Fluoroscopy Dose and Time During Vertebral Augmentation for Spine Pain Due to Malignant Fractures

Simple Summary

Vertebral augmentations are procedures used to alleviate pain from vertebral fractures, including those caused by cancer. These procedures are most performed under fluoroscopy guidance. Therefore, radiation exposure for both patients and health care providers is unavoidable. Historically, the risk of medical complications arising from radiation exposure in VA procedures is not clear. Some studies have measured the level of radiation dose and time encountered during these procedures in osteoporotic compression fractures. However, there are no data for radiation exposure in patients with compression fractures associated with cancer. In these observational cohort data, several factors proper to cancer population seemed to correlate to a higher time and dose of radiation. This included elevated body mass index, the presence of sclerotic bone lesions proper to certain tumors, the complexity of the compression fractures, and the higher bone density of male and younger patients. Other variables are attributed to the availability of the resources in each setting and the experience of the providers.

Abstract

Background: Vertebral augmentation (VA) procedures are used to treat painful vertebral fractures caused by malignancies, but there are few data on the radiation exposure for patients and proceduralists during these VA procedures. We retrospectively examined the radiation dose exposure during VA procedures and defined the characteristics of patients who underwent such procedures. Methods: We conducted a retrospective observational cohort study including patients with cancer who experienced axial back pain from compression fractures caused by malignancies. Participants were identified using an electronic medical records database and must have had evidence of stable vertebral compression fractures upon imaging and documentation of a clinical evaluation. We collected data on patient demographics, fluoroscopy time (FT) and dose (FD) during the procedure, the volume of polymethylmethacrylate injected, and reported complications. Results: Overall, 140 patients were included. Their median age was 69, and they were mostly men (n = 79). The most common diagnosis was multiple myeloma (41.4%). Most patients had a single-level compression fracture of the thoracolumbar spine. The mean FT was 233.80 s, with higher FTs for patients with an elevated body mass index and patients younger than 60 years. The average FD was 157.98 mGy, with higher FDs for patients with an elevated BMI and for male patients. Pain relief was not associated with FT or FD. Conclusions: Patients with cancer who underwent VA experienced longer FT and higher FD compared to their non-cancer counterparts in the literature. However, we found multiple confounders for this relationship.

1. Introduction

Vertebral augmentation (VA) procedures, e.g., vertebroplasty and kyphoplasty with and without expandable intravertebral implants (EII), are used to treat painful vertebral fractures, including those caused by malignancy. VA involves injecting polymethylmethacrylate (PMMA) or other substances into the fractured vertebral body, stabilizing the fracture, and relieving the associated pain.

Image guidance is necessary to ensure proper placement of PMMA within the anterior third of the vertebral body. To this end, the most used imaging guidance is fluoroscopy, which is performed by trained fluoroscopy technicians. Image-guided VA is an advanced procedure that can be lengthy, resulting in longer fluoroscopy times (FTs) and, thus, higher total fluoroscopy doses (FDs) compared to most interventional fluoroscopy-based procedures for spinal pain. Such increases in FT and FD are especially prevalent at our institution, a teaching hospital, given the high number of trainees that perform VA procedures under the direction of an experienced interventionalist [1]. Although several studies have measured the level of radiation exposure during the most common fluoroscopy-guided procedures for non-malignant spine pain, such as epidural steroid injections or facet joint injections [2,3], there are few data on the radiation exposure for patients with cancer who undergo VA procedures for painful pathologic vertebral fractures. Furthermore, the risk of medical complications arising from radiation exposure in VA procedures for both the patient and the proceduralist is not entirely clear.

Thus, we evaluated the level of radiation exposure during VA procedures for patients with cancer who have spine pain from pathologic fractures, as determined by reviewing the FT and FD documented in their medical records. We also characterized patient and fluoroscopy characteristics, such as patient demographics, level of augmentation, VA technique, and pedicle approach. These data may assist interventional pain providers in determining if the level of the radiation risk in their own VA practice is within an acceptable range. It can also further help elucidate safety strategies for decreasing radiation exposure for both patients and health personal in the procedure suites.

2. Materials and Methods

2.1. Study Setting and Population

This retrospective cohort study was performed at The University of Texas MD Anderson Cancer Center, a tertiary care academic care center. Longitudinal patient data are regularly collected during cancer treatment. Patients were selected from an electronic medical record database for patients whose progress was monitored by our Department of Pain Medicine. On presentation, all included patients had a chief complaint of axial back pain secondary to compression fractures of the thoracic or the lumbar spine due to metastatic bone lesions. Our institutional review board approved the study with a waiver of consent (protocol #2022-0781).

2.2. Selection of Patients

Using medical billing codes and codes from the International Classification of Diseases, 9th and 10th editions, we searched our clinical database for patients who underwent VA procedures between 30 January 2018 and 30 June 2022. The patients received care from various oncology subspecialists.

Eligibility criteria included a cancer diagnosis; evidence of stable thoracic or lumbar vertebral compression fracture upon computed tomography (CT) or magnetic resonance imaging (MRI) of the spine; a clinical evaluation by our pain specialists confirming axial pain originating from compression fracture(s) and supporting the need for VA; and a VA procedure performed in an operating room. In all, 176 patients were identified. Of these patients, 140 met the inclusion criteria and were enrolled in our study (Figure 1).

2.3. VA

VA, sometimes referred to as cementoplasty, refers to the transcutaneous injection of PMMA, an elastic silicone-based polymer into the fractured vertebral body. The procedure is often performed with fluoroscopic guidance; however, VA injections in smaller bone structures, such as those in the cervical and upper thoracic spine (namely, T1–T5), sometimes require CT guidance. However, such modality is rarely available or accessible for specialists outside of Interventional Radiology. Thus, in this study, we only included fluoroscopy-based VA procedures. At our institution, technicians operated the C-arms using alternating uniplanar anterior–posterior and lateral views.

2.4. Variables Studied

During a thorough medical record review, we collected patient demographics, including sex, age, and cancer diagnosis. FT was measured from initial to final usage of fluoroscopy during the procedure, and FD was measured as the total dose of radiation issued (these variables are automatically registered on the fluoroscopy database as part of the medical records). To determine what variables, if any, were associated with increased VA radiation exposure, we collected data on body mass index (BMI), which are collected as part of the general medical assessment on all patients. As such, 25 or higher is considered overweight, and a BMI of 30 or higher is considered obese. We also reviewed the vertebral levels augmented; the type of procedure performed (i.e., vertebroplasty or kyphoplasty with or without expandable intravertebral implants); the approach used (i.e., unipedicular, bipedicular, or extra-pedicular); the volume of PMMA injected; and any immediate reported complications.

2.5. Statistical Analysis

Summary statistics, including mean, SD, median, and range, for continuous variables, such as age as well as frequency counts and percentages for categorical variables, such as sex, are provided. The Wilcoxon signed-rank test was used to evaluate statistical differences in pain scores before and after VA. The Fisher exact or Chi-square test was used to assess associations between categorical variables. The Wilcoxon rank-sum test was used to assess the difference in a continuous variable between patient groups. SAS, version 9.4 (Cary, NC, USA), was used for all the analyses.

3. Results

3.1. Study Group

Of the 176 patient medical records initially reviewed, 36 were excluded owing to incomplete documentation of FT or FD. These latter gaps were noteworthy. The final study population, therefore, consisted of 140 patients (Figure 1). All patients had at least one symptomatic compression fracture of the thoracic, lumbar, or sacral spine, which was confirmed by MRI, CT, or both. The medical records of all included patients had findings of a physical examination performed by a pain specialist that described axial pain without signs or symptoms of nerve or spinal cord compression. Of the included 140 patients, 15 underwent a VA procedure at another institution; all 15 of these patients had fractures in vertebral bodies different from the locations of the fractures of the patients who underwent VA treatment at our institution.

3.2. Demographic Data

Demographic data for the 140 included patients are summarized in

Table 1. Patients’ characteristics (N = 140).

Characteristic	n	%
Age
<60	26	18.6
≥60	114	81.4
Sex
Male	79	56.4
Female	61	43.6
Race
White	109	77.9
Black	13	9.3
Asian	7	5
Other	11	7.9
Cancer diagnosis
Multiple myeloma	58	41.4
Pancreatic	4	2.9
Gastro-esophageal	4	2.9
Gynecologic	4	2.9
Thyroid	1	0.7
Other cancers	20	14.3
Prostate	9	6.4
Lung	12	8.6
Breast	6	4.3
Renal	8	5.7
Colorectal	5	3.6
Melanoma	5	3.6
Hepatobiliary	4	2.9

. The median age was 69 years (range, 38–92). There were 61 women and 79 men. Most patients (77.9%) were White. The most common diagnoses were multiple myeloma (n = 58, 41.4%) and lung cancer (n = 12, 8.6%), with the remaining diagnoses comprising solid and liquid cancers of diverse etiologies.

None of the patients had a documented osteoporosis-related fracture or other factors that might have contributed to the vertebral injury (e.g., trauma). Most patients had a single compression fracture of the lumbar spine (n = 39, 27.9%) or of the thoracic spine (n = 29, 20.7%). The spine levels augmented included those from the thoracic spine (as cephalad as T4) to the sacral spine (as caudad as the sacrum). In most cases, either one or two levels at the lumbar spine were augmented per procedure (n = 52, 36.4%). The most common techniques used were percutaneous vertebroplasty (n = 65, 46.4%) and balloon kyphoplasty (n = 55, 39.29%), and most patients (n = 81, 57.86%) underwent a unipedicular VA approach (

Table 2. Vertebral augmentation levels, technique, and approach used.

Characteristic	n	%
Levels augmented
L1–2	52	36.4
L3–4	10	7.1
S1	1	0.7
T1	29	20.7
T1, L1–2	14	9.9
T1, L3–5	8	5.7
T2	5	3.6
T2, L1–2	9	6.4
T2, L3	1	0.7
T3	3	2.1
T3, L1–2	2	1.4
T3, L3	2	1.4
T4, L1	2	1.4
Technique
Kyphoplasty	55	39.3
Kyphoplasty with EII	2	1.4
Vertebroplasty	65	46.4
Vertebroplasty with EII	1	0.7
EII	17	12.1
Pedicle approach
Unipedicular	81	57.9
Bipedicular	48	34.3
Combined	11	7.9

Abbreviation: EII, expandable intravertebral implants.

). Patients older than 60 years were more likely to undergo unipedicular VA compared to patients younger than 60 years (n = 68 [59.6%] and n = 13 [50%], respectively).

3.3. Main Results

The patients’ baseline demographic and clinical characteristics were generally similar. The mean pain score (as evaluated by the numerical rating scale) was reduced by 3.56 (SD, 2.58) after the VA procedure. The median pain score was reduced by four (range, −3–10) (p < 0.0001). These findings suggest that the VA procedures were effective for pain control.

Overall, the mean FT was 233.80 s (SD, 126.28 s; median, 198.75 s). The mean FD was 157.98 mGy (SD, 101.14 mGy; median, 144.66 mGy). FT and FD were directly proportional to BMI. Whereas a mean FT of 139.95 s and a mean FD of 15.04 mGy were correlated with a mean BMI of 22.7, a mean FT of 303.7 s and a mean FD of 210.79 mGy were correlated with a mean BMI of 30.45. FT was longer (p = 0.048) for patients younger than 60 years (mean, 261.1546 s; SD, 135.4471 s; median, 231.7 s) compared to patients 60 years old or older (mean, 227.5588 s; SD, 123.8808 s; median, 191.25 s). Differences in FD among patients younger than 60 years and those 60 years and older were not statistically significant (

Table 3. Association of FD and FT with patient age, sex, and pain relief.

(A). Correlation of FD and FT with patient age.
	Age	n	Mean	SD	Median	p value
FD, mGy
	<60	26	153.9827	101.7622	132.22	0.149
	≥60	114	158.8884	101.4324	145.925
FT, seconds
	<60	26	261.1546	135.4471	231.7	0.048
	≥60	114	227.5588	123.8808	191.25
(B). Correlation of FD and FT with patient sex.
	Sex	n	Mean	SD	Median	p value
FD, mGy
	Male	79	165.9843	102.3922	160.00	0.049
	Female	61	147.6077	99.3825	119.51
FT, seconds
	Male	79	232.2357	127.4361	221.10	0.150
	Female	61	235.8213	125.7996	193.50
(C). Correlation of FD and FT with patient pain relief.
	Difference in pain scores	n	Mean	SD	Median	p value
FD, mGy
	<5	69	161.3165	96.2517	150.72
	≥5	43	155.8321	92.4816	145.19	0.5193
FT, seconds
	<5	69	243.5565	134.5512	223.60
	≥5	43	218.0837	112.2511	182.90	0.8397

Abbreviations: FD, fluoroscopy dose; FT, fluoroscopy time. Pain scores 0–10 VAS; mGy, milli-gray.

A).

FD was higher (p = 0.049) for the male patients (mean, 165.9843 mGy; SD, 102.3922 mGy; median, 160.00) compared to the female patients (mean, 147.6077 mGy; SD, 99.3825 mGy; median, 119.51). Differences in FT between male and female patients were not statistically significant (Table 3B).

There were no significant correlations between the volume of PMMA injected and FT or FD or between pain relief and FT or FD (Table 3C).

The incidence of complications from our VA procedure was reported to be 10%, mainly from mild PMMA extravasation, and without clinical implications. This did not reach a statistical difference between the groups with and without complications, as the p value for FD was 0.5205, and 0.4637 for FT. The complications listed included PMMA extravasation n = 8 (one vascular, one to the spinal canal), balloon rupture n = 1, and inability to advance the trocar n = 3.

4. Discussion

VA is a minimally invasive technique for the percutaneous stabilization of painful vertebral fractures, including those fractures that are pathological from malignancy. There is a variability in adverse events caused by VA [4,5,6,7,8], due in part to the need for intraoperative fluoroscopic visualization and to differences in the experience level of the proceduralist, as well as to fracture morphology and vertebral level, among other factors.

Due to this complexity and the length of time required for such procedures, examination of the ionizing radiation exposure is imperative, as potential deterministic (harmful effects likely to occur after exceeding dose thresholds) and non-deterministic (stochastic) effects from such exposure are concerning for the patient and for the proceduralist. For instance, latent effects seen from ionizing radiation exposure include the development of cataracts after a fractionated or cumulative dose of 800 rem (0.01 rem = 0.1 mGy), with a latency period of 8 years [9,10], with concerns that lens opacification may actually occur with even lower cumulative doses [11]. The International Commission on Radiological Protection and the National Council on Radiation Protection and Measurements define the dose limits for VA procedures to keep both workers and patients safe from the harmful effects of ionizing radiation [12,13,14].

In comparison to non-malignant conditions, VA procedures performed in cancer-related fractures expose some differences that might be reflected in a higher FT and FD. Here, we discuss some factors that might contribute to our findings.

4.1. Patient Confounders

The literature lacks documentation of radiation exposure of VA procedures for pathologic vertebral fractures. For instance, several studies have published data only on the non-cancer population. In a meta-analysis that did not include patients with cancer, Kopparapu et al. [15] reported a mean FT of 294 ± 198 s for vertebroplasty, which was determined to have lower radiation exposure for both the operator and the patient compared to kyphoplasty. In another study of non-cancer patients, Boszczyk et al. [16] reported a mean FT per level of 124 s (range, 36–258 s) for single fractures and 102 s (range, 36–174 s) for multiple fractures. In our study, we found a longer overall mean FT of 233.8 s and a higher overall mean FD of 157.98 mGy for VA procedures.

Interestingly, like the results displayed by Manchikanti et al. [3], we found a significant association between increased radiation exposure and overweight status, defined as a Body Mass Index (BMI) of 25 or above (mean BMI of 26.67 for our study). FT was found to be longer particularly in patients with BMI > 30.45. This might be idiosyncratic of our population.

Furthermore, commonly distorted vertebral anatomy in cancerous lesions (e.g., posterior wall disruption, osteolysis, destruction of pedicles), requires a tailored approach to each individual VA without sacrificing the efficiency of the procedure [17]. In our study, most patients underwent a unipedicular VA approach (57.86%). Of high importance is the presence of sclerotic bone lesions, which, as opposed to lytic bone lesions, oftentimes require longer FT [18] due to the higher bone density that the instrumentation must pass through. This seems to correlate with the observed increased FT in our patients aged less than 60. In this subgroup, there was a higher number of sclerotic lesions from prostate cancer, breast cancer, carcinoid, medulloblastoma, neuroblastoma, adenocarcinoma of the gastrointestinal tract, Hodgkin lymphoma, small cell lung cancer, pulmonary adenocarcinoma, and medullary thyroid carcinoma.

Finally, the higher bone density of men on average compared to women may explain the association with higher FD [19] as well, which our study also found.

4.2. Equipment Confounders

Simultaneous biplanar fluoroscopy has been associated with decreased radiation exposure by allowing for the use of a single optimal setting throughout the procedure, thus limiting the need to readjust the settings for a second plane of view [20]. Similarly, the use of an O-Arm, a portable imaging device that provides real-time 2D and 3D images, has resulted in higher precision and accuracy, a lower need for surgical revision and additional imaging, and no radiation exposure for the surgical team [21]. In our institution, both technologies are available in neurological and interventional radiology suites but not available in the interventional pain fluoroscopy suite.

FT and FD can be lowered when surgeons operate the C-arms using foot switches to control the radiation exposure [22]. However, these procedural settings are not available in many facilities, including ours. Lastly, kyphoplasty has greater technical demands compared to vertebroplasty due to the use of an inflatable bone tamp, therefore resulting in a higher need for fluoroscopy assistance.

4.3. Proceduralist Confounders

Interestingly, our findings may have been impacted by a learning curve effect of the trainees, even when they were supervised by an experienced proceduralist, as demonstrated by the higher FT and FD values observed in July and August of the academic year (mean FT, 265.80 s; mean FD, 165.90 mGy), when trainees are beginning their fellowship training and therefore are less experienced in performing VA procedures, compared to the months of May and June (mean FT, 225.75 s; mean FD, 145.85 mGy), towards the end of their fellowship training.

4.4. Limitations

This study had some limitations. Inadequacies in the data acquisition could not be addressed due to the retrospective nature of the study. In addition, a discrepancy in the size of the comparison groups could not be reconciled, in part due to our limitation to cancer-related fractures only compared to external data. Lastly, different institutions, different clinicians, and different levels of experience and expertise are major limitations greatly influencing the time and dose of fluoroscopy.

5. Conclusions

Our data suggest that there is greater radiation exposure with VA procedures in cancer patients compared to non-cancer patients, demonstrating that proceduralists must remain vigilant regarding radiation safety and even more so with the former. Numerous factors, such as patient variables, fracture complexity, technique, proceduralist experience, and various imaging equipment availability, may contribute to FT and FD.

Variables associated with decreased FT and FD values.

• Use of biplanar fluoroscopy
• C-arms operated by surgeons using a foot switch.
• Use of pulsed imaging
• Vertebroplasty
• Presence of experienced proceduralists.