Feasibility of radiosurgery for patients with spinal tumors treated in lateral decubitus position: A case series from Memorial Sloan Kettering Cancer Center

Abstract

Introduction

Often in clinical practice radiation oncologists encounter patients who require treatment to the spine commonly in the setting of metastatic disease. These metastases usually cause pain, immobility, or neurologic deficits mandating expedited therapy to alleviate the suffering of our patients. Spine radiosurgery techniques have been used extensively for palliation purposes; however, given the patients’ deteriorating condition or pain and inability to tolerate anesthesia the radiation oncologist is often left with the conundrum of how to best set up his or her patient in preparation for radiosurgery if supine is not a viable option. In the Memorial Sloan Kettering Cancer Center several patients have been treated successfully in the lateral decubitus position to overcome this set-up issue. In this report, the feasibility of the lateral decubitus set-up for patients who benefit from radiosurgery to the spine when and if they cannot tolerate standard supine position is explored.

Objective

To report on a retrospective case series of three patients with a total of four lesions who were treated with radiosurgery for spinal metastases while set up in the lateral decubitus position.

Methods and materials

This is a retrospective case series of 3 patients who were treated with radiosurgery to the spine for palliation of painful metastatic foci. Patients were treated in the lateral decubitus position in 1-5 fractions in order to be eligible for this retrospective case series. Their set-up data, and clinical outcomes were then compared with historic controls.

Results

Patients who were treated in the lateral decubitus position were set up reliably and reproducibly. Additionally clinical outcomes on routine follow-up and imaging, and toxicity profiles also corroborated the utility of this treatment set-up.

Conclusions

Routinely employing optical surface tracking during patient setup followed by KVCBCT prior to treatment delivery along with intra-fractional monitoring is safe and effective while utilizing the lateral decubitus position for the treatment of spinal metastases for patients who cannot tolerate the supine position. Finally the patient follow-up also corroborated that treatments were successful thus lending credence to the safety, ease, effectiveness, and feasibility of this patient set-up.

Introduction

Metastatic spinal cord disease affects between 5-15% of all patients with cancer and there are more than 20,000 cases diagnosed in the United States annually [1-2]. Untreated spinal cord metastases inevitably progress causing paralysis, sensory loss, incontinence, pain [3], and certainly a decreased quality of life. While most patients will succumb to their diseases within the first year of diagnosis of metastatic spinal disease [4] there is a growing population of long-term survivors of metastatic spinal disease, which is a cogent argument for treating these patients with radiosurgery versus conventional external beam radiation in order to administer higher BEDs with more precise targeting especially in the setting of radioresistant histologies and often in the setting of re-irradiation. In general the main presenting symptom of metastatic disease in the spine is pain. Spine SRS/SBRT has been used for many years with great success to control patient’s pain, neurologic deficits and to restore patients’ quality of life therefore if supine positioning is not possible another position should be utilized so as to be able to proceed with therapy.

There are many compelling advantages favoring non-invasive radiosurgery techniques for spinal metastases compared to conventional external beam fractionation such as rapid and durable pain control and bone marrow preservation with tighter margins in likely already debilitated patients [4]. In addition, patient convenience plays a great role as well when choosing to employ a radiosurgery technique. Also, expediting the completion of radiation enables patients to receive systemic therapy as quickly as possible. When radiosurgery is employed, patients generally meet with the radiation oncologist on day one, are simulated on day two, and receive/complete their treatment on day three (if single fraction schedule is planned). This is in stark contrast to the most commonly used schedule in the United States of 30Gy in 10 fractions [4-6]. Convenience and patient comfort are of utmost importance when palliating patients’ pain which is also why radiosurgery has gained such momentum in these cases. Additionally it is widely believed that a painful or uncomfortable set up could cause an increase in the overall time that the patient is on the treatment couch or simulator therefore a further step that at times may be taken is offering the patient his/her treatment in the lateral decubitus position especially given that a comfortably resting patient is less likely to fidget possibly increasing time on the simulator or treatment couch and likely decreasing fidelity.

Now that patients have their set-up with standard laser alignment supplemented by treatment setup verification with optical surface tracking which is then verified with KV cone beam CT prior to delivery of treatment plus intra-fractional patient monitoring, we are confident that our treatment delivery is accurate on the order of 1-2mm despite whether or not the patient is set up in the supine or lateral decubitus position and therefore we have employed this technique when it is occasionally necessary.

Historic controls demonstrate efficacy of pain control with radiosurgery for spinal metastases with a median of approximately two weeks till pain control [4]. In another report by Ryu, pain relief was found to last a median of 13.3 months [6]. In other reports control of the spinal lesions was achieved in greater than or equal to 90% of patients [4, 7-9].

Materials and Methods

In this retrospective case series we studied patients who underwent radiosurgery to the spine in 1-5 fractions while treated in the lateral decubitus position. Our patients were CT simulated in the lateral decubitus position with standard isocenter and alignment tattoos with an alpha cradle from neck to ankles formulated. In order to achieve precise localization of tumor and spinal cord with reliable positioning and immobilization, patients were also placed into a custom-made non-invasive rigid fixation device that was formulated by one of our physicists that has four adjustable paddles (two anterior and two posterior) that can be fixed so as to provide support to the patient’s back and front and ensure maximal reproducibility (please refer to figures 1-3) [10]. In addition patients underwent a CT myelogram prior to CT simulation to help contour the spinal cord.

When planning these cases in general the spinal cord dose is limited to 12Gy or 14Gy if GTV coverage is poor. The esophagus’ dose is limited to 14Gy to 2.5cc, and a heart max point dose of 22Gy. In terms of dosimetry, the lateral decubitus plans were able to achieve all of our standard planning criteria for one example of our planning scan with isodose lines depicted (Figure 4).

Figure 4.

planning scan that depicts the isodose lines for patient number 2 who received 2400cGy in one fraction (see red isodose line).

Figure 1.

(a) Custom built adjustable paddles keeping patient #1 in place during simulation and treatment with alpha cradle extending from chin to ankles.

Once the patient was placed on the treatment machine, optical surface tracking was utilized to ensure that the patients’ outer contour was acceptable and that the patients’ set-up could proceed to the next steps. (figures 1-2). At this point a KV cone beam CT was performed and the bony as well as soft tissue anatomy was verified to overlap on both the reference and localization scans. Then after appropriate shifts were made if needed, ExacTrac® was utilized to verify and then maintain overlay of bony anatomy. Then if the set-up appeared acceptable the dose was delivered if no further shifts were needed (see Figure 3). Patients were then treated with IGRT-IMRT with either 6 or 15MV photons with seven to nine static beams with dynamic MLCs or with VMAT. For details of patient characteristics, clinical information please see Table 1-2.

Figure 3.

KVCBCT prior to delivery of dose of patient #2 Note that patient had disease in right paraspinal area that is well covered in the PTV. Note, zero margin of PTV into spinal cord as contoured based on myelogram.

Table 1.

Patient Details

Patient details	Patient 1	Patient 2	Patient 3A	Patient 3B
patient age	60	40	67	67
Patient sex	female	female	male	male
KPS prior to therapy	70-80%	70%	unknown	unknown
Primary tumor site	Endometrium	rectal	Thymic carcinoma	Thymic carcinoma
Histology (resistant vs. sensitive)	intermediate	intermediate	intermediate	intermediate
Spinal location of disease	L2-L5	T5/T6	T2	T6
RT dose per fraction	600cGy	2400cGy	600cGy	600cGy
RT total dose	3000cGy	2400cGy	3000cGy	3000cGy
Number of fractions	5	1	5	5
Primary or secondary RT	secondary	primary	Secondary (pt had Hodgkin Lymphoma treated in 1969)	Secondary (pt had Hodgkin Lymphoma treated in 1969)
Toxicity	Grade 1 dry desquamation Grade 2 fatigue	none	none	none

Figure 2.

optical surface tracking for patient #2. On the right of the figure is the reconstruction of the patient’s outer contour with the patient facing to the right with the right arm seen crossing the chest below her face. On the left of the figure are the positioning parameters that must be met in order to allow the treatment to proceed. The green bars on the left indicate that they have been met optically and treatment may commence.

We then reviewed the individual fractions of radiation delivered for each patient and studied the amount of time that the patient was in the treatment room, the number of KVCBCTs that were obtained and the number of shifts as well as the extent of each shift made that were necessary to successfully treat the patients’ lesions in order to assess whether or not their set up was preasonable. These data were obtained primarily by reviewing the brain lab reports.

Results

On average we found that the time on the machine per treatment lasted 33 minutes (range 16-76 minutes) with trends towards quicker on-machine-time as the patients’ number of fractions increased with the first treatments lasting longest. Next we looked at the shifts that were required prior and during therapy using orthogonal x-rays as well as KVCBCT and we found that on average each patient underwent approximately 2 KVCBCT per fraction (range 1-4) along with an average of 1.5 shifts per fraction. A shift took into account vertical/longitudinal/lateral/roll/pitch/yaw changes that were required to match the patient’s localization scan with his or her reference CT scan as well as the orthogonal x-rays with the planning DRRS. On average for all sixteen fractions in this retrospective case series there was a vertical shift required of approximately 5.1mm, as well as longitudinally 6.2mm, laterally 8.1mm, and rotationally 0.82degrees. Please see Table 3 for raw data taken from each patient’s chart.

Table 3.

On Treatment Machine Details

Patient	# KVCBCT	# SHIFTS	#MV images	# orthogonals	Vertical shifts (cm)	Longitudinal shifts (cm)	Lateral shifts (cm)	Rotational shifts (degrees)	Avg all shifts performed (cm)	Avg. rotation (degrees)	Apprx time on machine (minutes)
#1 (L2-L5)
1st	3	2	2	2	+0.99 -0.19	-0.33 -0.68	+0.52 -0.05	n/a	0.3		50
2nd	1	1	1	1	-0.87	-0.39	-0.39	n/a	.55		20
3rd	3	2	2	2	+0.51 -0.06	+0.31 -0.35	+2.17 -0.52	n/a	1.0		36
4th	2	1	2	1	-0.07	+0.50	-0.31	n/a	.3		24
5th	2	2	2	2	-0.41 +0.45	+0.07 -0.80	+3.11 -0.20	n/a	1.7		34
#2 (T5/T6)
	3	3	1	7	-0.60 +0.08 +0.17	-1.35 +0.10 +0.11	+0.94 +0.24 +0.66	+1.8º -1.0 º 0 º	.48	1.4 º	56
#3A. (T6)
1st	2	1	1	5	+0.337	+0.281	-0.30	0 º	3.07		55
2nd	1	1	0	3	+1.08	+0.186	+0.922	0 º	0.13		18
3rd	2	2	0	1	-0.53 +0.65	-0.115 -.02	+0.57 +0.57	0 º	0.6	0 º	22
4th	2	1	Not avail	Not avail	-0.58	+0.09	+0.02	+4.2 º	0.5	4.2 º	22
5th	4	1	Not avail	Not avail	-1.01	-0.21	-0.03	+5.0 º	1.0	5.0 º	41
#3B. (T2)
1st	7	6	1	2	+0.36 +0.11 -0.03 -0.17 +0.27 +0.46	-0.56 +0.23 -0.07 +0.53 -0.39 -0.19	+4.18 +0.03 +0.08 +3.43 +0.09 +0.78	3.0 º +2.5 º -2.5 º 0 º +0.8 º -0.9 º	0.67	1.6 º	76
2nd	2	2	0	1	-0.506 +0.397	+0.625 -0.052	-0.639 -0.347	0 º	0.4		22
3rd	2	1	Not avail	Not avail	-0.860	+1.228	+0.775	0 º	0.95		20
4th	1	1	Not avail	Not avail	-0.779	+0.86	-0.107	0 º	0.58		23
5th	1	1	Not avail	Not avail	-0.267	+0.748	-0.177	+1.2 º	0.39	1.2 º	16

Patients treatment and outcomes

All patients were treated in the lateral decubitus position. For patient 1 with metastatic endometrioid adenocarcinoma of the uterine corpus, she received radiosurgery 3000cGy in 5 fractions to L2-L5 from 12/16/2011 to 12/22/2011 and she passed away on 10/30/2012 from a pulmonary embolism. This patient was found to have responded very favorably to her spine radiosurgery clinically and she was able to stop all narcotics and had an increase in her KPS of ten points. Patient 2 had metastatic rectal cancer and underwent spine radiosurgery on 5/23/2014 to T5/T6 (single fraction 2400cGy x1) and she passed away from metastatic disease in Sept. 2015) with stable disease in the radiosurgery (T5/T6) volume [see Figure 1 (b) and 4 for the set-up, isodose lines, and on treatment images that demonstrate the successful execution of this plan]. The third patient was referred for metastatic thymic carcinoma and completed radiosurgery 3000cGy in 5 fractions to two separate sites, T6 and T2 over the course of 1/23-1/29/2014 and 1/30-2/5/2014 respectively This patient passed away in April, 2014 from hemorrhagic brain metastases treated at an outside hospital. Patient outcome details are summarized in tables 1-2.

Figure 1.

(b) demonstrating the custom built rigid non-invasive stabilizing device with alpha cradle underneath. Patient #2 is seen lying in lateral decubitus position. The Reddish glow on the patient’s skin (in color photograph) is the optical surface tracking helping to ensure accurate patient set-up.

Patient as well as tumor and prior treatment characteristics in this small case series were very similar to our patients who are generally treated with radiosurgery for spinal metastatic disease while lying in the supine position and all initially presented with back pain. In Memorial Sloan Kettering Cancer Center’s spine clinic, pain control (and current medications) are documented on routine history and physical as well as during follow-up and based on review of the follow-up documentation, the treatments led to excellent control of pain and minimal side effects, lending credence to the excellent dosimetry and patient set-up. Pain control was rapidly achieved with patients decreasing their narcotic intake along with either increasing or stable KPS with excellent palliation of symptoms with durable responses. In no case has the metastatic spinal disease been the cause of the patients’ demise, rather it is the systemic illness that results in the patients’ ultimate decline.

Compared with historic data [4, 11-12] our outcomes data compare well with prior reports with respect to overall survival and local control and certainly compare favorably with an historic paper that compared surgery and radiation vs. radiation alone for metastatic spinal cord compression and found that the better arm (surgery and radiation) had a median overall survival of just 126 days [12]. While our follow-up is short and our case series is small our results are encouraging nevertheless as they demonstrate a survival of approximately one year except for one patient who succumbed to progression of systemic disease despite local control of spinal disease.

Our results demonstrate feasibility of delivery of radiosurgery in the lateral decubitus position based on treatment review, clinical outcomes, as well as dosimetric profiles. Additionally, treatment in the lateral decubitus position has been reported for other body sites, such as whole breast [13] with positive reports of dosimetry and outcomes and potentially decreased toxicity. Our case series is the first to report on this set-up position for patients who undergo radiosurgery for metastatic disease to the spine. Our report suggests that in certain clinical situations that is when patients cannot tolerate traditional supine set-up, lateral decubitus position is a reasonable option given the findings of our dosimetry, encouraging optical surface tracking data, and good correlations between KV cone beam CT obtained prior to delivery of dose and initial planning CT scan in addition to excellent clinical response to treatment. While this report represents a small case series, these results are encouraging as to the safety, feasibility, and efficacy of this treatment set-up.

Conclusions

Despite the short follow-up of just 3 patients with four lesions, we believe that we have sufficient evidence to support the safety and feasibility of utilizing the lateral decubitus position for a patient who cannot tolerate the supine position. Given the favorable clinical outcome (rapid pain relief with minimal side effects and durable pain control), as seen in our case series, comparable dosimetric profiles, relative ease of simulation and reproducibility with KVCBCT and intra-fraction motion monitoring, this treatment technique will be employed as needed at our center for patients who cannot tolerate the supine position.

While the lateral decubitus position may pose a challenge for all involved in the care of these patients, it may impart a large benefit to the patient who can undergo his/her radiosurgery more comfortably in this position. In terms of patient set-up, the lateral decubitus position was found to be reproducible with small errors easily detected and corrected prior to delivery of dose. In addition, now that flattening filter free mode has gained wider acceptance, we can further reduce treatment time. This report also raises the possibility of whether treating spinal tumors with protons (on a 90 degree static gantry) is feasible with patients treated in the lateral decubitus position.

Table 2.

Endpoints

Endpoints	Patient 1	Patient 2	Patient 3A	Patient 3B
Complete pain relief	YES	YES	YES while on same pain meds as prior to RT	YES while on same pain meds as prior to RT
Notes	By 2nd month patient stopped almost all narcotics. initially unable to walk, –post-treatment walks with walker, continues to have residual LE weakness	Stopped taking pain medications, had progressive dz in lungs and brain and received palliative conventional RT to both sites	Patient continued taking tramadol and fentanyl patch at 2 months post –RT	Patient continued taking tramadol and fentanyl patch at 2 months post –RT
KPS post-therapy	80%	80%	80%	80%
MRI Response	Stable disease	Stable disease	Stable disease	Stable disease

Complete pain relief: pain score of 0 at three months post treatment

Partial pain relief: reduction in pain score of ≥3 (improvement of at least 3 points) and no increase in narcotics

Stable response: pain score within 2 points of baseline

Progressive response: increase of at least 3 points from baseline.