Focused ultrasound for treatment of bone tumours

Abstract

Purpose: Focused ultrasound (FUS) is a modality with rapidly expanding applications across the field of medicine. Treatment of bone lesions with FUS including both benign and malignant tumours has been an active area of investigation. Recently, as a result of a successful phase III trial, magnetic resonance-guided FUS is now a standardised option for treatment of painful bone metastases. This report reviews the clinical applications amenable to treatment with FUS and provides background on FUS and image guidance techniques, results of clinical studies, and future directions. Methods: A comprehensive literature search and review of abstracts presented at the recently completed fourth International Focused Ultrasound Symposium was performed. Case reports and older publications revisited in more recent studies were excluded. For clinical studies that extend beyond bone tumours, only the data regarding bone tumours are presented. Results: Fifteen studies assessing the use of focused ultrasound in treatment of primary benign bone tumours, primary malignant tumours, and metastastic tumours meeting the search criteria were identified. For these clinical studies the responders group varied within 91–100%, 85–87% and 64–94%, respectively. Major complications were reported in the ranges 0%, 0–28% and 0–4% for primary benign, malignant and metastatic tumours, respectively. Conclusions: Image-guided FUS is both safe and effective in the treatment of primary and secondary tumours. Additional phase III trials are warranted to more fully define the role of FUS in treatment of both benign and malignant bone tumours.

• Ablation
• bone tumour
• high intensity focused ultrasound
• image guidance
• pain palliation

Introduction

Bone tumours, whether benign or malignant, often present challenges with local disease control and pain. Several bone tumour clinical studies have been conducted in the last decade, focused on both curative and palliative treatments. One emerging modality in the treatment of bone tumours is high intensity focused ultrasound (FUS), which causes thermal damage of tissue using focused sound waves. The FUS treatment is guided in real time by magnetic resonance (MRgFUS) or ultrasonography (USgFUS), which provides a feedback mechanism to increase treatment accuracy. This technique is non-invasive and does not use ionising radiation, which contrasts with standard approaches such as surgery or radiotherapy used in cancer treatment. The aim of this paper is to review bone tumour clinical studies using both USgFUS and MRgFUS, including benign tumours, primary malignancies, and metastatic bone tumours.

The vast majority of primary bone tumours are benign. Many are asymptomatic and remain undetected until radiographic examinations are performed for other reasons. Thus, the actual incidence of benign bone tumours is currently indeterminate [1]. One of the most common is osteoid osteoma, comprising nearly 10% of benign bone tumours [2]. It consists of a nidus, or core of growing cells, surrounded by a thick bony shell that is normally found in the appendicular skeleton [3]. They are predominantly seen in children, teenagers, and young adults, with more than 80% of tumours seen in individuals between the ages of 5 and 30 years [4]. The presence of a benign bone lesion does not necessarily mean lack of aggressiveness; histologically benign lesions may be locally highly aggressive, causing severe discomfort and pain [5].

Malignant primary bone tumours constitute 0.2% of all malignancies in adults and approximately 5% of childhood malignancies [1]. The most common primary malignant bone tumour is osteosarcoma, representing 35% of total primary bone malignancies [6]. The 5-year overall survival rate for bone sarcoma is 68%. Osteosarcomas usually develop in areas where the bone is growing quickly, such as near the ends of the long bones, and are most common in children and young adults. These particular tumours require aggressive local treatment and at times may require amputation.

Bone metastases are the most common source of pain in cancer patients [7]. Autopsy studies have shown that up to 85% of patients with breast, prostate and lung cancer have bone metastases at the time of death. Complications due to skeletal metastases, including intractable pain, fracture, and decreased mobility, can dramatically reduce performance and quality of life and lead to depression and anxiety [8]. Bone metastases portend a poor prognosis, with patients generally facing a median survival of 3 years or less. In lung cancer, survival is typically measured in months, whereas patients with breast cancer or prostate cancer often live for several years. Bone metastases are generally classified as osteolytic, which is characterised by destruction of normal bone, or osteoblastic, characterised by deposition of new bone. This distinction is not absolute since metastatic lesions can contain both osteolytic and osteoblastic components [9].

There are two distinct treatment approaches for bone tumours: curative and palliative. For thermal ablation techniques in particular, curative treatment aims for complete coagulation necrosis of the primary lesion. In palliative treatment of primary or secondary bone tumours, therapeutic goals include pain palliation, tumour reduction, prevention of impending pathological fractures, and/or tumour decompression. The denervation of the periosteum, which contains pain-reporting nerve fibres, is considered a major factor in pain palliation. This explains the rapid relief following ultrasound treatment which is characterised by enhanced power deposition in bone relative to surrounding soft tissues [10].

Clinical applications of thermal ablation therapy are expanding rapidly due to the ability to produce immediate obliteration of either small tumours or benign disease, when the target volume is located in an anatomic region with sufficient biological reserve and/or separation from surrounding critical normal tissues. Either cryogenic temperatures (<−40 °C) or high temperatures >50–60 °C produce complete ablation of the tissue vasculature and most importantly disrupt and coagulate all structural proteins in the target cells. These effects are non-subtle and irreversible, leading to immediate and complete cellular death. In lower dose regions, in the transition zone of a thermally coagulated lesion, there is a percentage of cells that survive the thermal insult. In this border region, adjunctive therapies such as radiation or chemotherapy may be helpful in further expanding the radial penetration of effective therapy by providing some differential effect on weaker tumour cells relative to the surrounding normal tissue host.

There are numerous technologies available to produce a high intensity heat focus at depth in tissue. Invasive approaches include interstitial or intracavitary placement of radiofrequency (RFA), microwave (MWA), laser, or thermal conduction sources into the target region (Table 1). These implantable heat sources are well characterised in the literature [11–14], and for many tissue sites such as liver, breast, prostate, kidney and lung, they can heat deep tissue targets effectively without unacceptable normal tissue complications [4,15]. Insertion of heat sources directly into bone tissue is difficult, however, such that a non-invasive approach with precise control of the focal zone at depth is required. Taking advantage of the short wavelength and deep penetration of ultrasound pressure waves, beams may be combined from large external transducer arrays to produce an intense focal hot spot at depth, non-invasively; thus the name high intensity focused ultrasound (FUS).

Table 1. Thermal ablation techniques for bone cancer treatement.

Bone tumour	USgFUS	MRgFUS	RFA	MWA	Laser ablation	Cryoablation
Primary malignant osteosarcoma	[19,33,34,46]	–	–	[57]	[58]	–
Other primary malignant tumours	[19,33,34,46]	[5,19,33,46,49]	[59]	[60]	–	[61]
Primary benign osteoid osteoma	–	[17,48,50]	[62,63]	[64]	[65]	[66]
Other primary benign tumours	–	[67]	[68–71]	[72]	–	[73]
Bone metastasis	[33]	[16,18,20,21,36, 43,53–55]	[74, 75]	[76]	–	[77]

Methods

This section reviews the FUS method, image guidance techniques, and FUS devices used in ongoing clinical studies. Furthermore, we address primary and secondary end points typical in the analysed clinical studies as well as the tools for outcome assessment.

High-intensity focused ultrasound

In general, ultrasound systems produce an acoustic pressure wave from one or more piezoelectric transducers that operate at a frequency in the range of 0.2–4.0 MHz. Attenuation of ultrasound energy as it propagates through biological tissue results in a rise of temperature in the treated volume. Coagulative necrosis and cell death occur within seconds at temperatures in the range of 65–85 °C [16], so that the ablative treatment must be well localised to the tumour target and avoid surrounding normal tissues. To achieve the required focusing and rapid elevation of tissue temperature at depth, the ultrasound energy is usually intensified at a focal spot by using multiple intersecting ultrasound beams that converge on the target. Typical clinical focused ultrasound systems produce high intensity focal volumes of 0.2–20 mm³ at depth, while spreading out the ultrasound energy over a large surface area under the transducer array so as to have negligible effect on tissues outside the focus. Because of the high temperatures achieved in the high intensity focus, sonications are generally limited in duration to only a few seconds. This reduces the potentially detrimental thermal smearing and energy dissipation effects to surrounding tissue due to blood perfusion. Because of these technical issues, ablation of a typical tumour requires sequential tiling of multiple sonications to create homogeneous thermal damage and coagulative necrosis of the entire target volume [16,17].

With focused ultrasound systems, tissue in the path of the ultrasound beams outside the focus is warmed, but only to sub-lethal temperatures. Between successive sonications, the heat deposited in intervening tissue is dissipated by normal tissue conduction and perfusion cooling effects [18]. Due to the low thermal conductivity and high ultrasound attenuation in the periosteum and bone cortex, it is possible to use lower energy levels when treating bone compared to treatment of well-perfused soft tissue. This provides an improved safety profile by reducing thermal damage around the treated bone site [19–21].

Image-guided FUS

Over recent decades there have been numerous external ultrasound array systems for focused heating in the body [22–27], including one system specifically optimised for small animal treatments [28]. At present there are three commercial non-invasive ultrasound heating systems that have been integrated within either US or MR imaging systems for image-guidance of FUS treatments in humans: the ExAblate system (InSightec, Tirat Carmel, Israel) based on the General Electric MR platform [29,30], the Sonalleve FUS system (Koninklijke Philips Electronics, Eindhoven, the Netherlands) based on the Phillips MR platform [31,32], and the Haifu system (Model JC, Chongqing Haifu Medical Technology, Chongqing, China) coupled with a B-type ultrasonography system [33].

In USgFUS, the US imaging probe is situated in the centre of therapeutic transducer array and provides real-time sonography feedback. This way, the user can target the lesion to be treated, guide the US energy deposition, and assess the extent of acute coagulation necrosis in the treated tissue [19,33,34]. MRI adds to focused ultrasound therapy the advantages of high resolution lesion localisation, real time temperature monitoring, and post-treatment tissue evaluation. Lesion visualisation is achieved through the intrinsic high tissue contrast of MRI, while temperature monitoring is achieved through use of specific MRI sequences (proton resonance frequency shift method (PRF)) performed during the ablation [17].

Clinical studies

This review summarises clinical studies to date assessing use of therapeutic ultrasound for bone tumours, including both benign and malignant primary lesions, as well as metastatic disease. Case reports were excluded from analysis. In instances of published updates of previous studies, only the more complete publications were analysed. For clinical studies that extend beyond bone tumours, only the data regarding bone tumours are presented [34,35]. The fourth Focused Ultrasound Symposium (12–16 October 2014 in North Bethesda, MD, USA) reported significant advances in the FUS treatment of bone tumours, and abstracts from that recent meeting were considered under the aforementioned inclusion criteria. The following sections summarise the most relevant aspects of clinical trial design: inclusion criteria for patient selection, end points, and instruments for outcome assessment.

Inclusion criteria for patient selection

For safe and effective ablation therapy, carefully defined patient selection criteria are critically important. The anatomy of the human skeleton necessarily involves many critical neighbouring structures. Patients with bone tumours in difficult to access locations or in close proximity to major nerves, e.g. spinal tumours, have been typically excluded from clinical studies. Some studies also excluded patients with tumours that are larger than 10 cm [19], close to joints [36], close to blood vessels [33], in weight-bearing bones [18,20], or with locally advanced disease [35].

Patients with bone metastases were included in published studies only if they reported moderate or severe pain, typically a score of 4 or greater out of 10 for worst pain in a 24-h period. Furthermore, patients with more than 1–2 metastatic sites were excluded since this type of pain is believed to be better treated with a systemic rather than focal approach [4]. In the case of MRgFUS, MR compatibility must be considered, and patients with claustrophobia or pacemakers were excluded from this image guidance technique. Some studies aimed for primary treatment, so patients with previous radiotherapy were excluded [37].

The Karnofsky Performance Status Scale is widely used to quantify the functional status of cancer patients [38]. One study in particular did not consider patients with a Karnofsky score lower than 70% [19]. Finally, some patients were enrolled only if there was no other treatment technique available or feasible [18–21,33,35].

End points and instruments for outcome assessment

The primary end points of a clinical study are typically safety (phase I) and efficacy (phases II and III). Safety is assessed by occurrence of adverse events, which can be divided into minor and major complications [39]. The definition of efficacy is dependent on the type of tumour and intent of treatment. For primary bone tumours, either malignant or benign, efficacy is analysed in terms of local tumour control, survival, recurrence, and/or conservation of the diseased limb. On the other hand, treatments for bone metastases aim for pain palliation and improved quality of life, where quality of life is considered a secondary end point. Geiger et al. also distinguish technical and clinical end points [17]. The technical end points address short-term complications from the FUS procedure, and will be considered under the umbrella of safety. In palliative treatments, pain progression is also considered.

Different instruments were used to quantify each target outcome, including questionnaires for pain and quality of life assessment, as well as anatomical imaging (CT, MRI, or CE-MRI), functional imaging (PET/CT, SPECT, or scintigraphy), biochemical (e.g. alkaline phosphatase) and histopathology analyses for determination of response. Each clinical study used one or more of these instruments to evaluate clinical outcome.

Safety: The safety end point is evaluated in terms of incidence and severity of complications due to the FUS treatment. These adverse events must be assessed by clinical and imaging examinations, where the latter are especially relevant to detect occult adverse events. Some authors (e.g. [19]) subdivide this endpoint into minor and major complications, where the latter are events that lead to substantial morbidity, disability, increased level of care, or results in hospital admission or substantially lengthened hospital stay [40]. We standardised this classification for all studies.

Local tumour control: Local tumour control is generally accessed via imaging (SPECT, MRI or PET/CT), and may also be assessed with biochemistry and histopathology analysis. Several parameters can be analysed: inflammatory status, bone remodelling and/or remineralisation, tumour vascularity and perfusion, tumour cellular survival and activity, coagulative necrosis of the target area, and particularly tumour size. For primary malignant bone tumours, the presence and condition of metastatic lesions are also evaluated, e.g. using chest radiography and/or CT to define lung metastases [19]. Furthermore, biochemical markers for cancer are also assessed in all clinical studies that address primary malignancies [19,33,34]. To evaluate treatment efficacy via local tumour control, lesion changes were evaluated by Napoli et al. [37] in accordance with the MD Anderson (MDA) criteria [41]. These criteria include quantitative and qualitative assessments of the behaviour of bone metastases. The response is divided into four categories: complete response (CR), partial response (PR), progressive disease (PD), and stable disease (SD). Quantitatively, these criteria define PR as a decrease of 50% or more in the sum of the perpendicular measurements of a lesion, and PD as an increase of 25% or more in this sum.

Pain palliation: Patient response to treatment was measured by using the Brief Pain Inventory (BPI), a validated numeric scale for the evaluation of pain in cancer patients [42]. In this evaluation, patients were asked to rate their worst pain with allowed responses ranging typically from a score of 0 to 10 (0 = no pain, 10 = pain as bad as you can imagine). Two different names were used for this 11-point scale: numerical rating scale (NRS), and visual analog scale (VAS). Geiger et al. [17] also used the terminology VAS, but for a 1–10 range. Finally, Li et al. [33] used a 4-point pain scale (0–3) named verbal rating scale (VRS).

Another assessment of palliative therapy is derived from pre- and post-procedural pain as defined by the use analgesics [40]. In order to quantify its use, a morphine equivalent daily dose (MEDD) intake form was used in some studies [43]. In the case of metastatic tumours, both pain scales and MEDD followed the standards published by the International Bone Metastases Consensus Working Party (IBMCWP) [44]. This consensus defines a minimum follow-up of 2 weeks, 1 month, and then monthly until 6 months after delivery of FUS treatment. Longer follow-up is considered for patients with prolonged survival. Complete response is defined as a pain score of zero (VAS or NRS) at the treated site with no concomitant increase in analgesic intake (stable or reducing analgesics in daily oral morphine equivalents). Partial response is defined as any of the following: 1) pain reduction of 2 or more at the treated site on a 0–10 scale without analgesic increase, 2) analgesic reduction of 25% or more from baseline without an increase in pain. Pain progression is defined as an increase in the pain score of 2 or more points above baseline at the treated site with stable analgesic use, or an increase of 25% or more in daily oral morphine equivalent compared to baseline with the pain score stable or 1 point above baseline [44].

One study [36] did not account for the interference of analgesics in pain, but similarly to IMBCWP guidelines, considered a minimum VAS reduction of 3 points to be clinically significant. Another study proposed its own guidelines for treatment response based on guidance from World Health Organization standard [33].

Quality of life: Quality of life (QoL) is considered an important secondary end point in the majority of clinical studies that address painful bone metastases. Pain interference with daily living is evaluated on a 0–10 scale (0 = no interference, 10 = completely interferes) with questions concerning general activity, mood, ability to walk and work normally, relations with other people, sleep, and enjoyment of life [4]. This assessment of functional interference related to pain was monitored using different questionnaires. The first is described in the BPI-QoL [42] and the second [Quality of Life Questionaire (QLQ-BM22)] was developed by the European Organization for Research and Treatment of Cancer [45].

Results

The summary of all clinical studies meeting the above criteria is presented in Tables 2–4 for primary benign, primary malignant and metastatic bone tumours, respectively. Based on the search criteria, 15 clinical studies were selected, which included both USgFUS and MRgFUS treatments. The average treatment time was 2 h, ranging from 2 min [35] up to 18 h [19]. The procedure time was in part dependent on the type of anaesthesia or conscious sedation, but also on tumour size and blood supply. The number of treatments per site varied: the majority being a single treatment and occasionally two treatments for both benign tumours and metastatic tumours. Primary malignant tumours required up to seven sessions [19,33,46]. Details on treatment planning of image-guided FUS are outside the scope of this review, and can be found elsewhere for MRgFUS [16] and USgFUS [33].

Table 2. Clinical studies of image-guided FUS for primary benign bone tumours.

Study	Patients	Primary tumour	End points	Device + guidance	Follow-up	Outcome assessment	Outcome (a, average; m, months)
Napoli, 2014 [48]  Geiger, 2014 [17]	29	Osteoid osteoma	Palliation, Efficacy, Safety	ExAblate+ MR	1, 6, 12, 24 months	IBMCWP guidelines, Imaging	93% CR, 7% PP (12–24 m) aVAS = 7.9 → 0.7 (12–24 m) 0% AE (24 m)
Arrigoni, 2014 [49]	12	Fibro-osteitis (7) Periosteal chondroma (2) Others (3)	Palliation, Efficacy	ExAblate+ MR	6, 12 months	VAS, Imaging	92% OR, 8% NR, 0% PP (12 m) aVAS = 7.8 → 2.1 (12 m) AE: 66% minor, 0% major (12 m)
Bazzocchi, 2014 [50]	7	Osteoid osteoma	Palliation, Safety, QoL	ExAblate+ MR	1,3, 6, 12 months	VAS, QoL	86% CR, 14% PR, 0% PP (1–12 m) aVAS = 7.5 → 0.3 (1–12 m) 0% AE (1–12 m)

AE, adverse event; CR, complete response; IBMCWP, International Bone Metastasis Consensus Working Party; MR, magnetic resonance; QoL, Quality of Life; OR, overall response; PP, pain progression; PR, partial response; NR, no response; RR, recurrence; VAS, Visual Analogue Scale.

Table 3. Clinical studies of image-guided FUS for primary malignant bone tumours.

Study	Patients	Primary tumour	End points	Device + guidance	Follow-up	Outcome assessment	Outcome (a, average; m, months; y, years)
Wu, 2004 [34]	44	Malignant (44) Stage IIb (34) Stage IIIb (10)	Palliation (stage IIIb), Limb conservation (stage IIb), Safety, Survival	JC Haifu + US	10–39 months	Kaplan-Meier method, Biochemical analysis, Amputation rate, AE	11% RR (led to amputations) AE: 45% minor, 8% major (1–12 m) SR: 85% (10–39 m)
Li, 2010 [33]	13	Osteosarcoma (12) Fibrous histiocytoma (1)	Palliation, Safety, Survival	JC Haifu + US	1, 4, 6 weeks, 4, 6 months	VRS, AE, Imaging, Biochemical analysis	85% OR, 8% RR (4–6 m) 46% CR, 39% PR (4–6 m) aVRS 1.85 → 0.08 (4–6 m) AE: 68% minor, 0% major (4–6 m) SR: 100% (1 y), 85% (2 y), 69% (3 y), 39% (5 y)
Chen, 2010 [19]	80	Osteosarcoma (63) Other sarcomas (13) Others (4)	Efficacy, Safety, Survival	JC Haifu + US	1,2,3,4,5 years	Kaplan-Meier method, Imaging, Blood analysis, AE	86% CR, 14% NR (2–8 w) 6% RR (2 y), 12% RR (5 y) AE: 100% minor, 28% major (5 y) SR: 90% (1 y), 72% (2 y), 61% (3 y), 51% (4 y) and 51% (5 y)
Wang, 2013 [46]	11	Osteosarcoma (4) Ewing’s sarcoma (2) Chondrosarcoma (2) Others (3)	Palliation, Local tumour control	JC Haifu + US	1,3, 6, every 6 months	Imaging	87% OR (1 d), 0% RR and 0% PP at the time of publication AE: 100% minor, 0% major

AE, adverse event; CR, complete response; NR, no response; OR, overall response; PP, pain progression; PR, partial response; RR, recurrence; SR, survival rate; US, ultrasound; VRS, Verbal Rating Scale.

Table 4. Clinical studies of image-guided FUS for painful bone metastases.

Study	Patients (lesions)	Primary tumour	Endpoints	Device + guidance	Follow-up	Outcome assessment	Outcome (a, average VAS; m, months; y, years)
Liberman, 2009 [21]	31 (32)	Breast (11) Kidneys (6) Prostate (5) Others (9)	Palliation, Safety	ExAblate + MR	3 days, 2 weeks, 1, 3, 6 months	IBMCWP guidelines	72% OR, 4% PP (3 m) 36% CR, 36% PR, 24% NR (3 m) aVAS = 5.9 → 1.8 (3 m) AE: 0% (3 m)
Li, 2010 [33]	12	Liver (5) Lung (4) Kidney (1) Others (2)	Palliation, Safety, Survival	JC Haifu + US	1, 4, 6 weeks, 4, 6 months	VRS, Imaging, Biochemical analysis	75% OR, 8% PP (4–6 m) 42% CR, 33% PR (4–6 m) aVRS = 1.75 → 0.17 (4–6 m) SR: 83% (1 y), 17% (2 y), 0% (3 y) AE: 68% minor, 0% major (4–6 m)
Napoli, 2013 [37]	18	Lung (5) Breast (4) Kidney (3) Others (6)	Palliation, Local tumour control	ExAblate + MR	1,3, 6 months	IBMCWP guidelines, BPI, MDA criteria	89% OR (3 m) 72% CR, 17% PR, 11% PP (3 m) aVAS = 7.1 → 1.0 (3 m) AE: 0% (3 m)
Hurwitz, 2014 [43]	112	Breast (34) Lung (17) Prostate (15) Others (44)	Palliation, Safety, QoL	ExAblate + MR	2, 5, 7, 14, 30, 60, 180 days	IBMCWP guidelines, BPI	64% OR, 23% CR (3 m) aNRS reduced 3.6 points (3 m) AE: 51% minor, 4% major (3 m) 60% of AEs solved in 1 d
Huisman, 2014 [51]	11 (12)	Kidney (3) Colorectal (3) Breast (2) Others (3)	Safety, Efficacy, Technical feasibility	Sonnalleve + MR	3 days, 1 month	IBMCWP guidelines	67% OR (1 m) 11% CR, 56% PR, 0% PP (1 m) aVAS = 8 → 4 (1 m) AE: 18% minor, 0% major (1 m)
Meyer, 2014 [53]	104	Subset from  Hurwitz et al, 2014 [43]	Efficacy, Safety	ExAblate + MR	3 months	IBMCWP guidelines, BPI	Responders, AE, Sonication pain: 68%, 45%, 28% (FUS) 71%, 57%, 50% (FUS + chemo)
Zaccagna, 2014 [54]	72 (87)	Not provided	Palliation, QoL	ExAblate + MR	1, 3, 6 months	IBMCWP guidelines, QLQ-BM22	47% CR, 40% PR, 13% PP (6 m) aVAS = 6 → 2 (6 m) AE: 0% (6 m)
Bazzocchi, 2014 [55]	39 (57)	Breast (15) Kidney (5) Others (18)	Palliation, Safety, QoL	ExAblate + MR	1, 3, 6, 12 months	VAS AE	aVAS = 5.2 → 1.2 (12 m) VAS = 0 (1 m) in 31% patients AE: 5% minor, 0% major (12 m)

AE, adverse event; BPI, Brief Pain Inventory; CR, complete response; IBMCWP, International Bone Metastasis Consensus Working Party; MDA criteria, MD Anderson criteria; MR, magnetic resonance; NR, no response; NRS, Numeric Rating Scale; OR, overall response; PP, pain progression; PR, partial response; QLQ, Quality of Life Questionaire; QoL, Quality of Life; RR, recurrence; SR, survival rate; US, ultrasound; VAS, Visual Analogue Scale; VRS, Verbal Rating Scale.

Primary benign bone tumours

In 2013, Napoli et al. reported on the use of MRgFUS to treat six patients with limited joint function and reduced quality of life due to painful osteoid lesions [47]. These patients were included in a recent 24-month follow-up, which included a total of 29 patients with osteoid osteoma [17,48]. Patients received therapy using MRgFUS, delivered toward the nidus, identified on MRI and/or CT. The treatment was well tolerated and no adverse events were recorded with follow-up up to 24 months. Complete clinical response occurred in 27/29 patients as defined by absence of pain and no intake of analgesics. VAS decreased dramatically in these patients, from 7.9 at baseline to 0.7 on average at the last follow-up (12-24 months). Two patients reported pain recurrence with average VAS = 5 requiring subsequent CT-guided radiofrequency ablation. Imaging evaluation with CE-MRI demonstrated oedema and hyperaemia decrease in every lesion associated with complete response. At CT, bone remodelling was evident in all complete responders, and nidus fading was demonstrated in 15/27 patients. No complications were observed.

Arrigoni and his colleagues [49] treated 12 painful epiphyseal benign bone lesions with MRgFUS. After the treatment, 11 patients had pain regression with a mean VAS that decreased from 7.8 to 2.1 on average, 12 months after treatment. A patient with periosteal chondroma did not experience improvement. With 12-month follow-up, eight patients did not show any signs of oedema on MRI. No substantial changes were found in CT images, but in three cases they observed a recovery of the normal morphological structure of bone. No major complications were observed.

From March 2013 to May 2014, seven consecutive patients with superficial osteoid osteomas of the lower limb were treated with MRgFUS [50]. The mean VAS at the baseline was 7.5. In all patients VAS dropped to 0 after 1 month. In six patients VAS remained at 0 during the follow-up, while in 1 patient VAS dropped from 9 to 0 after 1 month, but rose to 2 after 3 months (6-month control available, no recurrence documented). No adverse events were observed.

Primary malignant bone tumours

The first FUS treatment performed in bone was in China, in a tibial osteosarcoma on December 1997. Through October 2001, a total of 1038 patients with solid tumours had undergone extracorporeal USgFUS ablation in 10 Chinese hospitals; 153 were bone tumours and 44 were reported [34]. Among them, FUS was performed as a limb-salvage treatment in combination with neoadjuvant chemotherapy in 34 patients (Enneking’s stage IIb). The remaining 10 patients were stage IIIb (nine patients with lung metastasis) and were treated with FUS alone with palliative intent. Histopathological examination demonstrated clear evidence of tumour destruction and regrowth of normal bone in the treated region. Follow-up diagnostic imaging revealed that there was no, or reduced, blood supply, and no uptake of radioisotope in the FUS-treated tumour, both indicating a positive therapeutic response and absence of viable tumour (no response rate available). Furthermore, both CE-MRI and bone scans indicated complete coagulation necrosis of the treated tumours. At 6–12-month follow-up, imaging showed obvious regression of the lesion and the region of induced coagulation necrosis. Most frequently, the non-enhancing treated volume decreased by less than 20–50% in volume. The follow-up range was 10–40 months, and the overall survival rate was 85% (38/44). One patient with stage IIb disease, and five patients with stage IIIb disease died as a result of distant metastases. Five patients underwent amputation due to local recurrence. Few major complications were observed during the follow-up period (8%), including pathological fractures in three patients, peripheral nerve damage in two, restricted joint movement in one, and epiphyseal separation in another one.

Li et al. evaluated 13 patients diagnosed with primary bone tumours, primarily osteosarcomas, treated with USgFUS [33]. At a 6-month follow-up PET/CT and SPECT showed no abnormal radioactivity concentrations in the tumour areas; the areas became cold lesions of the size and shape of the original bone tumour. The pain was quantified on a scale ranging from 0–3, and decreased, on average, from 1.85 at baseline to 0.08 at the follow-up. The response rate for patients with primary bone tumours was significant: six (46%) patients had CR, five (38%) had PR, one (8%) had moderate response, and one (8%) had PD. The overall response rate was 85%. The 1-, 2-, 3-, and 5-year overall survival rates were 100%, 85%, 69%, and 39%, respectively. No abnormalities in electrocardiogram, liver and kidney functions or blood electrolytes were detected. After two months, alkaline phosphatase and lactic acid dehydrogenase returned to normal levels [33].

In 2010, Chen et al. reported long-term follow-up results of a non-randomised clinical trial [19]. The group used USgFUS for the treatment of 80 primary malignant tumours, including 60 stage IIb and 20 stage IIIb (Enneking staging) patients. FUS combined with chemotherapy was performed in 62 patients with osteosarcoma, one with periosteal osteosarcoma and three with Ewing’s sarcoma. The remaining 14 patients with chondrosarcoma, malignant giant cell tumour of bone, sarcoma of the periosteum or with unknown histology received FUS alone. Follow-up images demonstrated complete ablation of the tumours in 69 patients and greater than 50% tumour ablation in the remaining 11 patients. The overall 5-year survival rate was 51%, with 64% and 16% for patients with stage IIb and III disease, respectively. Among the patients with stage IIb disease, long-term survival rates were substantially improved in 30 patients that received combined treatment with FUS and chemotherapy (86% 5-year survival rate), in comparison with the survival rates for 24 patients who did not finish chemotherapy and six patients who underwent partial ablation only (36% 5-year survival rate). Only five of the 69 patients who underwent complete ablation had local cancer recurrence during the follow-up period (5–87 months). All patients experienced mild pain and among other adverse events, 28% were major complications [39] where 11 of these patients required surgery and eight presented severe peripheral nerve damage. In these cases the distance between the damaged nerves and the tumour margin was less than 10 mm, suggesting that 10 mm is a reasonable safety margin to avoid nerve damage in FUS treatments.

The most recent USgFUS malignant bone tumour treatment series reported in the literature was in 2013, by Wang et al. [46]. Eleven patients with primary malignant tumours of the bony pelvis received USgFUS for both palliative (seven patients) and curative purposes (four patients). The efficacy of FUS ablation was only assessed by CE-MRI. With median follow-up of 22 months (11–154 months), seven patients who received palliative ablation died of metastatic disease. Enlargement of residual tumour was observed in all patients receiving palliative FUS ablations 6–24 months after study entry. The remaining four patients who received ablation with curative intent were alive at the time of the publication. Local recurrence was observed in one patient receiving curative FUS ablation, which was retreated by additional FUS ablation; no local recurrence was observed thereafter. Significant coagulative necrosis was obtained in all patients, with an average volume ablation ratio of 87% (range 65–100%). Complete tumour necrosis was achieved in all patients receiving curative FUS ablation. No major complications were encountered.

Metastatic bone tumours

Liberman et al. in 2009 published the first multicentre clinical study on the use of FUS for pain palliation of bone metastases [21]. This report includes previous pioneer work from Catane et al. [18] and Gianfelice et al. [20], comprising 31 patients and 32 bone lesions; 3-month follow-up was available for 25 out of 31 patients. Eighteen patients (72%) had a significant reduction in pain (>2 points) and nine (36%) reported a VAS score of 0. The average VAS score decreased from 5.9 prior to treatment to 1.8 at the 3-month follow-up; with 52% of patients reporting substantial pain relief within 3 days; 24% had no response and one patient (4%) experienced worsened pain levels. A reduction in opioid usage was reported in 67% of patients with recorded medication data. No major complications were reported.

Li et al. also reported on 12 patients with bone metastases treated with USgFUS [33]; 96% of the patients suffered pain before the procedure. Pain decreased from 1.75 at baseline to 0.17 on a 0–3 scale at the last follow-up (4–6 months). Response rates were classified by MRI or PET/CT and not pathologically: five (42%) had complete response, four (33%) partial response, one (8%) had progressive disease, with the overall response rate as 75%. The 1-, 2-, 3-, and 5-year overall survival rates were 83%, 17%, 0%, and 0%, respectively.

Napoli et al., in 2013, reported a prospective, single-arm research study with 18 patients treated with MRgFUS for painful bone metastases [37]. The pain severity score changed significantly from a baseline average of 7.1 to 1.1 at 3-month follow-up. A score of 0 for pain severity, without medication intake, was reported by 13/18 patients (72%) at final follow-up, consistent with a complete response to treatment. CT examinations demonstrated increased bone density with restoration of cortical borders in 5/18 patients (28%). According to the MDA criteria [41], a complete response to treatment was observed in 2/18 patients (11%), a partial response in 4/18 patients (22%), stable disease in 10/18 patients (56 %) and progressive disease in 2/18 patients (11%). No treatment-related adverse events were recorded during the study.

The results of the first phase III clinical trial on bone tumours were published in 2014 by Hurwitz et al. [43]: 147 patients with metastatic bone pain, typically refractory to radiation and other pain interventions, were randomised to MRgFUS treatment or placebo treatment. Patients randomised to placebo underwent the same procedure as those receiving MRgFUS treatment but without energy deposition. The pain response rates 3 months after treatment were 64% in the MRgFUS-treated arm vs. 20% in the placebo arm. Complete pain relief was observed in 23% of treated patients, compared to 6% of patients who received placebo treatment. Approximately two-thirds of responders experienced significant pain relief – as defined as a decrease in worst NRS score by 2 points or more – within 3 days of treatment, establishing the ability of MRgFUS to induce fast pain response. This response was accompanied by a similarly rapid improvement in patient function scores. The most common complication was pain during MRgFUS treatment (32%) and major complications (third-degree skin burn, fracture) occurred in 4% of treated patients. However, one fracture was outside the treated area, and the skin burn was due to a violation on the inclusion criteria protocol. Furthermore, the majority of adverse events (60%) were transient and resolved on the treatment day, and 51 patients (46%) had no adverse events.

In 2014, Huisman and colleagues reported the first experience with volumetric MRgFUS for palliative treatment of painful bone metastases [51]. In this technique the focal spot is iteratively switched to well-separated positions around a circular trajectory and sonications are repeated continuously until ablative temperatures are achieved throughout increasing diameter concentric circles. The resulting volumetric temperature rise contrasts with more traditional raster scan iterative sonications where each overlapping focal ablation is followed by a cooling period. The goal is improvement in ablation zone homogeneity, energy efficiency, and overall treatment time [52]. Pain inventories were implemented at 3 days and 1 month after treatment for 11 patients. At 3 days after volumetric MR-FUS ablation, NRS pain scores decreased significantly from 8 to 6 on average, and a response was observed in 6/11 patients (55%). At 1-month follow-up, nine evaluable patients had NRS pain scores decreased significantly compared to baseline, from 8 to 4 on average, and six out of nine patients obtained pain response. The overall response rate was 67% with 0% pain recurrence at 1-month follow-up. No treatment-related major complications were observed.

The phase III trial as reported by Hurwitz et al. was subject recently to a retrospective analysis of the safety of combination MRgFUS with active systemic chemotherapy [53]. Chemotherapy data were available for 104 patients. Patients were followed for 3 months. Ninety patients were treated without chemotherapy, and 14 were treated with chemotherapy. There was no significant difference between the response rates of the chemotherapy group (71%) and the non-chemotherapy group (68%). The overall adverse event rates were 57% for chemotherapy patients and 45% for non-chemotherapy patients. Sonication pain was not significantly different between the groups, with 50% pain in the chemotherapy group and 28% pain in the non-chemotherapy group. Remaining adverse event rates were not significantly different (p = 0.17).

A prospective, single arm, multicentre study was performed to evaluate the efficacy of MRgFUS for pain palliation of bone metastasis in patients who had exhausted radiotherapy or refused other therapeutic options [54]. A total of 72 patients with painful bone metastases were enrolled; 34/72 patients (47%) reported complete response to treatment and discontinued medications, 29/72 patients (40%) experienced a pain score reduction >2 points, consistent with partial response. The remaining nine patients (13%) had recurrence after treatment. Significant differences between baseline (6) and follow-up (2) average VAS values and medication intake were observed. Similarly, a significant difference was found for QLQ-BM22 between baseline and follow-up (p < 0.05). No treatment-related adverse events were recorded.

Finally, Bazzocchi et al. evaluated the clinical outcome of 39 patients (57 lesions) with painful bone metastases that were treated with MRgFUS [55]. Nine patients had a single bone metastasis, while five showed no other distant metastasis. The follow-up schedule included 1-, 3-, 6-, and 12-month evaluations; 45 lesions were evaluated after 1 month, while 31 lesions reached the 3-month (54%), 17 the 6-month (30%) and eight the 12-month (14%) follow-up points. Four patients died during follow-up and three lesions required retreatment. On a lesion-based approach, average VAS score at baseline was 5.2 decreasing to 2.5 at 1 month, and to 2.0, 2.1, and 1.2 after 3, 6 and 12 months respectively. In 14/45 lesions (31%) the VAS dropped to 0 1 month after the treatment; VAS persisted at the 0 level in eight patients up to 3 months, in five patients up to 6 months, and in three patients up to 12 months. The major determinant of MRgFUS success was lesion size, with smaller lesions corresponding to higher efficacy in terms of pain relief – possibly due to more efficient tumour debulking. Two treatment-related adverse events were reported: a single case of small skin burn, and one case of prostate inflammation in a patient treated to the ichiopubic ramus.

Discussion

The use of image-guided FUS has increased significantly in the past decade. The ability to deliver treatments accurately and non-invasively is a great advantage of image-guided FUS for the treatment of bone tumours. Contrary to ionising radiation, image-guided FUS treatments can be repeated without the risk of cumulative dose effects in normal tissues outside the target. Also, thermal dose may be confined precisely to a small treatment volume (typically 0.2–20 mm³) without the need for invasive implants such as microwave, radiofrequency, laser, and cryoablation techniques (Table 1). Furthermore, the extent of treatment is readily controlled in three dimensions, which provides safe, non-invasive conformal treatments that spare healthy tissue.

Efficacy

Standardisation of validated assessment instruments facilitates comparisons of clinical studies. In this respect, the IBMCWP guidelines are seeing increasing use for evaluation of palliative treatments, which allows direct comparison of treatment outcomes. However, not all studies followed the same criteria, making comparison more challenging. For instance, Geiger et al. (2014) used a 10-point pain scale (1–10), Li et al. (2010) a 4-point pain scale (0–3), while the remaining studies used the IBMCWP 11-point pain scale (0–10). In the case of curative treatments, there are greater challenges comparing different studies since the number of patients varied widely and different end points were reported such as imaging, biochemical analyses, and survival rates. Taking into account overall response inclusive of complete and partial responses as a baseline for comparisons, the responders group varied within 92–100%, 85–87% and 64–87% for primary benign, primary malignant, and metastatic tumours, respectively. In treatments with a curative aim, the recurrence rate was 0–14%, and in palliative treatments the pain progression was 0–13%. These results demonstrate the efficacy of FUS for both palliative and curative purposes in the treatment of bone tumours.

Safety

Similar to response, challenges exist in assessing safety across studies. Different studies used different classifications for adverse events. Therefore, for summary purposes we defined adverse events according to minor and major complications [39]. Overall, studies of benign tumours reported 0–66% minor and 0% major complications. Patients treated for primary malignant tumours presented higher complication ranges: 45–100% minor and 0–28% major complications. Finally, patients with metastatic tumours presented complication ranges similar to the first group: 0–51% minor and 0–4% major complications. The most frequently observed complications were mild skin burn – usually resolving by 1–2 weeks after FUS – and sonication pain during treatment. Several authors inferred that skin burns were most likely due to lack of operator experience or not following established protocols. The lack of adverse events in some studies [17,37,47,48,50,54,55] may relate to the limited number of patients. In clinical studies with a large number of patients, one USgFUS study [19] reported a major complication rate of 28% (22/80) for patients with primary malignant tumours. This was in large part due to lack of guidelines to protect normal structures, e.g. there were eight serious nerve injuries as treatment guidelines did not limit proximity to nerves. Other primary malignant studies implemented a 1-cm tumour margin limit and major complications were limited to 8%. On the other hand, MRgFUS produced only 4% (4/112) major complications in patients with bone metastases. Overall, the data provides strong evidence that FUS is safe for treatment of bone metastases and primary benign lesions, while other applications are still under investigation.

Sonication pain was the most common complaint during conscious sedation or epidural anaesthesia. General anaesthesia provides an advantage in terms of alleviation of sonication pain. Also it ensures intraoperative immobilisation which facilitates ultrasound targeting. However, the use of anaesthesia does not allow patient feedback on pain, and that could risk serious normal tissue damage [33]. As such, the optimisation of patient anaesthesia relative to complication rates requires further investigation, especially in treatment sites near critical tissue structures.

US vs. MR guidance

Both MR and US imaging have been used to guide FUS in the treatment of bone tumours. Both imaging methods have advantages and disadvantages. Ultrasound imaging devices are less expensive, widely available, capable of real-time visualisation, and clinically proven in the treatment of organs such as liver and kidney which move with respiration. The main disadvantage of US is poorer imaging resolution than MRI, especially in areas that have air or bone. Furthermore, nerves cannot be visualised by ultrasound imaging, and are thus difficult to avoid in the beam path. This is especially relevant because nerves are sensitive to ultrasonic energy and tumours are often located adjacent to nerves.

MRI can provide three-dimensional imaging with better resolution, which allows accurate tumour delineation. Also, MR allows continuous monitoring of temperature distribution within the treatment zone and thus measures both normal and target tissue temperature rise during FUS exposure. Since rapid protein denaturation occurs above 60 °C, sonications lead quickly to coagulative necrosis which is readily imaged by MR. The combination of these features – high resolution pre-treatment target visualisation, MR thermometry during treatment to spare adjacent healthy tissue, and immediate post-treatment verification of effective treatment zone – makes MR guidance a highly attractive component of focused ultrasound therapy.

Despite the advantages, MR adds high cost, long treatment time, and problems in tracking moving targets, e.g. ribs, which may limit application of MRgFUS. MR temperature imaging is limited primarily to the soft tissues adjacent to bone, since the PRFS method does not work well for calculating temperature rise in low water-content bone cortex, bone marrow and fat tissue. However, the thermal feedback received is sufficient to allow control of actual sonication location and minimisation of damage to non-target tissues [21]. This approach affords the treating physician a ‘closed loop’ monitoring of the treatment in real time, resulting in a high level of safety and efficacy [18]. Also, the reduced major complications associated with MRgFUS will reduce the overall hospitalisation time. Thus, although more expensive initially, MRgFUS can decrease overall patient care cost [17]. In any case, direct comparison of safety and efficacy between MRgFUS and USgFUS is not possible since USgFUS was used mainly for primary malignant tumours whereas MRgFUS was used for benign and metastatic tumours. The only study on bone metastases was a subset from Li et al. [33] which did not show significant differences in terms of adverse events and response rates.

Future directions

The success of FUS ablation in the clinical studies summarised above demonstrates the promise of image-guided FUS and suggests further investigation is needed in larger numbers of patients with primary and secondary bone tumours. Three clinical trials are currently recruiting to evaluate MRgFUS in pain palliation of bone metastases: NCT01833806, NCT01964677 and NCT01586273. Other aspects could be addressed such as the correlation between site of metastases or primary tumour type and treatment outcome, local effect on the progression of bone metastases, long-term durability of pain palliation, and long-term durability of bone strength. Furthermore, the low procedure morbidity and short treatment time suggests that MRgFUS may also be a viable treatment option for patients with multifocal bone metastases. As pointed out by Gianfelice and his colleagues [20], treatment for multifocal disease can be repeated with no obvious increase in treatment morbidity. Expansion of indications for FUS inclusive of spinal treatments whether invasive [56] or non-invasive will significantly expand the number of patients who may benefit from this treatment modality. Finally, additional randomised studies for malignant lesions using multimodality approaches including FUS are warranted.

Conclusions

Focused ultrasound is a treatment modality with both emerging and established roles in treatment of bone tumours. While most research has focused on MR guidance, both MR and US guidance have been studied with promising results. As phase III trials are yet to be completed for treatment of primary tumours, further study is warranted, particularly for primary malignant tumours, before FUS can be considered standard of care across all clinical applications related to bone lesions. Comprehensive review of the studies to date, however, indicates that use of FUS to treat primary benign tumours, primary malignant tumours, and metastatic tumours in bone is both safe and effective.

Declaration of interest

Mark D. Hurwitz has done consulting work for InSightec; David Vrba acknowledges funding from the EU project COST action TD1301 and the Czech Scientific Foundation, project 13-29857P. The authors alone are responsible for the content and writing of the paper.