GammaTile^®: Surgically targeted radiation therapy for glioblastomas

Across adult and pediatric populations, malignant neoplasms are associated with substantial morbidity and mortality [1]. Although progress has been made in the treatment of select cancers, glioblastoma remains an exception [2]. Glioblastoma is the most common malignant primary brain tumor in adults, with an incidence rate of 3.2/100,000 [3]. The mainstay of therapy is a multimodal strategy of neurosurgical resection followed by administration of concurrent chemoradiation therapy 4–8 weeks after initial surgery [2,4,5]. More recently, tumor treating fields has emerged as a US FDA-approved treatment option in combination with chemoradiation for newly diagnosed glioblastomas or in the recurrent setting. The expected overall survival has remained 14–21 months over the past several decades [5,6].

Notably, most glioblastoma progression or recurrence occurs locally, in regions immediately adjacent to the resection cavity [7–9]. While glioblastomas often appear as a discrete entity on MRI, surgical and autopsy studies reveal microscopic tumor cells extending at least 2 cm away from the visible tumor [7,9,10]. The density of these microscopic cells is greatest near the resection cavity and 50–70% of glioblastoma patients suffer tumor growth adjacent to the resection cavity during the 4–8-week recovery period [11–13], which prognosticates poor survival [14]. Not unexpectedly, the majority of available studies suggest a delay in initiation of radiation therapy beyond this recovery period is associated with poorer survival outcomes [14,15]. In patients who initiate concurrent chemo- and radiation- therapy within the 4–8-week recovery period, >80% of recurrences occur adjacent to the resection cavity [16,17]. These results suggest that therapeutic platforms that augment local control have the potential to improve clinical outcomes. Recognizing the importance of local control in glioblastoma treatment, NRG-BN001 is a multi-institutional clinical trial that aims to determine whether radiation dose escalation targeting regions surrounding the resection cavity improves survival when combined with temozolomide treatment.

Brachytherapy emerged as an attractive option in this context. Brachytherapy refers to the implantation of interstitial or intracavitary radioactive sources adjacent to the target tissue [18,19]. The notion of brachytherapy was proposed only a few years after the initial discovery of natural radioactivity over 100 years ago [20] and predates other therapeutic radiation techniques. Brachytherapy continues to be a major therapeutic platform for prostate, breast, gynecologic, ocular and other non-CNS neoplasms [21–23]. A key lesson learned during this century of brachytherapy development is an appreciation for the critical importance of the physical properties of the radiation source. Consequently, radioactivity emitting sources utilized in the past have mainly been replaced by safer and more efficacious isotopes, in other words, capable of delivering a more targeted dose and exhibiting shorter half-lives [24–26].

Pertaining to brachytherapy for CNS tumors, the first application of brachytherapy dates back to 1936, when radon was used as treatment for an intrasellar tumor [27]. Later, other sources such as iridium-192 (¹⁹²Ir), gold-198 (¹⁹⁸Au) and iodine-125 (¹²⁵I) were tested, with ¹²⁵I becoming the most commonly utilized isotope in recent years [28,29]. However, the aggregate of available studies suggests that ¹²⁵I is associated with increased risk for radiation necrosis, which has been hypothesized to be related to a combination of the long half-life time of ¹²⁵I and cavity dynamics [9,30–34].

In recent years, cesium-131 (¹³¹Cs) has emerged as a promising isotope for brachytherapy for CNS tumors. While ¹³¹Cs shares many characteristics with ¹²⁵I, the half-life of ¹³¹Cs is significantly shorter than that of ¹²⁵I (9.7 vs 59.4 days, respectively). This shortened half-life translates into improved ease of use [34], improved efficacy [35], as well as a superior safety profile [24,36].

There is long-standing interest in brachytherapy as a means of improving local control for patients afflicted with glioblastoma [37]. Despite promising institutional experiences [38], interest in brachytherapy waned after failure of two randomized controlled trials to demonstrate improved survival after ¹²⁵I brachytherapy [39,40]. While these studies failed to meet primary survival end points, there were nevertheless signals of efficacy. For instance, in the Laperriere study [39], a “tail” of longer-term survivors was noted in the ¹²⁵I brachytherapy treated arm, a finding reminiscent of the landmark temozolomide trial and recent immunotherapy trials [6,41]; this tail was not observed in the comparative arm. Moreover, improved survival was noted in the Brain Tumor Cooperative Group, with an approximate 1-month survival difference that did not reach statistical significance (p = 0.101) [40]. It should be noted that FDA clearances of carmustine wafers and temozolomide were based on survival differences of similar magnitude [42]. Moreover, in a separate study, Laperriere et al. examined clinical specimens obtained after ¹²⁵I brachytherapy and showed decreased cellularity and increased necrosis in these samples relative to samples from the comparative arm [43], suggesting brachytherapy contributed to improved local control.

In addition to these observations, there have been two developments that fueled a resurgence of interest in brachytherapy. First, the efficacy of concurrent temozolomide therapy raises the possibility that the addition of brachytherapy would further amplify efficacy, a possibility supported by two recently published case series [13,44]. Second, the advent of GammaTile^® (GT), a device with ¹³¹Cs radiation emitting seeds embedded in a resorbable collagen-based carrier tile has significantly lowered the technical barrier to radiation planning and surgical application.

Emerging data continues to demonstrate a favorable efficacy and safety profile of GT, described generically as surgically targeted radiation therapy (STaRT), against a spectrum of CNS tumors, including recurrent meningiomas [45], high grade gliomas [13] and brain metastases (Table 1) [18,46–48]. Here, we review the technical specifications of STaRT, describe our preliminary clinical experiences and discuss opportunities and limitations pertaining to clinical translation of STaRT as a glioblastoma therapy.

GammaTile^®

GT is an FDA-cleared brachytherapy platform where titanium encapsulated ¹³¹Cs seeds (Model CS-1, Rev. 2, IsoRay Medical Inc., WA, USA) are embedded in a resorbable collagen-based matrix (Saturable DuraGen Matrix, Integra Lifesciences, NJ, USA) tile (described as STaRT), providing a more modular system than previously available (Figure 1). GT is approved as treatment for both newly diagnosed malignant brain tumors and recurrent brain tumors irrespective of histology. A single tile measures 2 cm × 2 cm and contains four radioactive ¹³¹Cs seeds with a half-life time of 9.7 days and mean photon energy of 30.4 KeV [50,51]. The modular nature, the pliability and tissue adherent collagen matrix maximizes likelihood of conformal radiation delivery, facilitates dosimetric planning and expedites surgical implantation. The tissue offset of 3 mm provided by the tile dimensions reduces the likelihood of focal necrosis around the sources. The source delivers a low dose rate, which, when combined with the short half-life, affords a favorable safety profile. GT delivers 120–150 Gy at the cavity surface and maintains 60–80 Gy at 5 mm depth, exceeding the standard dose of external beam radiation therapy (EBRT) by 1.5-twofold [45,46].

Advantages of GT

The novelty of GT is in its design, which has the following advantages.

• Embedding ¹³¹Cs (half-life of 9.7 days) within a bioresorbable collagen matrix tile (described as carrier tile brachytherapy or STaRT) eliminates the need for subsequent surgical removal, which contributes to improved quality of life for the patients [49]. The use of a resorbable matrix ensures minimal intracranial residue after implant;

• ¹³¹Cs, which is considered low-dose rate brachytherapy, has a better adverse effect profile than other commonly used isotopes, such as iodine-125 (¹²⁵I) [52].

• The carrier material functions as a spacer and implanted compensator and avoids direct contact of the ¹³¹Cs seeds with the brain parenchyma while the seeds are active, potentially preventing harmful interactions and reducing the risk of necrosis. The 3 mm of tissue offset provides a clinically useful dose attenuation as a direct result of the inverse square law [53];

• From a neurosurgical perspective, implant of STaRT is akin to the use of other resorbable collagen matrices routinely implanted during surgery [54]. As such, GT implant does not require modification of routine surgical maneuvers or lengthen surgical duration;

• The design and the characteristic adherence of the collagen matrix which makes up the carrier tile, minimizes seed migration after implantation to maintain inter-source spacing after closure. As such, the design affords more uniform coverage of the target area [55];

• The collagen carrier tile remains intact for the duration of approximately 6 weeks or more (equal to or greater than 4 half-lives of ¹³¹Cs) and holds the resection cavity in a fixed configuration. Resection cavity contraction could lead to overlapping of traditionally used individual or Tyvek suture enclosed radioactive seeds. Such overlap increases the risk for brachytherapy morbidity, including radiation necrosis. The reduced risk for resection cavity contraction minimizes the risk for such consequent dose inhomogeneity [34];

• STaRT circumvents any delay between surgical resection and radiation therapy. Typically, a delay of 4–8 weeks between neurosurgical resection and the initiation of EBRT is anticipated in order to allow for surgical wound healing. There are select centers that initiate therapy prior to 4 weeks postoperatively. However, some delay in chemoradiation after surgery is universal. Implantation of STaRT bypasses this delay to initiate adjuvant RT at the time of surgery.

• EBRT requires patient immobilization through face mask application. Some patients experience claustrophobia in this context; the use of STaRT eliminates this risk while ensuring 100% patient compliance with their radiation treatments;

• STaRT decreases the burden on the patient and to the healthcare system. Daily visits to the hospital are required for EBRT, which compromises the quality of life for the patient and increases the burden on the healthcare system. STaRT eliminates the need for daily visits. In the face of a pandemic, such as the one currently challenging healthcare delivery, access to cancer surgery is still vital [56]; GT placement eliminates the need for up to 30 or more visits to healthcare facilities for radiation therapy (RT), reducing this exposure risk to a vulnerable population;

• Cost-effective modeling suggests that GT may be more cost effective than EBRT for the healthcare system [57];

• There is no requirement for expensive equipment or lengthy training. Precise targeting is accomplished through visualization of the surgical bed and placing the tiles typically takes a few minutes, so there is no steep learning curve.;

• In EBRT, the radiation beam must travel through healthy tissue, which carries the potential to harm non-tumorous tissue. STaRT is localized to limit radiation delivery to the tumor affected parenchyma [53]. This localized delivery reduces possible side effects and neurocognitive decline associated with EBRT [18]. Also, STaRT minimizes the likelihood of treatment related hair loss, which occurs after EBRT.

Clinical experience

Candidates for STaRT meet with both the treating radiation oncologist and neurosurgeon. Radiation safety precautions are reviewed by the medical physicist with the patient prior to the surgery. The number of tiles to be implanted is determined based on the anticipated postoperative surface contours of the tumor bed. The number of tiles is then custom-ordered and is available within a week of request. The tiles are received by the radiation oncology department and handled in accordance to institutional policy. On the day of surgery, the tiles are brought to the operating room by the medical physicist, who performs radiation safety checks throughout the implantation process. In our institution, maximum safe resection of tumor is verified through an intra-operative MRI before STaRT. It should be noted that intra-operative MRI is not required for STaRT. After maximal resection, the collaborating radiation oncologist scrubs in to sterilely unpack and prepare the STaRT implant, which is then handed to the surgeon. Tiles are placed into the resection cavity under microscope or loupe magnification. Most implantations take two minutes. Even in large resection cavities requiring a dozen or more tiles the implantation has been completed within five minutes. A thin cut head computerized tomography (CT) and MRI are performed in the postoperative setting to provide the basis for dosimetric calculations (Figure 2). At the completion of the cranial repair, no special precaution is required of nursing staff during the patient’s hospitalization since native cranium is able to block the emission of STaRT implant to exposure levels compatible with outpatient discharge (typically considered to be less than 6mRm/h at 1 m [58]). These limited levels of exposure compare favorably to the NCRP dose limit recommendations assuring safe levels of exposure to caregivers and medical personnel. In our practice, STaRT has been well tolerated and patients are discharged home on postoperative day 1 or 2.

Our most common use of STaRT has been in the recurrent glioblastoma setting. Many of these patients had several previous craniotomies and underwent multiple rounds of therapies, including radiation and bevacizumab. A significant portion of our patients was treated with high-dose corticosteroid immediately prior to the resection. Despite the anticipated heightened risk for wound related complications in this population, surgical resections with STaRT implantations have been associated with an excellent safety profile. A dedicated report to describe our experience will be presented elsewhere.

Limitations in clinical translation as a glioblastoma therapeutic

While the rapid dose falloff of ¹³¹Cs-based STaRT increases the safety profile in terms of wound healing and radiation induced neurological morbidities, there are limitations in terms of treatment for the microscopic tumor cells that extend beyond 2 cm of the resection cavity where STaRT is implanted. Ultimately, while STaRT affords an increased likelihood of local control, meaningful improvements in clinical outcomes requires synergy with other forms of adjuvant therapy [59,60].

To maximize the probability of efficacy, gross total resection (or near gross total resection) will be required, given tumor cells more than 5–8 mm distant to the resection cavity where STaRT is implanted are unlikely to benefit from this therapy. The availability of intra-operative MRI and 5-aminolevulinic acid may facilitate maximal surgical resection and facilitate STaRT efficacy in this context. Initiation or resumption of postoperative adjuvant systemic treatment can be undertaken as soon as medically cleared.

The available literature suggests that a subset of glioblastoma patients suffered from microscopic tumor infiltration into the brainstem. These patients exhibit poor survival [61] and are unlikely to meaningfully benefit from treatments aimed to boost local control.

When applied in the newly diagnosed setting, the safety of combining STaRT with the standard-of-care temozolomide/EBRT remains an unresolved matter (Tables 2 & 3) [62,63]. Thoughtful safety studies involving dose titration of STaRT are warranted in this regard. One possible approach would be to deliver the typical EBRT ‘boost’ treatment via STaRT immediately at the time of resection such that 90% of the boost dose would be delivered to the area at greatest risk of harboring residual disease, prior to initiation of wider field EBRT treatments. This approach offers the advantages of immediate treatment, a more condensed treatment time frame and fewer clinical visits. Moreover, it allows potential for dose escalation in the treatment of radio-resistant glioblastomas.

Future perspective

In many non-CNS tumors, brachytherapy has improved disease control and improved clinical outcomes. Thoughtful application of brachytherapy to CNS tumors is likely to produce similar results. The new GT platform which utilizes modular collagen-based carrier tiles to maintain spacing and provide a tissue offset of ¹³¹Cs brachytherapy seeds, holds great potential in this regard. The modulatory properties lower technical barriers for clinical application and allow for more accurate coverage. The ¹³¹Cs source and the offset affords a favorable safety profile and rapid dose tapering, minimizing risk for adverse events such as radiation necrosis and wound compromise. At the same time, cancers with tumor cells beyond the range of STaRT delivery will require integration of supplementary therapy. Safety of STaRT in the context of standard EBRT warrants consideration as an upfront treatment.