A 59 year old otherwise healthy gentleman presented in 2007 with a prostate-specific antigen (PSA) level of 5.4 ng/mL (elevated versus prior levels). Prostate biopsy dated 7/27/2007 revealed Gleason score 3 + 4 = 7 adenocarcinoma from 10 % of the left lobe and Gleason score 3 + 3 = 6 adenocarcinoma from 10 % of the right lobe (“overall” Gleason score 3 + 4 = 7). His prostate was palpably normal (Clinical stage T1c). His I-PSS score was 11/35 (c/w mild preexisting voiding symptoms) and his presenting Sexual Health Inventory Matrix (SHIM) score was 25/25, indicating full potency. He was very concerned about loss of potency with treatment. After careful consideration of a number of potentially curative surgical and radiotherapeutic treatment methods, he selected CyberKnife SBRT monotherapy as his treatment of choice. Due to relatively limited efficacy data in 2007, he agreed to receive his prostate SBRT treatment as a participant in an IRB-approved clinical trial [32].

On September 7, 2007, he received 4 transperineally implanted intraprostatic gold fiducials to serve as his SBRT image guidance reference structure (Fig. 11.1). His prostate measured 59.6 cc on ultrasound imaging, with no other specific abnormally visualized. Fiducial-to-fiducial coregistered CT and MRI-based treatment planning was accomplished 1 week later, with a Foley catheter in place for both studies to identify the location of the urethra. The prostate and seminal vesicle bases were defined primarily from the MRI image set and contoured contiguously, with all contours also reconciled with the reference CT image set, to form the Gross Target Volume (GTV).

An asymmetric margin expansion was then applied to form a Clinical Target Volume (CTV) that covered the entire prostate and potential extracapsular extension contiguously. Specifically, this consisted of GTV + 2 mm on the right, GTV + 5 mm on the left (the side that harbored Gleason 7 disease and MRI-abnormality; felt to be at higher risk of extracapsular extension), contiguously also surrounding the seminal vesicle bases by 2 mm, and shaved to zero mm against the rectum. There was no further CTV to PTV expansion (CTV = PTV) (Fig. 11.2).

Although it could be argued that at least a mm or two of extra CTV to PTV margin expansion could also be applied to account for issues such as image interpretation/registration uncertainty and potential interfraction/intrafraction prostate distortion, the 2 Gy/fraction radiobiological dose equivalent of the Virtual HDR protocol is 9,400 cGy for an α/β [alpha/beta] ratio of 3 Gy (a commonly assumed rectal α/β [alpha/beta] ratio value). As such, any added margin expansion is potentially injurious to adjacent tissue, and we thus choose to err on the side of “tight” margins on this protocol. The published sub-millimeter tracking accuracy of the CyberKnife delivery device and its ability to correct beam aiming in the translational and rotational dimensions (“6D”) further support the reasonability of “tight” margins when using this technology [14, 15].

He subsequently received protocol SBRT treatment on consecutive days, from 10/15/2007—10/18/2007, using the CyberKnife device as the SBRT delivery mechanism. The prescribed dose was 3,800 cGy/4 fractions, prescribed to the 49 % isodose line, creating an intraprostatic maximum dose (“Dmax”) of 7,755 cGy within the left lobe (“HDR-like” intraprostatic dose escalation) to a PTV that measured 148.2 cc,. Adjacent tissues (bladder, urethra, rectal wall and rectal mucosa) were also subject to “HDR-like” dose limitations per protocol (120, 120, 100 and 75 % of prescribed dose, respectively). Minor protocol deviations were accepted on the bladder, rectal wall and rectal mucosa Dmax levels to assure full PTV coverage, while the urethra dose level was fully protocol compliant.

Acutely, he had grade I GU and GI toxicity, including an I-PSS symptom score increase of 6 points and some rectal discomfort at the 2 week follow-up visit, fully resolved to baseline by the 1 month follow-up visit. There was no chronic GU or GI toxicity. His sexual potency was fully maintained, with a SHIM score of 25/25 at all short-term and long-term follow-up visits, out to 5 years.

The PSA level steadily decreased from 5.4 ng/mL pre-treatment to 1.0 ng/mL by 6 months post-treatment, increasing to 2.3 ng/mL at the 1 year follow-up (“PSA bounce”), thereafter decreased to 0.4 ng/mL by the 2-year follow-up visit, followed by a second minor “PSA bounce” to 0.5 ng/mL at the 30 month follow-up, steadily decreasing after that to a final PSA nadir level of 0.1 ng/mL by the 4 year and 5-year follow-up visits. The PSA response curve for this case is illustrated in Fig. 11.3.

The SBRT planning MRI had shown a 61 cc prostate with diminished left lobe T2-weighted signal intensity and gadolinium hyperperfusion, c/w stage T2b disease. Follow-up MRI at 2 years revealed resolution of abnormalities and a prostate volume shrinkage to 35 cc (reduction of 43 %). He remains free of complications, with fully maintained potency function and improved voiding function (baseline I-PSS score of 11/35 pre-SBRT decreased to 2/35 at the 5 year follow-up visit). He will continue to have annual PSA-based follow-up visits.

The predominance of published literature to date describes SBRT monotherapy against low- and intermediate-risk prostate cancer cases [33–35]. The initial cohort was recently updated by King et al. [33] In that study 67 low and low-intermediate-risk patients were treated with SBRT to a dose of 36.25 Gy/5 fractions, to a planning target volume (PTV) created by an anisotropic margin expansion of the prostate, by 3 mm posteriorly and 5 mm elsewhere, prescribed to cover >95 % of the PTV with the prescription dose, using the CyberKnife as the delivery device.

The 4-year actuarial biochemical disease-free survival was 94 % (median follow-up 2.7 years), with 2 biopsy proven local failures, and a reassuringly low rate of grade 3 or higher toxicity (3 % GU, 0 % GI). The few observed cases of grade 3 GU toxicity were preceded by urologic instrumentation. Grade 1–2 GU and GI toxicity was seen in 28 and 12.5 % of patients, respectively, with a statistically significantly lower incidence of low grade toxicity in patients that were treated every other day (q.o.d.) as opposed to daily (q.d.) (p = 0.001 for gastrointestinal and p = 0.007 for genitourinary). It was concluded that SBRT efficacy compared well with other treatments and that the incidence of severe toxicity was low, causing the authors to suggest that “current evidence supports consideration of stereotactic body radiotherapy among the therapeutic options for localized prostate cancer.”

A pooled dual institutional study by Freeman and King, restricted to a smaller cohort of 41 CyberKnife SBRT patients treated to 35–36.25 Gy/5 fractions with a significantly longer 5-year median follow-up was also reported. They reported similar findings—93 % 5-year biochemical disease-free survival and only a single case of grade 3 GU toxicity, with no Grade 3 or higher GI toxicity [34].

A different study reported by Katz et al., evaluated the efficacy of two different CyberKnife SBRT dose-fractionation regimens (35 Gy/5 fractions versus 36.25 Gy/ 5 fractions) in two groups of patients with predominantly low-risk prostate cases—41 consecutive patients treated to 35 Gy/5 fractions versus 41 matched subsequent cases treated to 36.25 Gy/5 fractions, with 54 and 48 months median follow-up in each group, respectively [35]. The crude biochemical disease-free survival rate measured 97.5 % with each dose regimen, with no significant difference in PSA nadir (0.1 ng/mL at 4 years) or toxicity rates. The incidence of late grade 2 urinary toxicity was 5 and 10 % in the 35 Gy versus 36.25 Gy, groups, respectively (p = NS), with an identical 5 % rate of grade 2 GI toxicity in each group. The authors concluded that “overall, the highly favorable PSA response, limited biochemical failures, limited toxicity, and limited impact on quality of life in these low- to low-intermediate-risk patients are supportive of excellent long-term results for CyberKnife delivered SBRT” and their work suggested that there is no efficacy loss with the lower dose level. It should be pointed out though, that “local relapse” may be a very late event, meaning their conclusion of equivalence is at best “hypothesis generating,” and requires longer-term confirmation before the efficacy of the lower dose arm may be considered truly identical to the more commonly used dose level of 36.25 Gy/5 fractions.

Based on ample successful brachytherapy precedent, a different and more aggressive prostate CyberKnife SBRT method and fractionation scheme, known as “Virtual HDR” SBRT, was described in an IRB-approved protocol in 2006, with comparative dosimetry analysis versus actual HDR brachytherapy published in 2008 along with very preliminary clinical results [30, 32]. Using this approach, prostate CyberKnife SBRT treatment plans are deliberately designed to mimic HDR brachytherapy as closely as possible; using an actual HDR brachytherapy schedule of 3,800 cGy/4 fractions, previously described by Grills et al. [22].

Beyond simply mimicking the dose-fractionation aspect of HDR brachytherapy, this method also applies substantially identical intraprostatic and periprostatic isodose morphology, including “HDR-like” urethra, bladder and rectal maximum “D_max” dose limitation constraints. Additionally, to qualify as “Virtual HDR” a prostate SBRT treatment plan is required to produce significant intraprostatic dose heterogeneity, just as is seen with actual HDR brachytherapy, with intraprostatic D_max dose escalation greater than 150 % of the prescribed dose, requiring prescription to an isodose line less than 67 %, and with the majority of such treatment plans prescribed to an isodose line of 50–55 %. This translates to a typical intraprostatic maximum dose level approaching twice the prescription dose level. This dose escalation region is designed to superimpose primarily over the peripheral zone of the prostate—the region that most frequently harbors the largest volume of intraprostatic cancer cell burden [33]. As imaging techniques such as multi-parametric MRI or evolving PET/CT methods more accurately define the “dominant intraprostatic lesion” (“DIL”), this technique may also be modified to restrict the intraprostatic dose escalation to the actual DIL, though the preponderance of “Virtual HDR” experience gained to date has been with the use of dose escalation to essentially the entire peripheral zone [36–39].

The Virtual HDR series was most recently updated in abstract form in 2012, reporting 51 patients (63 % low-risk; 37 % intermediate-risk), with a median follow-up of 3.5 years (range 6—60 months) [40]. Actuarial 4-year biochemical disease-free survival in this series is 98 % (both definitions), with a 100 % local control rate and a median 4-year PSA nadir of 0.1 ng/mL. There is no acute toxicity in this series greater than grade 2 in either the GU or GI domain. Late toxicity is primarily observed in the GU domain; with 18 and 6 % rates of late grade 2 and 3 GU toxicity, respectively. All observed cases of grade 3 GU toxicity are comprised of urethral strictures requiring surgical correction—a comparable incidence versus the predicate actual prostate HDR brachytherapy series upon which this series was based [41]. There is a statistically significant correlation of increased baseline I-PSS score with an increased rate of grade 2 or higher late GU toxicity, suggesting that caution is warranted in the use of this specific technique and dose-fractionation in patients that have significant preexisting obstructive uropathy (p = 0.0039; I-PSS score as continuous variable versus maximum grade GU toxicity—previously unpublished data). Corresponding rates of late grade 2 and 3 GI toxicity in this series are 2 and 0 %. Patient accrual is scheduled to continue in this series for the foreseeable future, and a larger multi-institutional series that began in 2007, evaluating the same technique, has also completed accrual [42].

The UCSF group has similarly described a “HDR-like” prostate CyberKnife SBRT approach, albeit with less extreme intraprostatic dose escalation versus the Virtual HDR method (UCSF series prescribes to the 60–80 % isodose line, equating to moderately less intraprostatic dose heterogeneity), using the same 3,800 cGy/4 fraction monotherapy fractionation scheme as described above, as well as a 1,900 cGy/2 fraction prostate boost in conjunction with wider field “conventional” pelvic radiotherapy to 45–50 Gy for patients felt to be at higher risk of extraprostatic disease [31]. This series describes 38 patients (20 SBRT monotherapy and 18 SBRT boost).

With a maximum follow-up of 43.5 months, the median observed PSA nadir in this series is 0.35 ng/mL, likely not the true nadir, as their median follow-up is only 18 months. Their late grade 2 and 3 GU toxicity rates measure 8 and 5 %, respectively, with a 3 and 0 % incidence of late grade 2 and 3 GI toxicity, respectively—extremely similar to the levels observed in the San Diego Virtual HDR series described above. The extra toxicity nuance added by the UCSF series is the finding of no apparent difference in the incidence of grade 2 or higher GU or GI toxicity whether this SBRT method is used as monotherapy or as a boost in conjunction with large volume “conventional” pelvic radiotherapy, though absolute numbers of patients in this series remains small, with less than mature follow-up, such that this has to be regarded as a “preliminary” finding of toxicity equivalence between methods.

Experience with gantry-based SBRT regimens as monotherapy for low- to intermediate-risk prostate cancer has also emerged, though the efficacy result seems less established with this approach, due to greater variability of reported outcomes. The initial gantry-based SBRT approach described in contemporary literature has been the so called “SHARP” (Stereotactic hypofractionated accurate radiotherapy of the prostate) regimen described by Madsen et al., in 2007 [43]. This method applied a dose of 33.5 Gy/5 fractions prescribed at the isocenter, using a fixed beam non-coplanar IGRT approach, treating the prostate with a 4–5 mm expansion margin from the prostate to the block edge (which equates to a far smaller margin between the prostate and “quasi full dose” 90 % isodose line). The authors assumed a prostate cancer α/β [alpha/beta] ratio of 1.5, equating to a 2 Gy/fx radiobiological dose equivalent of 78 Gy at the isocenter (author note—This decreases to a total “conventional” fractionation dose equivalent 64 Gy at the 90 % isodose line, which is at or barely beyond the margin of the prostate; lower still if the α/β [alpha/beta] ratio actually exceeds 1.5 Gy). Forty low risk patients were prospectively enrolled and evaluated in this study.

The biochemical disease-free survival rate of the “SHARP” study is unclear, as the result is substantially different depending on which specific definition of biochemical relapse is applied—70 % disease-free survival at 4 years using the ASTRO biochemical relapse-free definition versus 90 % using the Phoenix definition, with the higher biochemical relapse rate implied by the ASTRO definition potentially explained by “delayed PSA bounces” according to the authors. The most common PSA nadir level observed in this study was 0.6–1.0 ng/mL, with 74 and 32 % of all patients achieving a nadir level below 1.0 and 0.5 ng/mL, respectively. Comparison of the PSA nadir values in this series versus those reported in external beam radiotherapy literature caused the authors to conclude that this regimen likely has a biologic potency in the range of 73–78 Gy of “conventionally” fractionated radiotherapy.

Toxicity was reasonably low in this study for both the GU and GI domains. There was a single acute GU grade 3 toxicity and no delayed toxicity in either domain higher than grade 2, with 20 and 7.5 % delayed grade 2 GU and GI toxicities, respectively. The authors concluded that there is probably room for dose escalation. An update of this series with 5-year median follow-up was provided in abstract form in 2010, reporting 71 and 93 % 5-year biochemical relapse-free survival using the ASTRO and Phoenix definitions, respectively, with a median PSA nadir value of 0.65 ng/mL, implying no further degradation in efficacy to 5 years [44].

The substantial uncertainty in biochemical relapse-free survival with the SHARP regimen depending on relapse definition and the low DFS rate imputed by the ASTRO definition suggests the possibility that this regimen may not have an acceptable rate of efficacy. This concern is amplified by the fact that the suboptimal result in this series occurred in spite of restricting the treatment to low-risk patients. At best, the long-term efficacy of this specific SBRT regimen remains uncertain. Acknowledging this concern, the SHARP study authors have suggested consideration of dose escalation in both their initial and follow-up reports [43, 44].

A subsequent gantry-based prostate SBRT series described by the UT Southwestern group has employed a substantially more aggressive approach, sequentially evaluating dose levels of 45 Gy/ 5 fractions, 47.5 Gy/5 fractions and 50 Gy/5 fractions in cohorts of 15 low-risk and intermediate-risk patients, with escalation to the each succeeding dose level contingent upon the demonstration of reasonable acute safety of the previous dose level [45]. A total of 45 patients were described in their report. The treatment was delivered to a volume that included the prostate + 3 mm uniform margin expansion to form the PTV, with the dose prescribed to cover >95 % of the PTV, delivered using an axial plane IMRT method with daily image guidance, either by a helical Tomotherapy device or by a step and shoot MLC-based linear accelerator device.

At 12–30 months of follow-up, PSA control by the Phoenix definition (nadir + 2 ng/mL) is 100 %. The maximum grade 2 GU and GI toxicity incidence rates of 31 and 18 %, respectively, and maximum grade >3 GU and GI toxicity rates of 4 and 2 %, respectively; overlap the toxicity incidence of other contemporary radiotherapy series. This suggests a non-excessive severe toxicity rate in spite of aggressive SBRT dosing, with the caveat that a maximum follow-up of 30 months is inadequate to fully define the late toxicity incidence. As such, the late toxicity result of this approach must still be regarded as preliminary. The authors postulate that the relatively low toxicity incidence observed to date in spite of aggressive dosing, is due to tight margins, strict dose limitation of all adjacent normal tissues (bladder, rectum and urethra) and the use of an endorectal balloon to stabilize the rectum and displace its lateral and posterior walls out of the high dose volume. The study has now expanded to multiple centers and is ongoing.

Another gantry-based prostate SBRT series of 65 low-risk patients has been described in abstract form by Mantz et al., using a Varian Trilogy cone beam CT (CBCT) volumetric-guided IMRT approach, with Calypso electromagnetic tracking to provide 4D localization, with an action level of 2 mm [46]. They deliver a dose of 40 Gy/5 fractions (q.o.d. schedule) to the PTV (exact expansion from the prostate not described in the abstract). With 36 months of median follow-up, the PSA response is excellent (median PSA nadir 0.3 ng/mL at 3 years) and there is a zero incidence of reported grade 3 or higher toxicity in the GU or GI domains. As with other gantry-based reports, this series seems encouraging, with further follow-up and more complete reporting necessary to fully validate the results.