Quality and safety in stereotactic radiosurgery and stereotactic body radiation therapy: can more be done?

Abstract

Stereotactic radiosurgery (SRS) has been an effective modality for the treatment of benign and malignant cranial disease for 50 years. Increasingly, the stereotactic approach, ablative doses of radiation delivered in a highly focused manner to a target of interest, is being applied in a number of extracranial disease sites. Stereotactic body radiation therapy (SBRT) holds significant potential for improving tumor control rates across a range of locations and histologies. Both SRS and SBRT require specialized technology, meticulous procedures, and dedicated personnel. Several recent high-profile medical radiation events have generated considerable attention within the media, and serve to remind the profession that close attention to ongoing quality improvement is a fundamental responsibility. The purpose of this manuscript is to provide some recommendations for SRS / SBRT processes and procedures that may be beneficial in understanding and reducing risks inherent to the modalities.

INTRODUCTION

The field of stereotactic radiosurgery was pioneered in 1951 by the Swedish neurosurgeon, Lars Leksell, when he applied the methodology of stereotactic surgery to the delivery of external beam irradiation [1]. The accomplishment was particularly remarkable given that stereotactic surgery was itself in its infancy, having been applied to humans only four years earlier. While radiosurgery was initially indicated for pain and arteriovenous malformations, the approach was subsequently adopted for the treatment of malignant disease. Throughout the 1960’s and 1970’s the field progressed slowly, hampered primarily by the lack of optimal radiation sources and imaging modalities. Leksell and others subsequently abandoned the use of low energy x-rays in favor of protons. Through the mid-1980’s, however, stereotactic radiosurgery remained a highly specialized and relatively obscure procedure, practiced at only a handful of centers worldwide. In those early years, few recognized the true clinical potential the approach held for the treatment of cranial neoplasms, much less for extracranial disease.

The field changed radically with the invention of the Gamma Knife in 1968 [2] and the development of linac-based radiosurgery beginning in the early 1980’s [3-8]. Dedicated, low-cost stereotactic radiosurgery apparatus provided many clinicians with both the capabilities and confidence to effectively treat malignant as well as benign disease. By the mid-1980’s, the success of cranial SRS as an efficient, potent means of local tumor treatment prompted a growing interest in the application of analogous strategies of high-dose-per-fraction treatment in the treatment of extracranial disease. Pioneering efforts in stereotactic body radiation therapy (SBRT) occurred at two institutions, though with radically different approaches. Ingmar Lax and Henric Blomgren, working at the Karolinska Hospital, devised a mechanical device that could comfortably and reproducibly immobilize a patient as well as dampen breathing-related internal organ motion [9]. Their Stereotactic Body Frame (SBF) incorporated a stereotactic fiducial system in which the coordinates of a desired target could be defined from an imaging study. Dose was routinely delivered in one to four fractions (i.e., hypofractionated). In parallel, a methodology for radiosurgery of targets involving and adjacent to the spine was developed at the University of Arizona by Alan Hamilton and Bruce Lulu [10]. Patients were placed within a shallow, rigid box in a prone position, and small clamps were attached under local anesthesia to one or more spinous processes adjacent to the intended target. The clamp system provided a fiducial mechanism for stereotactic localization while simultaneously keeping patients rigidly fixed throughout the procedure. Stereotactic management of extracranial disease has become increasing popular with the emergence of image guidance technologies in the mid-to-late 1990’s, eliminating the requirement for rigid fixation while introducing localization uncertainties comparable to those associated with frame-based cranial applications [11-15]. Inroom 2D and 3D x-ray guidance is now commonplace in radiotherapy, and the extension to SBRT applications has logically followed.

QA REOMMENDATIONS AND DISCUSSION

By most accounts, stereotactic radiosurgery and stereotactic body radiation therapy are safe and effective. SRS has been used for decades in the treatment of brain metastases, with several studies demonstrating a clear benefit in local control [16, 17]; the addition of whole brain radiotherapy also confers a survival benefit [17, 18]. Acute side effects, including headache, screw-site infection, and short-term exacerbation of neurologic symptoms are relatively minor and readily managed [19]. Late side effects, including radiation necrosis, brain edema, and exacerbation of preexisting, or development of new neurologic deficits occur in less than 5% of patients [19]. Five year local control rates following SRS or fractionated stereotactic radiation therapy (FSRT) for acoustic neuromas exceed 95% [20]. Current doses of 13 Gy (single fraction) or ~50 Gy (in 1.8 Gy fractions) yield excellent rates of hearing preservation as well as very low rates of facial and trigeminal neuropathies [20]. Similarly, excellent rates of local control can be expected following either SRS or FSRT treatment of meningiomas, though the grade and location of these tumors plays a significant factor in both tumor control and potential complications [21-1238].

SBRT is a much more recent discipline, with unique technological and clinical considerations. Nevertheless, initial clinical results from prospective single institution, and more recently, multi-institutional clinical trials of SBRT have documented similar high rates of tumor control coupled with a low incidence of serious toxicity despite the high dose fractions of radiation being delivered. In the setting of early stage lung cancer, Baumann et al and Timmerman et al reported three year local control rates of 92% and 97.6%, respectively [24, 25] Incidence of grade 3 / grade 4 toxicity were 28.0% / 1.8% and 12.7% / 3.6% respectively. Similar results have been reported for SBRT in the setting of metastatic lung [26] and liver [27] tumors.

The treatment of tumors of the vertebral column using stereotactic techniques, often referred to as spinal radiosurgery, presents unique technical and clinical challenges, largely due to the close proximity of the spinal cord. The radiation sensitivity of the spinal cord is not well documented, and the implications of treatment-related spinal neuropathy are quite profound. In spite of these challenges, spinal radiosurgery is an active and growing discipline. Maximum doses of 10-14 Gy, given in a single fraction are generally considered safe [28-31], though data remain sparse.

SRS / SBRT is fundamentally different from conventional radiotherapy in that the intent is to deliver an ablative dose that overcomes all of the abilities of a cancer cell to defend itself. It is inherently aggressive, much like a surgical approach, and therefore an increased rate of acute complications compared with conventional radiotherapy could be expected. Despite this, reports of complications are relatively few. In the original clinical SBRT paper, Blomgren et al concluded that three of nine patients with primary intrahepatic tumors likely suffered treatment-related death following SBRT. The lack of a systematic approach to total dose and dose per fraction was likely a key contributor in these complications. More recently, Timmerman et al reported grade 5 toxicity in 6 of 70 patients with early stage lung cancers treated with SBRT [32]. Hilar and pericentral location were strongly predictive of severe toxicity. Based on these observations, a lower SBRT dose is now recommended when treating centrally-located lung tumors. Similarly, SBRT treatment-related spinal cord toxicity is exceedingly rare; a review of the handful of reported cases of spinal cord myelopathy has been summarized by Kirkpatrick et al [31]. Hoppe et al observed acute skin toxicity in 26 of 50 NSCLC patients treated with SBRT [33]. Toxicity was correlated with physical factors, namely, the use of too few beams, and a bolus effect from the treatment couch and immobilization devices. In contrast to the early SBRT applications, much of the experience gained over the last decade has been within the scope of single or multiple institution prospective phase I / phase II clinical trials. The data gleaned from these studies has been extremely helpful to elicit important clinical and technical factors, and as a result, reports of serious complications are rare.

Modern localization techniques incorporating image guidance are capable of localization accuracy on the order of 1 mm for cranial sites and 2 mm for extracranial sites, independent of additional uncertainties caused by respiratory motion in some disease sites [13-15, 34-40]. Similarly, modern dose algorithms are now sufficiently robust to calculate dose with high accuracy and precision in all anatomical sites [40-42]. There are an abundance of guidance documents available to assist practitioners in initiating and maintaining high quality stereotactic programs [40, 43-51].

The growing clinical experience and technical advances, therefore, support the assertion that SRS and SBRT are effective and can be performed with a modest risk of complications. Still, many in the community are aware of recent events, publicized in the media, in which serious errors have occurred. These include: a calibration error on a radiosurgery linac that affected 77 patients in Florida in 2004-2005; identical errors in measurement of output factors affecting 145 patients in Toulouse, France between April 2006 and April 2007 [52-54], and 152 patients in Springfield, Missouri between late 2004 and late 2009; a single incident in France in which the backup jaws for an AVM treatment with a small circular collimator were set to 40 cm x 40 cm [53]; and an error in a cranial localization accessory that affected seven centers in France, Spain and the U.S. While no side effects related to the Florida calibration error have been reported, that is not the case with several of the other events. Gourmelon et al reported at 31% 12 month actuarial rate of trigeminal neuropathy in 32 acoustic neuroma patients overdosed in the Toulouse accident [54]. In contrast, despite a mean overdose of 61.2%, no treatment-related morbidity was observed in the 33 patients treated for brain metastases [52]. Clearly, the consequences of dosimetric errors are dependent on location and histology, as well as time following irradiation. In the case of the patient treated with the incorrect backup collimator setting, a subsequent dosimetric evaluation indicated that a large portion of normal brain received doses in excess of those intended for the AVM. The patient developed an oseotracheal fistula requiring surgery, experienced a hemorrhage and subsequently died.

Additionally, a review of the Nuclear Regulatory Commission (NRC) Radiation Event Report Notification database yielded 13 radiosurgery-related events from 2005 to present, 12 of which resulted in a deviation from the original prescription. These are summarized in Table 1. While patient outcome is not described on the NRC site, several of the events listed, including treatment of the wrong location with single fraction doses as high as 90 Gy, would likely be accompanied by significant morbidity. Wrong-site errors continue to plague all medical disciplines, and are not unique to radiotherapy [55]. Clearly, adherence to rigid quality assurance guidelines could minimize these errors, which can have catastrophic consequences.

Table 1.

List of radiosurgery events reported to the NRC during the period 2005-2010

Event Description	Treatment Implication
Patient orientation entered incorrectly at MR Scanner	Wrong location treated
Fiducial box not seated properly during CT imaging	Wrong location treated
Malfunction of automatic positioning mechanism following re-initialization	Wrong location treated
Right trigeminal nerve targeted instead of left	Wrong location treated
Facial nerve targeted instead of trigeminal nerve	Wrong location treated
Mistake in setting isocenter coordinates	Wrong location treated
Head not secured to stereotactic device (2 events)	Wrong location treated
Selected collimators did not match planned	Wrong dose/distribution delivered
Physician mistakenly typed 28 Gy instead of 18 Gy into planning system	Wrong dose delivered
Physicist calculated prescription to 50% isodose instead of 40%	Wrong dose delivered
Microphone dislodged, causing stereotactic device to break	Treatment halted after 2 of 5 fractions
Couch moved during treatment	None; personnel interrupted treatment

The NRC estimates that 60% or more of radiotherapy misadministrations are due to human error [56]. Indeed, further inspection of each event described above suggests that every one was partially, if not completely, the result of human error; certainly every event could have been prevented by added diligence on the part of the staff. Prevention of errors such as those described above has become a subject of much discussion over the past few years. Several guidance documents aimed at understanding and radiotherapy risks mitigating radiotherapy errors have been forthcoming recently from national and international organizations; these include: the World Health Organization (WHO), the International Commission on Radiological Protection (ICRP), the National Health Service (NHS) of the United Kingdom and the Alberta Heritage Foundation for Medical Research [57-60]. A list of some of the common factors contributing to radiotherapy incidents has been summarized from these documents:

• •Lack of training, competence or experience
• •Inadequate staffing and/or skills levels
• •Fatigue and stress, staffing and skills levels
• •Poor design and documentation of procedures
• •Complexity and sophistication of new technologies
• •Over-reliance on automated procedures
• •Poor communication and lack of team work
• •Inadequate infrastructure and work environment
• •Changes in processes

The WHO has suggested a number of general preventative measures aimed at reducing radiotherapy errors:

• •A thorough quality assurance program to reduce the risks of systematic equipment and procedural-related errors;
• •A peer review audit program to improve decision making throughout the treatment process;
• •Extensive use of procedural checklists;
• •Independent verification through all stages of the process;
• •Specific competency certification for all personnel;
• •Routine use of in-vivo dosimetry.

The WHO acknowledges that while radiotherapy-related errors are not uncommon, the radiotherapy-related error rate compares favorably with the rate of other medical errors [57]. Additionally, the NHS analysis indicated that in 80% of errors identified, the patient was not expected to suffer any adverse clinical effects [59]. Nevertheless, the high-dose nature of SRS/SBRT suggests a potential for increased rate and severity of adverse clinical events. In light of the spate of recent errors described above, we have expanded the general preventative measures provided above into a more specific set of recommendations aimed at assisting institutions and practitioners contemplating or currently performing clinical SRS and/or SBRT procedures. These are not intended to be exhaustive but to supplement existing guidance that may be available from professional organizations or published in the existing literature. For succinctness these are categorized into program planning, commissioning, and patient-specific aspects:

Program Planning – considerations when starting an SRS / SBRT program and/or expanding an existing program to new disease sites.

• •Establish clinical program goals, specify disease sites, identify program specialists, and develop guidelines for treatment, follow-up and assessment following nationally accepted clinical standards [40].
• •Determine the resources required, personnel, technology, time, needed to carry out the program goals; AAPM Task Group 101 recommends such a feasibility study [31]. There must be adequate resources in place to meet the demands of the SRS/SBRT program, and staff must be provided sufficient time to carry out the necessary tasks. ASTRO and ACR have provided guidance on minimum staffing levels [43, 50]. AAPM Task Group 101 suggests additional personnel are necessary to initiate and support SRS/SBRT programs [40].
• •Develop processes for initial and ongoing training of all program staff. Program personnel must demonstrate knowledge and competence in their respective discipline through graduation from an approved educational program, board certification and licensure as appropriate. Program personnel should also maintain their skills by lifelong learning through continuing professional development. Program personnel must also receive site-specific SRS/SBRT training prior to involvement in an SRS/SBRT program [40, 43, 50].
• •Identify specific equipment and processes for simulation, immobilization, image guidance, management of organ motion, and treatment delivery [40].
• •Develop quality assurance processes that encompass all clinical and technical program aspects.
• •Establish clinical SBRT patient conferences for pre-treatment planning and post-treatment review.
• •Develop checklists for all SRS/SBRT processes. Use of checklist has the potential to reduce the risk of errors [56, 61].
• •Integrate SRS/SBRT processes for documentation and reporting, peer review, regular review of processes and procedures, updating clinical guidelines and recommendations, ongoing needs assessment, and continuous quality improvement within the formal quality management program.
• •Require practices to be ACR/ASTRO accredited [62].

Commissioning tests that characterize every performance aspect of the SRS/SBRT system/process prior to initiating a clinical program [38].

Commissioning tests must be performed in a manner that assesses both the individual and integrated (i.e., end-to-end) components that comprise the SRS/SBRT process [40, 43, 50].

• •Obtain an independent verification of the absolute calibration.
• •Obtain an independent assessment of measured beam data, particularly important given the challenges in small field measurements.
• •Perform a comprehensive commissioning of the treatment planning and delivery systems, incorporating the full range of delivery scenarios, parameters and techniques. Commissioning should include all aspects of simulation, including the use of localization / immobilization devices and 4DCT, as well as all aspects of motion management, as appropriate. Commissioning should be facilitated through the record and verify system and performed in clinical mode on the delivery device. This should be done in an end-to-end manner in addition to verifying operation of individual components.
• •Perform an evaluation of individual and end-to-end accuracy capabilities of the localization / image guidance system(s) for each and every disease site as appropriate.
• •Obtain verification of system commissioning from an independent organization such as the Radiological Physics Center. This should be repeated utilizing specialized phantoms for each and every disease site as appropriate.

Patient-Specific Quality Assurance – processes that ensure a commissioned radiosurgery system is capable of delivering an individual patient’s treatment as intended.

• •Each patient’s proposed course of treatment should be discussed and reviewed by the entire stereotactic team, preferably as part of a regular stereotactic conference.
• •The course of treatment, including anatomy definition, dose schedule, normal tissue constraints, CTV/ITV and PTV margins, should be designed based on established guidelines.
• •Standard institutional procedures should be available and followed for every stage of the treatment process. Checklists should be used at every stage of the treatment process, including but not limited to: simulation, planning, daily patient setup and localization, daily delivery, and quality assurance.
• •Appropriate program team members, including radiation oncologist(s), medical physicist(s) and radiation therapist(s) must be present as described by their responsibilities during treatment.
• •A radiation oncologist should confirm anatomy on reference images and review and approve the results of image guidance procedures prior to initiating each treatment.
• •An independent review of setup and treatment parameters should be performed by the physicist(s) and therapist(s) prior to initiating each treatment.

Patient-specific dosimetry should be performed prior to a patient starting treatment. This is most effectively done by mapping the approved plan onto a phantom, delivering the plan using the patient-specific treatment parameters through the record-and-verify system, and directly verifying both point dose(s) and planar distributions.

CONCLUSIONS

SRS and SBRT are critical components of the cancer treatment process, with demonstrated efficacy, as well as significant potential for future gains, in many disease sites. Both are specialized, technology-driven clinical procedures that require a high level of diligence for safe and effective application. There are a number of references focused on guiding the practice of SRS and SBRT. Similarly, there is now a significant body of work aimed at understanding risks in radiotherapy. In providing recommendations for SRS / SBRT processes and procedures, we have attempted to bring these elements together to assist the profession in the task of ongoing quality improvement.