Liver-directed stereotactic body radiotherapy can be reliably delivered to selected patients without internal fiducial markers—A case series

Abstract

Tumor motion and the lack of tissue-tumor contrast have been challenging parts of liver-directed stereotactic body radiotherapy (SBRT). In this study, we investigated the possibility of liver-directed SBRT without internal fiducials using breath hold technique and diaphragm matching technique. One hundred thirty-four volumetric images of 13 consecutive patients with either primary or metastatic liver tumors who underwent expiratory breath hold SBRT were compared and analyzed. Reproducibility of diaphragm position between fractions relative to bone was evaluated on image registration software. At median follow-up time of 13 months, 1-year and 2-year local control rates of index lesions were 90% and 72%, respectively. In comparison to diaphragm matching, a greater margin is required for bone matching technique for that 19 of 67 (28%) of all interfractional SI offsets were more than 6 mm, whereas 6 of 67 (9%) intrafractional SI exceeded 6 mm. Despite the small study size, our study showed that breath hold SBRT without internal hepatic fiducial is a valid approach for selected patients.

1. INTRODUCTION

Hepatic tumors, either primary or metastatic malignancies, can be effectively treated with stereotactic body radiotherapy (SBRT).¹ Yet liver-directed SBRTs remained technically challenging for the combination of lack of tissue contrast on computed tomography (CT) and respiration related tumor motion.²

Breath hold technique has been proposed to minimize organ motion, with expiratory breath hold being superior in terms of reproducibility.^3-5 Surrogate markers for tumor positioning during treatment such as invasive internal fiducial and radio-opacities either has inherent risks or are not readily available.^6-8 The liver dome, however, is always present, and has been shown to correlate with actual tumor position.⁹

Clinical outcomes on liver-directed SBRTs without implant fiducial or radio-opaque markers at index lesions are lacking. The purpose of this study is to illustrate the potential of liver-directed SBRTs using diaphragm as surrogate tumor positioning under expiratory breath hold.

2. METHODS

Institutional Research Ethics Board approval was obtained prior to study initiation (institution and certification number: IRB2019087). For this case series, electronic medical records and images of 13 consecutive patients, who underwent hepatic SBRT without internal fiducial between January 2018 and December 2020, were retrospectively collected. The institutional indication for hepatic SBRT was in accordance the reimbursement policies of the National Health Insurance program. In general, SBRT were offered when alternative locoregional treatment were considered inadequate on multidisciplinary board. Primary and oligometastatic tumors, defined as no more than five metastatic lesions body wide and no more than the involvement of three organs, were included.

Breath hold technique was done with Abches (APEX medical Inc., Tokyo, Japan), a breath cycle meter that consists of one main assembly with two sensing fulcrums and one range indicator, and a head-mount mirror (Fig. 1). This device provides real time visual feedback and has minimal compatibility issue.

Fig. 1.

The main assembly and head-mount reflective mirror (A). The dial on range indicator as viewed from the reflective mirror (B). Inhalation point (C) End-expiratory breath hold position (D).

2.1. End-expiratory breath hold technique and monitoring

Patient was simulated in supine position with arms above head. All CT scans were acquired with verbal stop-and-go technique, with patient resuming to regular breathing rhythms in between breath holds. Patients were instructed to hold their breath at a set range on the display (Fig. 1C,D). The process repeats itself till the end of image acquisition. All patients had triphasic simulation scans on breath holds in additional to simulation scan for target delineation.

2.2. Treatment planning and radiotherapy technique

Gross tumor volumes (GTV) were defined as tumors identified on simulation CTs. There was no expansion from GTV to clinical target volume (CTV). A 3-5 mm expansion was added to form the planning target volume (PTV). Dose prescriptions range from 36Gy in six fractions to 50Gy in five fractions. The highest allowable doses were prescribed while maintaining normal tissue constraints summarized in Table 1.

Table 1.

Treatment planning dose constraints of organ at risk

Organ	For 5 and 6 Fraction SBRT
Heart	Dmax < 32 Gy
Great vessels	Dmax < 52.5 Gy
Chest wall	D2cc < 32 Gy
Skin	Dmax < 32 Gy
Uninvolved liver	>700 mL of uninvolved liver < 15 Gy
Portal hepatis	Mean dose < 19 Gy
Doudenum	Dmax < 30 Gy
Small bowel	Dmax < 30 Gy
Stomach	Dmax < 30 Gy
Esophagus	Dmax < 35 Gy
Spinal cord	Dmax < 25 Gy

Elekta Versa HD with 6MV flattening-filter free photon beams (with dose rates up to 2400 MU/min) and on-board cone-beam CT (CBCT) delivered all treatments. Three volumetric CBCT images without contrast enhancement were acquired at each treatment fraction. The first CBCT images were used for setup correction to bony structures. The next CBCT were to reassure the correction and to make necessary shifts with respect to the diaphragm position. The last CBCT was done after treatment and was used to identify intrafractional changes on diaphragm position.

2.3. Interfractional reproducibility with bone matching technique

Position of the diaphragm may differ between fractions. Interfractional reproducibility was defined as the reproducibility of diaphragm dome positions between the planning CT and the first CBCT of each treatment fraction. A comparison between diaphragm matching and bone matching techniques using offline image registration was done to identify discrepancy in superior-inferior (SI), medial-lateral (ML), and anterior-posterior (AP). First, planning CT and first CBCT images of each fraction were registered using automatic image registration in the on-board viewing software. Vertebral bodies were matched to correct setup error with 6-dimensions of freedom. Then, the topmost part of the diaphragm on CT images was compared for translations in all orthogonal planes. Translation between the diaphragm dome positions on the two sets of images was compared.

2.4. Intrafractional reproducibility

The diaphragm position may differ between the start and end of the treatment. Using the same technique as described on preceding paragraphs, the position of the liver dome between the second and last CBCTs were compared for intrafractional translations.

2.5. Follow-up

Responses to treatment were evaluated on the modified Response Evaluation Criteria in Solid Tumors (mRECIST) for HCCs and RECIST for hepatic metastases.^10,11 Radiological studies were arranged at 3-month intervals and as clinically indicated. Local failure in target lesions was defined as increase in tumor dimension and an increase in tumor marker in the absence of new lesions.

2.6. Statistical analysis

Assessments of the offsets in all three orthogonal were expressed with group mean and standard deviations, along the absolute values. For reporting of local control rates with Kaplan-Meier analysis, we used R (version 3.2.1, R foundation for Statistical Computing, Vienna, Austria), with the KMsurv, survfit, and survival packages.

3. RESULTS

All of the 13 enrolled patients had Child-Pugh score A liver cirrhosis score and were treated at either recurrence or disease progression. Thirteen target lesions were treated. Relevant patient demographic is summarized in Table 2. Response rates were summarized in Table 3. Fig. 2 demonstrates cumulative local control rate of the target lesions. The median follow-up time was 13 months (range: 1.5 to 29 months). 1-year and 2-year local control rates of target lesion is 90% and 72%, respectively.

Table 2.

Clinicopathologic characteristics

Variable	Value
Patient (n)	13
Age (years)
Range	38–83
Median	57
Gender
Male	8
Female	5
Number of lesions treated per patient (n)
1	13
Primary cancer (n)	9
Metastatic cancer(n)	4
Pathology
HCC	9
Adenocaricinoma	3
Others	1
Primary site
Liver	9
Other gastrointestinal site	3
Uterine cervix	1
GTV volume (cc)
Median (range)	16 (5–50)
PTV volume (cc)
Median (Range)	43 (13–96)
Normal liver volume (cc)
Median (Range)	1135 (723–1588)
Number of treatment fractions (n)
4	1
5	9
6	3
Prescribed dose (Gy)
Range	36–50

GTV = gross tumor volume; PTV = planning target volume.

Table 3.

Tumor response of target lesion, as evaluated with RECIST and mRECIST

	mRECIST	Number (%)	RECIST	Number (%)
Primary liver tumors (n = 9)	CR	2 (22%)	NA
	PR	6 (67%)
	SD	1 (11%)
	PD	0 (0%)
Hepatic metastasis (n = 4)	NA		CR	2 (50%)
			PR	2 (50%)
			SD	0 (0%)
			PD	0 (0%)

CR = complete response; mRECIST = modified Response Evaluation Criteria in Solid Tumors; PR = partial response; PD = progressive disease; SD = stable disease.

Fig. 2.

Local control rate of target lesion with a median follow up of 13 months.

For interfractional reproducibility of liver dome relative to bone, 67 treatment CBCTs from 67 treatment fractions scans were compared to their simulation counterparts. Meanwhile, a total of 134 CBCTs from 67 treatment fractions were analyzed for intrafractional reproducibility. The mean and standard deviations of the translations were summarized in Table 4. Fig. 3 is a frequency chart of SI offsets in absolute value. Nineteenth of 67 (28%) of all interfractional SI offsets were more than 6 mm, whereas 6 of 67 (9%) intrafractional SI exceed 6 mm.

Table 4.

Interfractional and intrafractional liver dome reproducibility with end-expiratory breath hold

		Offsets (mm)
		SI	AP	ML
Average interfraction reproducibility	Mean	0.6	0.1	0.3
	SD	3.6	0.7	0.5
	Abs. mean	4.6	0.6	0.5
	Abs. SD	3.1	0.7	0.6
Average intrafraction reproducibility	Mean	1.2	0.1	-0.3
	SD	3.0	1.1	0.6
	Abs. mean	3.0	0.8	0.4
	Abs. SD	2.1	0.9	0.5

Abs = absolute; AP = anterior-posterior; ML = medial-lateral; SD = standard deviation; SI = superior-inferior

Fig. 3.

Frequency chart of fractional SI offsets in 3 mm intervals. (A) interfractional SI offsets (B) intrafractional SI offsets. SI = superior-inferior.

4. DISCUSSION

This study investigated the feasibility of liver-directed SBRT without the use of fiducial marker and reported on the resulting local control rate, which is often lacking in previous publication addressing breath hold techniques.^3,12-14 In this study, a 1-year local control rate at 90% compared favorably to reported series on primary and metastatic liver tumors.^1,15 There are rational explanations for our result. First, all patients in this study were treated with diaphragm matching technique. It has been shown that the diaphragm is a more reliable surrogate for tumor positioning as compared to bone.¹³ Second, breath hold technique has helped minimizing tumor motion and the resulting underdosing.¹⁶ Even for primary HCCs, a tumor with modest radiosensitivity, underdosing remains an issue.¹⁷ Nevertheless, the vicinity of organ at risk to the target lesion still complicated dose delivery. For the two patients who developed local failure in the target lesions, the lesions were in proximity of the heart, a major dose limiting organ.

We would like to emphasize that this study is not to encourage omission of fiducial implantation for all patients, but to illustrate the potential application of SBRT to patients who did not get fiducial markers. One shortcoming of this study is the number of patients analyzed. This is a direct result of patient selection and breath hold technique. Breath hold at end expiration may not be tolerable for those who require SBRT, accounting for two-thirds of patient requiring SBRT.³

Hepatic tumors have the greatest movement in the superior-inferior plane as the diaphragm is the dominant driver of respiration.^{18, 19} Shifts in the ML and AP planes were modest. In our study, the absolute mean and standard deviations in the SI direction were 4.6 ± 3.1 mm and 3.0 ± 2.1 mm for interfractional and intrafractional reproducibility, respectively. A higher proportion of interfractional SI offsets exceeding 6mm (28% vs 9%), which readily exceed the 5mm expansion from CTV to PTV. This discrepancy was described by Dawson et al using orthogonal imaging.^{3, 12} Our study has confirmed their findings with contemporary volumetric imaging technique and highlighted the need for Individualized monitoring.

To conclude, fiducial-free liver-directed SBRT is feasible for selected patients. The combination of expiratory breath hold and diaphragm matching techniques is encouraging.