Abbreviations: WET, water equivalent thickness; 2DWET, two-dimensional X-ray image-based WET.

Adaptive score

Increased parotid gland dose, spinal cord dose deviation, and worsening PTV coverage are important factors affecting a clinical oncologist's decision to implement ART [6, 14–16, 21, 30]. In this study, the left and right parotid gland Dmean values, spinal cord Dmax, V107, and D98 were selected as dose indices for head and neck ART. Changes in these indices depend on the geometry of the CTV, tumor size, and tumor location and thus vary greatly among patients during tumor shrinkage. Furthermore, the necessity of implementing ART increases in instances where a single index changes significantly, as well as in cases where multiple indices change slightly. To incorporate these factors into the ART evaluation, the AS was used to categorize and integrate the clinical impact of each dose index. In calculating the AS, the dose difference between the replacement and adaptive plans was assigned a categorical score (C) on a scale of 0 (no clinical impact) to 10 (substantial clinical impact) for the following indices: left parotid gland Dmean (C_{parotid L}), right parotid gland Dmean (C_{parotid L}), spinal cord Dmax (C_{spinal cord}), V107 (C_V107), and PTV D98 (C_D98). To accurately reflect the impact of the radiation oncologist's clinical judgment, as well as simplify the AS, the categorical score was increased by one in each of the following cases: for each 1 Gy increase in Dmean to the left or right parotid glands; for each 1 Gy increase in Dmax to the spinal cord; for each 1cm³ increase in V107; and for each 1% decrease in the PTV D98. Values exceeding 10 were classified as category 10. The AS was calculated from the categorical scores of each dose index using the following equation:

AS = (C_{parotid R} + C_{parotid L})/2 + C_{spinal cord} + C_V107 + C_D98

For example, if the right parotid gland Dmean increases by 1.0 Gy, the left parotid gland Dmean increases by 1.8 Gy, the spinal cord Dmax increases by 4.6 Gy, V107 increases by 0.22 cm³, and the PTV D98 decreases by 3.2%, then AS = 9.4.

Radiation oncologist’s evaluation

To assess the clinical relevance of the AS, dose distributions were reviewed by two head and neck experts in clinical oncology (MD1 and MD2). The review was performed with the dose distributions and dose–volume histograms (DVHs) of the initial and replacement plans displayed on the screen. On the basis of this information, MD1 and MD2 scored the necessity of ART implementation as follows: 1, no need for re-planning; 2, no need for re-planning, although dose distribution effects were observed; 3, re-planning is recommended owing to dose distribution effects; 4, re-planning is mandatory owing to significant dose distribution effects.

Statistics

Patient characteristics were summarized using descriptive statistics, such as mean and standard deviation (SD). Pearson product-moment correlation coefficients (r) were calculated to assess the association between the dosimetric evaluation, clinical oncologist’s evaluation, and 2DWET. To determine the cutoff value for AS that correlates most closely with the oncologist’s evaluation, logistic regression analysis was performed with AS as an explanatory variable and the oncologist’s evaluation level (1–2 vs 3–4) as an outcome variable. To evaluate the effectiveness of 2DWET, a receiver operating characteristic curve was constructed using a logistic regression model with 2DWET as an explanatory variable and the above-determined AS cutoff value as an outcome variable. All statistical analyses were performed using R (version 4.2.2) and SAS (version 9.4) software.

Adaptive plan outcome

Table 2 shows the change in each dose index at the time of ART implementation in 40 cases. On average ART was initiated after delivering 19.6 dose fractions (2 Gy/fraction), and the average improvement in parotid D50 was 1.47 Gy on the right side and 1.50 Gy on the left side. The average spinal cord Dmax improvement was 1.40 Gy, but the SD was 3.87 Gy (max 15.77 Gy) A 2.9% average improvement was observed in PTV D98. The V107 value was improved by 4.11 cm³, but the SD was 14.6 cm³.

Table Ⅱ. Patient characteristics.

		Initial plan (Mean ± SD)	*Delivered plan (Mean ± SD)
PTV high risk
	D2 (%)	104.5 ± 0.4	106.2 ± 1.2
	D50 (%)	102.0 ± 0.4	102.9 ± 0.8
	D98 (%)	96.1 ± 1.8	93.2 ± 3.7
	CI	0.94 ± 0.03	0.88 ± 0.07
	HI	0.72 ± 0.08	0.53 ± 0.22
PTV low risk
	D98 (%)	95.1 ± 2.3	90.5 ± 6.5
Spinal Cord
	D1cc (Gy)	36.9 ± 2.8	37.4 ± 3.2
	Dmax (Gy)	42.2 ± 2.8	43.4 ± 4.2
Parotid grand right
	Mean dose (Gy)	31.9 ± 10.4	33.3 ± 10.1
Parotid grand left
	Mean dose (Gy)	27.8 ± 7.5	29.3 ± 8.0
Oral cavity
	Mean dose (Gy)	30.1 ± 9.0	30.5 ± 8.6
Body
	Dmax (%)	106.2 ± 0.6	108.8 ± 1.9
	V105(cm³)	2.2 ± 3.0	27.5 ± 41.4
	V107(cm³)	0.0 ± 0.0	4.2 ± 14.6
Abbreviations: Gy, gray; PTV, planning target volume; Dmax, maximum dose; CI, Conformity Index; HI, Homogeneity Index; VX, the percentage of the organ volume that received X Gy or more; DX, dose received by the X% of the volume. *Delivered plan was obtained by replacing the initial plan's beam geometry to the rescan CT and recalculating.

Correlation between 2DWET and dosimetric index change

Figure 2 shows the correlations between 2DWET and each dose index and between weight loss and each dose index. No significant correlations were obtained between 2DWET and spinal cord Dmax (r = 0.11), right parotid gland Dmean (r = 0.08), or left parotid gland Dmean (r = 0.29); a moderate correlation between 2DWET and V107 (r = 0.37) was observed; and a strong correlation (r = 0.63) between 2DWET and PTV (high-risk) D98 was found. In contrast, a moderate correlation (r = 0.34) between weight loss and V107 was measured, but correlations between weight loss and other dose indices were lower.

Figure 2

Correlations between various dose metrics and two ART timing metrics (2DWET and weight).

Abbreviations: ART, adaptive radiation therapy; WL, weight loss; PG_L, left parotid gland; PG_R, right parotid gland; SP, spinal cord; D98, dose received by 98% of the PTV (high-risk); V107, total volume that received ≥107% of the prescription dose; r, Pearson product-moment correlation coefficient.

Correlation between the adaptive score and oncologists' evaluations

The correlation between the AS and ART scores assigned by head and neck radiation oncologists was 0.74 and 0.64 for MD1 and MD2, respectively, indicating a high correlation (Fig. 3). The ART scores assigned by the two radiation oncologists and the mean AS values were as follows: 2.92 ± 1.48 (ART) and 3.33 ± 1.51 (AS) for score 1; 5.40 ± 2.21 (ART) and 6.25 ± 2.49 (AS) for score 2; 10.78 ± 5.40 (ART) and 11.09 ± 3.76 (AS) for score 3, and 14.50 ± 4.27 (ART) and 14.75 ± 6.23 (AS) for score 4. Logistic regression analysis of the association between the AS and oncologist evaluation (Fig. 4) yielded concordance statistics of 0.89 and 0.91 for MD1 and MD2, respectively. The AS cutoff value (the threshold beyond which an oncologist should implement ART) was set at 7.5, which is the point at which Youden’s index is maximal.

Figure 3

Correlations between ART implementation decisions by two oncologists and adaptive scores. Oncologist scores: 1, no change from the initial plan; 2, no need for re-planning, although dose distribution effects were observed; 3, re-planning is recommended owing to dose distribution effects; 4, re-planning is mandatory owing to significant dose distribution effects; Abbreviations: AS, adaptive score; MD, medical doctor (oncologist).

Figure 4

Ability of the AS to predict ART implementation decisions by oncologists. The cutoff values for the logistic analysis were set at the score 3 as a boundary point for deciding whether to implement ART.

The 2DWET-based prediction of ART timing

Figure 5 shows logistic regression analysis of the association between 2DWET and AS using a cutoff value of 7.5. The concordance statistic was 0.74, indicating that 2DWET was a strong predictor of AS.

Figure 5

Ability of 2DWET to predict the AS. The cutoff value for the logistic analysis was set to AS = 7.5.

In this study, we proposed the AS as a new index to quantify the ART benefits for patients. In addition, the AS was monitored using the WET calculated from 2D X-ray images, and we evaluated whether the AS could be used as a surrogate for an oncologist’s decision to implement qualitative ART. All 40 cases presented had CT rescans at the discretion of the oncologists, and the rescan was performed after delivering 19.6 ± 5.5 dose fractions. Dewan et al. conducted a prospective study and found that a CT rescan performed at dose fraction 20 resulted in a significant reduction in the sizes of the parotid gland and treatment volume [16]. Zhang et al. performed weekly CT rescans on 13 patients with oropharyngeal cancer and reported that the optimal timing and frequency of ART was at weeks 1, 2, and 5 [31]. Similar to the above reports, the average timing of the CT rescan in our cohort was at approximately fraction 20. Gan et al. investigated the optimal timing of ART in 110 head and neck cancer patients and found that ART at week 3 provided optimal timing for 10 OARs. Furthermore, they reported that when ART was performed twice in the period from week 2 to week 5, dose fluctuations in the 19 OARs were limited to within 3 Gy [32]. However, in the cohort of our study, the benefits of ART in dose reduction of OARs varied greatly from patient to patient, being 1.5 ± 3.0 and 1.4 ± 2.2 Gy for the left and right parotid glands, respectively, and 1.2 ± 4.1 Gy for the spinal cord. Previous studies have also shown that the benefits of ART vary from patient to patient [33,34]. Therefore, fixed-schedule ART may lead to oversight of cases in which ART is highly necessary or to overexposure of patients who do not require additional CT scans. Belshaw et al. demonstrated that by using DIR and the anatomical details from CBCT images acquired during daily IGRT, it is feasible to accurately recalculate the initial plan and identify the risk of deviation in the spinal cord dose [35]. Additionally, online ART using magnetic resonance imaging (MRI)-guided linear accelerators (linacs) and ART-specific linac devices have seen increasing clinical use in recent years [36-38]. However, the prolonged treatment times of these techniques add to patient strain and elevate radiation exposure during daily CBCT scans. They also heighten the workload of medical physicists and treatment planners, who must recompute dose distributions and perform DIR; increase medical safety risks by necessitating frequent updates; and add to the workload of physicians, who must verify and approve the dose distributions [39]. We considered that triggered ART is the most balanced approach for improving the dose distribution, managing medical staff workload, and optimizing use of resources. Typical indicators for triggered ART encompass weight loss exceeding 10%, a significant reduction in tumor volume, an external contour decrease of more than 1 cm, and dose index changes surpassing 5% for the OARs and targets [40]. New surrogate methods have been suggested for tracking daily anatomical changes and assessing ART benefits [41]. Many studies necessitate the use of multiple CBCT or simulation CT scans for monitoring anatomical changes and determining the ART timing, with patient exposure and the additional workload for medical staff remaining unresolved challenges. Although weight loss as a monitoring method does not involve radiation exposure, weight loss did not always have high sensitivity and specificity for changes in dose indices in our cohort. In contrast, the 2DWET method could detect tumor shrinkage and external contour changes without CT images, and it strongly correlated with reduced PTV coverage in our cohort.

In this study, we proposed a new index (AS) to quantify the ART timing, which is qualitatively determined based on an oncologist's daily examination, palpation, and CBCT image review. The AS was calculated by categorizing and integrating changes in four dose indices: worse target coverage, increased spinal cord dose, increased parotid gland dose, and increased high-dose-range volume. These four dose indices are important factors influencing a radiation oncologist's decision to implement ART. The AS showed a strong correlation with the ART implementation decisions of two oncologists specializing in head and neck radiation therapy (r = 0.74, 0.64). Logistic regression analysis showed that AS = 7.5 was the optimal cutoff value for the implementation of ART. Using this value, the ART implementation judgments of oncologists MD1 and MD2 were predicted with sensitivities of 79.2% and 89.5% and specificities of 87.5% and 81.0%, respectively, suggesting that the AS could serve as a surrogate for oncologists' assessments. Previous studies have not provided a quantitative threshold at which ART should be initiated. However, the 2DWET method has a sensitivity of 63.2% and a specificity of 81.0% in detecting AS = 7.5, suggesting that it may be used as a quantitative indicator for daily ART implementation decisions. Using 2DWET to determine the timing of ART implementation in clinical practice, the 2D X-ray images acquired during daily IGRT can be used for monitoring anatomical changes. By performing CBCT scans less frequently, it is possible to reduce the treatment time and radiation exposure. Moreover, this would reduce the need for ROI replacement and dose distribution recalculations via DIR processing, thereby significantly reducing the workload. Recently, there has been increasing interest in developing a workflow for the implementation of ART in particle therapy for head and neck cancer [42,43]. Being a method for calculating WET based on 2D X-ray images, 2DWET is broadly applicable to both photon and particle therapies.

The limitations of this study include its sole focus on oropharyngeal and hypopharyngeal cancer cases and the restriction of the sample size to 40. The ROI used in the 2DWET analysis covered the whole neck to minimize observer bias; however, measuring the whole neck to determine WET introduces positioning errors, which may lead to an inaccurate representation of tumor reduction or alterations in body contours. The AS was determined using four key dose indicators; however, certain oncologists may prioritize other indicators, such as the oral cavity and pharyngeal constrictor muscles, in their evaluations. The decision to implement ART was performed by only two oncologists, which may bias the results.

We investigated the possibility of predicting the optimal ART timing using 2DWET and quantitative indicators that act as surrogates for radiation oncologists' judgments. An AS of 7.5 strongly correlated with the radiation oncologists' decision to implement ART and could therefore be used as a surrogate marker. Two-dimensional WET strongly correlated with a reduction in PTV coverage and had high sensitivity and specificity in detecting AS = 7.5. Therefore, 2DWET may be a powerful tool to predict and monitor the best ART timing.

adaptive radiation therapy (ART)

adaptive score (AS)

cone beam computed tomography (CBCT)

clinical target volume (CTV)

deformable image registration (DIR)

dose–volume histogram (DVH)

image-guided radiation therapy (IGRT)

intensity-modulated radiation therapy (IMRT)

organs at risk (OARs)

planning target volume (PTV)

region of interest (ROI)

volumetric modulated arc therapy (VMAT)

two-dimensional X-ray image-based water equivalent thickness (2DWET)

Ethics approval and consent to participate

The contents of the study, including the investigation procedure and the handling of patient information, were approved by the institutional review board of the National Cancer Center Hospital East (IRB No. 2020-282). Informed consent was not required for this planning study on anonymized patient data.

Consent for publication

Not applicable.

Availability of data and materials

The datasets used and analyzed during the current study are available from the corresponding author on reasonable request.

Competing Interests

The authors declare no conflict of interest.

Funding

This study was supported by the University of Tsukuba.

Authors’ Contributions

Concept and design: KH, KT, SM. Treatment planning creation: KH. Data analysis and interpretation: KH, KT, MW. Statistical analysis: MW. Important advice and critical discussion: KT, MW. Review of dose distribution (radiation oncologists): AM, KT. Research management and supervision: MI, TS. All authors read and approved the final manuscript.

Acknowledgements

The authors wish to express sincere gratitude to Yuichi Nagai, general manager of the Radiation Technology Department, for supporting the research facilities and providing the environment for conducting this research. We would like to express our deep gratitude to the Takaki Ariji and Hajime Oyoshi, Chief of Radiation Technology, for sharing his knowledge and skills and providing guidance in the treatment planning method for head and neck IMRT. We thank Edanz (https://jp.edanz.com/ac) for editing a draft of this manuscript.

Grégoire V, De Neve W, Eisbruch A, Lee N, Van den Weyngaert D, Van Gestel D. Intensity-Modulated Radiation Therapy for Head and Neck Carcinoma. Oncologist. 2007;12:555–64.
Chambers MS, Rosenthal DI, Weber RS. Radiation-induced xerostomia. Head Neck. 2007;29:58–63.
Hitchcock YJ, Tward JD, Szabo A, Bentz BG, Shrieve DC. Relative Contributions of Radiation and Cisplatin-Based Chemotherapy to Sensorineural Hearing Loss in Head-and-Neck Cancer Patients. Int J Radiation Oncology*Biology*Physics. 2009;73:779–88.
Pan CC, Eisbruch A, Lee JS, Snorrason RM, Ten Haken RK, Kileny PR. Prospective study of inner ear radiation dose and hearing loss in head-and-neck cancer patients. Int J Radiation Oncology*Biology*Physics. 2005;61:1393–402.
van der Geer SJ, Kamstra JI, Roodenburg JLN, van Leeuwen M, Reintsema H, Langendijk JA, et al. Predictors for trismus in patients receiving radiotherapy. Acta Oncol (Madr). 2016;55:1318–23.
Baudelet M, Van den Steen L, Tomassen P, Bonte K, Deron P, Huvenne W, et al. Very late xerostomia, dysphagia, and neck fibrosis after head and neck radiotherapy. Head Neck. 2019;41:3594–603.
Gujral DM, Nutting CM. Patterns of failure, treatment outcomes and late toxicities of head and neck cancer in the current era of IMRT. Oral Oncol. 2018;86:225–33.
Chen AM, Daly ME, Farwell DG, Vazquez E, Courquin J, Lau DH, et al. Quality of Life Among Long-Term Survivors of Head and Neck Cancer Treated by Intensity-Modulated Radiotherapy. JAMA Otolaryngology–Head Neck Surg. 2014;140:129.
Tribius S, Bergelt C. Intensity-modulated radiotherapy versus conventional and 3D conformal radiotherapy in patients with head and neck cancer: Is there a worthwhile quality of life gain? Cancer Treat Rev. 2011;37:511–9.
Barker JL, Garden AS, Ang KK, O’Daniel JC, Wang H, Court LE, et al. Quantification of volumetric and geometric changes occurring during fractionated radiotherapy for head-and-neck cancer using an integrated CT/linear accelerator system. Int J Radiation Oncology*Biology*Physics. 2004;59:960–70.
Ahn PH, Chen C-C, Ahn AI, Hong L, Scripes PG, Shen J, et al. Adaptive Planning in Intensity-Modulated Radiation Therapy for Head and Neck Cancers: Single-Institution Experience and Clinical Implications. Int J Radiation Oncology*Biology*Physics. 2011;80:677–85.
Bando R, Ikushima H, Kawanaka T, Kudo T, Sasaki M, Tominaga M, et al. Changes of tumor and normal structures of the neck during radiation therapy for head and neck cancer requires adaptive strategy. J Med Invest. 2013;60:46–51.
Heukelom J, Fuller CD. Head and Neck Cancer Adaptive Radiation Therapy (ART): Conceptual Considerations for the Informed Clinician. Semin Radiat Oncol. 2019;29:258–73.
Jensen AD, Nill S, Huber PE, Bendl R, Debus J, Münter MW. A Clinical Concept for Interfractional Adaptive Radiation Therapy in the Treatment of Head and Neck Cancer. Int J Radiation Oncology*Biology*Physics. 2012;82:590–6.
Castelli J, Simon A, Lafond C, Perichon N, Rigaud B, Chajon E, et al. Adaptive radiotherapy for head and neck cancer. Acta Oncol (Madr). 2018;57:1284–92.
Dewan A, Sharma S, Dewan AK, Srivastava H, Rawat S, Kakria A, et al. Impact of Adaptive Radiotherapy on Locally Advanced Head and Neck Cancer - A Dosimetric and Volumetric Study. Asian Pac J Cancer Prev. 2016;17:985–92.
Aly F, Miller AA, Jameson MG, Metcalfe PE. A prospective study of weekly intensity modulated radiation therapy plan adaptation for head and neck cancer: improved target coverage and organ at risk sparing. Australas Phys Eng Sci Med. 2019;42:43–51.
Luo Y, Qin Y, Lang J. Effect of adaptive replanning in patients with locally advanced nasopharyngeal carcinoma treated by intensity-modulated radiotherapy: a propensity score matched analysis. Clin Transl Oncol. 2017;19:470–6.
Delaby N, Barateau A, Chiavassa S, Biston M-C, Chartier P, Graulières E, et al. Practical and technical key challenges in head and neck adaptive radiotherapy: The GORTEC point of view. Physica Med. 2023;109:102568.
Chen AM, Daly ME, Cui J, Mathai M, Benedict S, Purdy JA. Clinical outcomes among patients with head and neck cancer treated by intensity-modulated radiotherapy with and without adaptive replanning. Head Neck. 2014;36:1541–6.
Castelli J, Simon A, Louvel G, Henry O, Chajon E, Nassef M, et al. Impact of head and neck cancer adaptive radiotherapy to spare the parotid glands and decrease the risk of xerostomia. Radiat Oncol. 2015;10:6.
Kumar A, Soni TP, Patni N, Jakhotia N, Singh DK, K R, et al. An Observational Dosimetric Study in Definitively Treated Primary Head and Neck Cancers: To Assess the Effect of Weight Loss and Change in Lateral Neck Dimensions on the Difference Between Dose Planned and Received by the Parotid(s) and Correlation with Adaptive Radiation Therapy. Adv Radiat Oncol. 2024;9:101446.
Ding GX, Munro P, Pawlowski J, Malcolm A, Coffey CW. Reducing radiation exposure to patients from kV-CBCT imaging. Radiother Oncol. 2010;97:585–92.
Ding GX, Munro P. Radiation exposure to patients from image guidance procedures and techniques to reduce the imaging dose. Radiother Oncol. 2013;108:91–8.
Ding GX, Alaei P, Curran B, Flynn R, Gossman M, Mackie TR et al. Image guidance doses delivered during radiotherapy: Quantification, management, and reduction: Report of the < scp > AAPM Therapy Physics Committee Task Group 180. Med Phys. 2018;45.
Hirotaki K, Moriya S, Tachibana H, Sakae T. Detection of anatomical changes using two-dimensional x‐ray images for head and neck adaptive radiotherapy. Med Phys. 2022;49:3288–97.
Brouwer CL, Steenbakkers RJHM, Langendijk JA, Sijtsema NM. Identifying patients who may benefit from adaptive radiotherapy: Does the literature on anatomic and dosimetric changes in head and neck organs at risk during radiotherapy provide information to help? Radiother Oncol. 2015;115:285–94.
Schwartz DL, Garden AS, Shah SJ, Chronowski G, Sejpal S, Rosenthal DI, et al. Adaptive radiotherapy for head and neck cancer—Dosimetric results from a prospective clinical trial. Radiother Oncol. 2013;106:80–4.
Castelli J, Thariat J, Benezery K, Hasbini A, Gery B, Berger A, et al. Weekly Adaptive Radiotherapy vs Standard Intensity-Modulated Radiotherapy for Improving Salivary Function in Patients With Head and Neck Cancer. JAMA Oncol. 2023;9:1056.
Brouwer CL, Steenbakkers RJHM, van der Schaaf A, Sopacua CTC, van Dijk LV, Kierkels RGJ, et al. Selection of head and neck cancer patients for adaptive radiotherapy to decrease xerostomia. Radiother Oncol. 2016;120:36–40.
Zhang P, Simon A, Rigaud B, Castelli J, Ospina Arango JD, Nassef M, et al. Optimal adaptive IMRT strategy to spare the parotid glands in oropharyngeal cancer. Radiother Oncol. 2016;120:41–7.
Gan Y, Langendijk JA, Oldehinkel E, Lin Z, Both S, Brouwer CL. Optimal timing of re-planning for head and neck adaptive radiotherapy. Radiother Oncol. 2024;194:110145.
Castelli J, Simon A, Rigaud B, Lafond C, Chajon E, Ospina JD, et al. A Nomogram to predict parotid gland overdose in head and neck IMRT. Radiat Oncol. 2016;11:79.
Castelli J, Simon A, Louvel G, Henry O, Chajon E, Nassef M, et al. Impact of head and neck cancer adaptive radiotherapy to spare the parotid glands and decrease the risk of xerostomia. Radiat Oncol. 2015;10:6.
Belshaw L, Agnew CE, Irvine DM, Rooney KP, McGarry CK. Adaptive radiotherapy for head and neck cancer reduces the requirement for rescans during treatment due to spinal cord dose. Radiat Oncol. 2019;14:189.
Kontaxis C, Bol GH, Lagendijk JJW, Raaymakers BW. Towards adaptive IMRT sequencing for the MR-linac. Phys Med Biol. 2015;60:2493–509.
Raaymakers BW, Lagendijk JJW, Overweg J, Kok JGM, Raaijmakers AJE, Kerkhof EM, et al. Integrating a 1.5 T MRI scanner with a 6 MV accelerator: proof of concept. Phys Med Biol. 2009;54:N229–37.
Lim-Reinders S, Keller BM, Al-Ward S, Sahgal A, Kim A. Online Adaptive Radiation Therapy. Int J Radiation Oncology*Biology*Physics. 2017;99:994–1003.
Chetty IJ, Fontenot J. Adaptive Radiation Therapy: Off-Line, On-Line, and In-Line? Int J Radiation Oncology*Biology*Physics. 2017;99:689–91.
Avgousti R, Antypas C, Armpilia C, Simopoulou F, Liakouli Z, Karaiskos P, et al. Adaptive radiation therapy: When, how and what are the benefits that literature provides? Cancer/Radiothérapie. 2022;26:622–36.
Gros SAA, Xu W, Roeske JC, Choi M, Emami B, Surucu M. A novel surrogate to identify anatomical changes during radiotherapy of head and neck cancer patients. Med Phys. 2017;44:924–34.
Sun B, Yang D, Lam D, Zhang T, Dvergsten T, Bradley J, et al. Toward adaptive proton therapy guided with a mobile helical CT scanner. Radiother Oncol. 2018;129:479–85.
Albertini F, Matter M, Nenoff L, Zhang Y, Lomax A. Online daily adaptive proton therapy. Br J Radiol. 2020;93.
Mackay K, Bernstein D, Glocker B, Kamnitsas K, Taylor A. A Review of the Metrics Used to Assess Auto-Contouring Systems in Radiotherapy. Clin Oncol. 2023;35:354–69.

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

Prediction and monitoring of adaptive radiation therapy timing using two-dimensional X-ray image-based water equivalent thickness

Status:

Version 1

Abstract

Background

Methods

Results

Conclusions

Figures

Background

Methods

Results

Discussion

Conclusions

Abbreviations

Declarations

References

Additional Declarations

Status:

Version 1