The role of radiation therapy in the treatment of metastatic castrate-resistant prostate cancer

Abstract

In the setting of castrate-resistant prostate cancer, patients present with a variety of symptoms, including bone metastases, spinal cord compression and advanced pelvic disease. Fortunately, a variety of radiotherapeutic options exist for palliation. This article focuses on these options, including both external beam radiotherapy and radiopharmaceuticals.

Introduction

Androgen deprivation therapy (ADT) forms the basis for management of metastatic or advanced prostate cancer not amenable to curative therapy. Response rates are 80–90% but the median duration of hormone sensitivity ranges from only 16 months in patients with distant metastases to 33 months in patients with localized disease [Hellerstedt and Pienta, 2002; Ross et al. 2008]. Almost 50% of the 3040 patients with newly diagnosed metastatic disease registered to the SWOG 9346 trial demonstrated castrate resistance within 7 months, and were thus ineligible to be randomized to a trial of intermittent therapy [Hussain et al. 2013]. Disease progression despite androgen blockade heralds the castrate-resistant phase of illness which may be complicated by a variety of symptoms from metastases and progressive disease. The aim of this article is to review the use of radiotherapy in the management of castrate-resistant prostate cancer (CRPC).

Localized external beam radiotherapy for bone metastases

External beam radiotherapy (EBRT), recognized for many years as effective palliation of bone metastases, may employ a variety of doses and fractionation schedules. Most commonly in North America, palliative treatment schedules for bone metastases include 3000 cGy in 10 fractions or 2000 cGy in 5 fractions, with single fractions (800 cGy) reported as being used less frequently [Ben-Josef et al. 1988; Chow et al. 2000; Fairchild et al. 2009]. A number of phase III trials have compared the effectiveness of multifraction versus single fraction treatments. In 2007, Chow and colleagues published a meta-analysis of 19 studies comparing a single fraction with multiple fractions for initial treatment of bone metastases for several tumor types, including prostate cancer, showing that the overall response, complete response, and rates of pathological fracture and spinal cord compression are similar [Chow et al. 2007] (Table 1). Although retreatment of an index lesion is 2.5 times more likely if a single fraction is used, the majority of patients are able to avoid more prolonged treatments without the need for returning for a repeat treatment. Indeed, many studies show no difference in duration of response between single fraction and multifraction schedules (4–12.5 months) [Price et al. 1986; Hoskin et al. 1992; Niewald et al. 1996; Gaze et al. 1997; Jeremic et al. 1998; Nielsen et al. 1998; Bone Pain Trial Working Party, 1999; Steenland et al. 1999; Hartsell et al. 2005].

Table 1.

Summary of meta-analysis of external beam radiation treatment schedules (single fraction versus multiple fraction) for bone metastases [Chow et al. 2007].

Metric	Number of studies	Pooled odds ratio (95% CI)	p	Pooled estimate of outcome
				Single fraction	Multiple fraction
Overall response	16	0.99 (0.95–1.03)	0.6	58%	59%
Complete response	13	0.97 (0.88–1.06)	0.51	23%	24%
Likelihood of retreatment	9	2.5 (1.76–3.56)	<0.0001	20%	8%
Pathological fracture	8	1.1 (0.61–1.99)	0.75	3.2%	2.8%
Spinal cord compression	5	1.44 (0.9–2.3)	0.13	2.8%	1.9%

CI, confidence interval.

Onset of pain relief occurs between 2 and 4 weeks and is similar between multifraction and single-fraction schedules [Niewald et al. 1996; Jeremic et al. 1998; Nielsen et al. 1998; Bone Pain Trial Working Party, 1999; Steenland et al. 1999; Hartsell et al. 2005]. Likewise, acute toxicity and quality of life are generally similar, although increased acute toxicity with multifraction treatments has been reported [Price et al. 1986; Gaze et al. 1997; Jeremic et al. 1998; Nielsen et al. 1998; Bone Pain Trial Working Party, 1999; Steenland et al. 1999; Hartsell et al. 2005]. These data indicate that excellent palliation of bone metastases can be achieved with single-fraction EBRT. Not only is this more convenient for patients but economic evaluations have shown that single fraction radiotherapy is more cost effective than multifraction regimens, chemotherapy and even analgesic regimens [van den Hout et al. 2003; Konski, 2004].

Reirradiation of painful bone metastases with EBRT is also possible, whether initial treatment has been single or multiple fractions. A meta-analysis of seven studies has shown a pooled overall response rate of 58% [95% confidence interval (CI) 49–67%], which is similar to that of initial treatment [Chow et al. 2007; Huisman et al. 2012]. Time to response averages between 3 and 5 weeks and mean duration of response ranges from 15 to 22 weeks [Huisman et al. 2012].

Stereotactic ablative radiotherapy (SABR) is an emerging technology that uses image guidance to deliver a high dose of extremely conformal EBRT in single or a few fractions (<5) with millimeter precision [Sahgal et al. 2012]. The aim of SABR is to achieve long-term control of spinal disease. Such treatment is time consuming for patients and resource intensive [Haley et al. 2011; Amdur et al. 2009], and consequently may not often be considered appropriate for patients with CRPC. Retrospective case series and phase I/II trials have shown that radiologic control of spinal metastases at 1 year ranges from 80.5% to 90% [Chang et al. 2007; Ahmed et al. 2012; Garg et al. 2012; Wang et al. 2012], with pain response being similar to conventional EBRT [Nelson et al. 2009; Lee and Chun, 2012].

In patients experiencing pathological fracture, surgery is often the first line of treatment, with postoperative radiotherapy recommended with fractionation schemes similar to primary palliation (3000/10–2000/5) [Chow et al. 2003]. After surgical stabilization of a pathological fracture, postoperative radiotherapy is associated with a higher probability of normal extremity use compared with surgery alone (53% versus 11.5%) and a decreased risk of second orthopedic procedures (15% versus 3%) [Townsend et al. 1994].

Hemibody irradiation

Hemibody irradiation (HBI) can be used to manage widespread bone metastases in CRPC. Three types of fields have been described: upper hemibody (UHBI) from mastoid process to iliac crest, lower hemibody (LHBI) from iliac crest to ankle, and mid-hemibody from top of the diaphragm to the bottom of the obturator foramen [Salazar et al. 2001]. Most case series reporting on the effectiveness of HBI have used single fractions of 800 cGy for LHBI and 600 cGy for UHBI (to limit lung toxicity). HBI provides high rates of response (63–96%) with a fast onset of pain relief (1–3 days) [Salazar et al. 1978, 1986, 1996, 2001; Hoskin et al. 1989; Zelefsky et al. 1989; Dearnaley et al. 1992; Poulter et al. 1992; Bashir et al. 2008; Berg et al. 2009]. Significant toxicities after single fraction HBI include ‘acute radiation syndrome’ characterized by severe nausea and vomiting, fever, tachycardia and hypotension [Salazar et al. 1978] and pneumonitis characterized by cough, dyspnea and characteristic radiographic changes [Van Dyk et al. 1981; Fryer et al. 1978]. Radiation dose reduction, lung shielding and management of toxicity with glucocorticoids, antinauseants and intravenous hydration have improved these complications [Salazar et al. 1978; Priestman et al. 1990; Bashir et al. 2008; Berg et al. 2009].

Multi-fraction regimens appear to be better tolerated and have a more durable response [Salazar et al. 2001]. However, given the hematologic toxicity of hemibody radiotherapy and the increasing availability of effective chemotherapeutic agents for CRPC, the role of hemibody radiotherapy has declined significantly.

Radiopharmaceuticals

Radiopharmaceuticals are radioactive isotopes or compounds that are injected intravenously to deliver a measurable dose of radiation to bone because of the calcium-mimetic properties of the compound [Lewington et al. 1991]. Single or repeated treatments are given to patients with prostate cancer with multiple symptomatic bone metastases on an outpatient basis and have been used alone or in combination with EBRT, chemotherapy and bisphosphonates [Porter et al. 1993; Quilty et al. 1994; Sciuto et al. 2002; Smeland et al. 2003; Lam et al. 2008; Fizazi et al. 2009]. The effectiveness and toxicity of radiopharmaceuticals depends on the characteristics of the radionuclide. Ideal radiopharmaceuticals deliver radiation over a range of millimeters to minimize bone marrow toxicity and should have a half life sufficient to provide therapeutic effect but short enough to avoid myelosuppression [Robinson et al. 1995].

Radium 223

Radium 223 (Ra 223) is a radiopharmaceutical that has generated much interest due to its unique physical properties and recent findings of a survival benefit [Nilsson et al. 2013; Parker et al. 2013a]. Ra 223 has a physical half life of 11.4 days and emits primarily α particles (two protons and two neutrons, identical to a helium nucleus) which are densely ionizing and deposit a substantial amount of energy over a very short range (<100 μm). α particles primarily cause double-strand DNA breaks resulting in nonrepairable damage to cellular DNA and more efficient cell kill. β particles and γ radiation, as are produced from other radionuclides, are less densely ionizing over a greater range in tissue [Nilsson et al. 2005].

Phase II trials have shown that Ra 223 is well tolerated and achieves response rates as high as 75% with duration of pain relief as long as 44 days. Ra 223 is the first radiopharmaceutical to demonstrate a clear survival benefit as reported by Parker and colleagues in an international phase III randomized trial (ALSYMPCA). A total of 922 patients with CRPC received six injections of Ra 223 (50 kBq/kg, every 4 weeks) or placebo. Eligibility criteria included symptomatic men with progressive CRPC with two or more bone metastases, no known visceral metastases and castrate serum testosterone (<1.7 nMol/liter). In both groups, 57% of patients had received docetaxel while the remainder were either not medically suitable or docetaxel was not available. Planned interim analysis found improved median survival in the Ra 223 group [14.9 versus 11.3 months, hazard ratio (HR) for survival = 0.7; 95% CI = 0.55–0.88, two-sided p = 0.002]. This study also demonstrated increased time to skeletal related events (15.6 versus 9.8 months, HR = 0.66; 95% CI = 0.52–0.83, p < 0.001) and improved quality of life in the Ra 223 group [Parker et al. 2013a and b]. Ra 223 and placebo shared similar rates of all adverse events (93% Ra 223, 96% placebo), adverse events of grade 3 or higher (56% Ra 223, 62% placebo) and serious adverse events (47% Ra 223, 60% placebo). Only 16% of study participants discontinued treatment due to adverse events in the Ra 223 arm compared with 21% in the placebo arm. Given the improvement in survival, excellent tolerance and ease of use (within a few weeks it can be disposed of without radiation precautions), Ra 223 is likely to become the radiopharmaceutical of choice.

Other radiopharmaceuticals

Strontium 89

Strontium 89 (Sr 89) has been in clinical use for over three decades. It has a half life of 50.5 days and emits β particles (electrons) with a maximum range in tissue of 6–8 mm (mean 2.4 mm) [Robinson et al. 1995; Windsor, 2001; Biersack et al. 2011]. A summary of pain response and toxicity for Sr 89 in CRPC is shown in Table 2.

Table 2.

Summary of response and toxicity of Sr 89 in castrate-resistant prostate cancer.

Response/toxicity		Outcome	Reference
Pain response	Overall response	34.7–78%	[Pons et al. 1997; Sciuto et al. 2002; Kraeber-Bodere et al. 2000; Oosterhof et al. 2003]
	Onset of response	2–3 weeks
	Maximal response	5–6 weeks
	Duration of response	2–6 months
Hematological toxicity	Nadir	4–6 weeks	[Quilty et al. 1994; Lee et al. 1996; Pons et al. 1997; Kraeber-Bodere et al. 2000; Turner et al. 2001; Sciuto et al. 2002]
	Recovery	3 months
	Grade 1/2	5.7–20%
	Grade 3/4	5–11%
Other common toxicity	Nausea and vomitingPain flare	13% (⩾grade 3 = 4%)10%	[Pons et al. 1997; Kraeber-Bodere et al. 2000; Turner et al. 2001; Sciuto et al. 2002; Oosterhof et al. 2003]

Sr 89 has been used with EBRT to increase the durability of pain response in patients with multiple bone metastases. The Trans Canada trial randomized 126 patients with CRPC to receive palliative EBRT (10 Gy single treatment to 30 Gy in 10 fractions) followed by a single injection of 400 MBq Sr 89 or placebo. There was no difference in pain response at the index lesion; however, quality of life was improved in the Sr 89 group due to a decrease in new bone metastases and analgesic use [Porter et al. 1993]. These results were confirmed by Quilty and colleagues who randomized 305 patients with CRPC to Sr 89 or placebo after either local EBRT or hemi-body radiotherapy (HBRT). Despite worse prognostic factors in the Sr 89 group, overall survival and pain response were similar between groups. At 3 months, more patients remained free of new sites of bone pain in the Sr 89 group (Sr 89 + local EBRT = 61.9% versus local EBRT = 41.7%, p < 0.05; Sr 89 + HBRT = 73.3% versus HBRT alone = 51%, p < 0.05) and fewer required retreatment [Quilty et al. 1994].

Sr 89 has also been combined with chemotherapy in patients with bone metastases from CRPC. A total of 105 patients were randomized to induction docetaxel/ketoconazole/vinblastine/estramustine followed by further docetaxcel alone or docetaxel and Sr 89. Time to progression (13.9 versus 7 months, p < 0.05) and overall survival (HR = 2.76, 95% CI = 1.44–5.29, p < 0.05) were improved with the addition of Sr 89. Toxicity tended to be higher in the Sr 89 group with respect to neutropenia, esophagitis, dyspepsia, gastritis and fatigue [Tu et al. 2001]. Sciuto and colleagues randomized 70 patients with bone metastases from CRPC to a single cycle of cisplatin with or without a concurrent single injection of 148 MBq Sr 89. Overall pain relief was greater in the Sr 89 group (91% versus 63%, p < 0.05), the median duration was longer (120 versus 60 days, p < 0.05) and progression on bone scan reduced (27 versus 64%, p < 0.05). Although there was no difference in overall survival, treatment was well tolerated [Sciuto et al. 2002]. In the recent TRAPEZE trial, 757 patients with CRPC were randomized to receive six cycles of docetaxel alone, with a single dose of Sr 89 or zoledronic acid or all three. Sr 89 was associated with a benefit for clinical progression-free survival (HR = 0.845; 95% CI = 0.72–0.99, p = 0.036, Cox regression) while zoledronic acid was associated with a significant benefit for skeletal related event-free interval (HR = 0.76; 95% CI = 0.63–0.93, p = 0.008). There was no effect of either agent on overall survival [James et al. 2013]. As monotherapy in patients with multiple bone metastases from CRPC Sr 89 is effective in decreasing pain at the cost of mild to moderate myelosuppression, Sr-89 may be combined with EBRT or chemotherapy to increase the durability of response.

Samarium 153

Samarium 153 (Sm 153) emits β particles with a maximum range in tissue of 2.5–3.1 mm (mean 0.6 mm), but it also emits low-energy γ particles (103 keV) that result in a small degree of extracorporeal exposure [Eary et al. 1993; Biersack et al. 2011]. The physical half life of 46.3 h is shorter than Sr 89, resulting in a faster delivery and response [Resche et al. 1997]. After a single injection, overall response rates for pain are 65–74% with onset at 1 week, and duration 2–3 months [Collins et al. 1993; Resche et al. 1997; Serafini et al. 1998; Tian et al. 1999; Sartor et al. 2004]. Grade 3 myelosuppression is significant, occurring in 3–12%, with blood counts reaching nadir by week 3–5 [Resche et al. 1997; Serafini et al. 1998; Tian et al. 1999; Sartor et al. 2004]. Despite this, several phase I studies have safely combined Sm 153 with docetaxel [Suttmann et al. 2008; Fizazi et al. 2009; Morris et al. 2009; Tu et al. 2009; Lin et al. 2011], and zoledronic acid [Lam et al. 2008]. Other radiopharmaceuticals such as Rhenium 188 and Rhenium 186 remain investigational and are not yet available for widespread use [Maxon et al. 1990; Palmedo et al. 2003; Liepe et al. 2003].

Spinal cord compression

Radiotherapy and corticosteroid treatment are the standard of care for malignant spinal cord compression (MSCC) [Loblaw et al. 2005], with no benefit for laminectomy after radiotherapy shown in either retrospective studies or a randomized controlled trial [Gilbert et al. 1978; Young et al. 1980; Sorensen et al. 1990]. Although a randomized controlled trial of surgery followed by EBRT compared with EBRT alone initially reported improved ambulation rates, time to ambulation and pain control in the surgery plus EBRT group [Patchell et al. 2005], this study included highly selected patients and was criticized for lack of external validity [Koch and de Keyser, 2006; Rades et al. 2010]. A subsequent retrospective matched pair analysis attempted to clarify these issues [Rades et al. 2010]. Using 11 potential prognostic factors to match cases, no difference was found in post-treatment motor function in either ambulatory patients (69% versus 68%) or nonambulatory patients (30% versus 26%) after surgery plus EBRT compared with EBRT alone. In addition, 10% of patients in the surgery group had significant morbidity after surgery, including wound infection requiring reoperation, bleeding, pneumonia and pulmonary embolism [Rades et al. 2010]. However, with appropriate patient selection, surgery can be recommended in good prognosis patients provided surgical consultation can take place within 24 h of the diagnosis of MSCC in order to preserve neurological function [Husband, 1998; Loblaw et al. 2012]. Patients with spinal instability secondary to metastases may also benefit from surgery as this allows for concomitant spinal stabilization [Loblaw et al. 2005].

Radiotherapy and steroids are the standard of care for MSCC if patients are not fit for surgery or if there is multilevel spinal disease. Post-treatment ambulation rates vary from 60% to 100% [Zelefsky et al. 1992; Marazano and Latini, 1995; Maranzano et al. 1996, 1997, 2005; Rades et al. 2004] in those with good ambulation prior to treatment and are approximately 38% in those with poor ambulation pretreatment [Maranzano et al. 1997]. Likewise, maintenance of sphincter function is also better in those with good pretreatment function (98% versus 14–44%) [Maranzano et al. 1997, 2005]. These results indicate that a high index of suspicion, prompt diagnosis and urgent referral for radiotherapy are essential for optimal outcomes. Local recurrence in the spine may be amenable to retreatment [Schiff et al. 1995; Sahgal et al. 2009]. The decision for retreatment must be made on a case-by-case basis, considering factors such as performance status, details of prior radiotherapy and interval since treatment. Acute toxicity for a second course of radiotherapy to the spine is expected to be similar to the initial course. The risk of late neurologic toxicity is dependent on total dose and may be significant for patients with sufficiently long life expectancy.

External beam radiotherapy for symptomatic pelvic disease

Although skeletal metastases are the most common cause of morbidity in patients with CRPC, pelvic symptoms predominate in 10–18% [Fossa et al. 1992; Otnes et al. 1995]. Locally advanced pelvic disease may manifest as obstructive or irritative urinary symptoms, hematuria, ureteral obstruction, renal impairment, pelvic pain or rectal obstruction, urgency and tenesmus [Megalli et al. 1974; Hernes et al. 2000; Din et al. 2009; Caravatta et al. 2012; Gogna et al. 2012]. Local therapies such as bladder irrigation, catheterization, transurethral resection of the prostate, fulguration of bleeding vessels and blood transfusions provide relief but often require repeated treatments [Crain et al. 2004]. Pelvic exenteration is associated with significant morbidity [Kamat et al. 2003] and systemic therapy is often ineffective for localized disease. Median survival in patients with CRPC referred for palliative radiotherapy ranges from 8 to 10 months [Fossa et al. 1992; Taylor et al. 1993] but may be higher if disease is confined to the pelvis [Fossa, 1987; Lankford et al. 1995]. EBRT can provide a durable palliative response in these patients with minimal toxicity.

The most common dose and fractionation regimens range from 2000 cGy in 5 fractions to 4000 cGy in 20 fractions [Lutz et al. 2007]. Large single fractions yield low response rates and significant toxicity. Spanos and colleagues attempted a phase I/II trial of 1000 cGy in one fraction given three times over 3 months but the response rate was only 41% and late grade 3–4 gastrointestinal toxicity was unacceptably high at 30% [Spanos et al. 1987].

Lower dose schedules limited to 1 week are associated with higher response rates and lower toxicity. In a study of 58 men with CRPC, 2000 cGy in five daily fractions produced an overall symptom response rate of 89% and complete response rate of 57% [Din et al. 2009]. Response rates for hematuria, urinary outflow obstruction, pain and rectal obstruction were highest at 6 weeks (overall response: 81%, 75%, 63%, 88%, respectively) but decreased by 7 months to 29, 54, 38 and 62.5% [Din et al. 2009]. Acute toxicity with a four- to five-fraction schedule is limited to grade 1–2 [Caravatta et al. 2012]. As up to half the patients in these series were lost to follow up or died of disease prior to response assessment, these response rates reflect the experience of those who survived [Din et al. 2009].

Durability of response may be increased with higher dose schedules of 4500–6000 cGy delivered over approximately 4 weeks. With these schedules, overall symptomatic response rate is 60–91% [Hindson et al. 2007; Gogna et al. 2012]. Gross hematuria resolves in 60–78% and 60–82% obtained significant relief of bladder outlet obstruction [Hindson et al. 2007; Gogna et al. 2012]. Median time to local-regional progression with higher dose schedules is longer (43 months) compared with shorter, lower-dose schedules (12 months) [Hindson et al. 2007; Din et al. 2009; Gogna et al. 2012]. Acute side effects with a higher total dose consist of grade 1 and 2 urinary (48%) and gastrointestinal (30%) toxicity [Gogna et al. 2012]. Late grade 3 and greater urinary and gastrointestinal toxicity is rare and occurs in 0–6% [Spanos et al. 1993; Furuya et al. 1999; Gogna et al. 2012]. Pelvic radiotherapy can be useful for palliation of symptoms, such as bladder outlet obstruction, hematuria and pain with mild acute toxicity. Shorter low-dose schedules can be used for patients with a limited expectation of survival but longer higher dose schedules are associated with a more durable response in patients with longer life expectancy.

Conclusion

In the setting of CRPC, progression of disease causes significant morbidity and impact on quality of life. Local EBRT palliates bone pain in the majority of men with minimal toxicity. Widespread bone metastases can be effectively managed with systemic radiopharmaceuticals such as Ra 223. In the setting of spinal cord compression, timely intervention is essential in order to preserve motor function. Radiotherapy is indicated either postoperatively following decompression or stabilization, or as monotherapy in those not suitable for surgery. In a subset of patients with CRPC, pelvic symptoms may predominate. EBRT treatment schedules should be tailored to fit the individual prognosis. Timely referral of patients is recommended to optimize treatment options and outcomes.