Osteonecrosis of the Jaw Associated with Antiangiogenics in Antiresorptive-Naïve Patient: A Comprehensive Review of the Literature

Objectives To review the available literature on medication-related osteonecrosis of the jaw (MRONJ) associated with antiangiogenics in antiresorptive-naïve individuals. Methods A literature search was performed using MEDLINE via PubMed, EMBASE, and Web of Science in December 2017. Results We identified reports describing a total of 35 antiresorptive drugs-naïve patients who developed antiangiogenic-related MRONJ. The mean age of these patients was 59.06 years and the F : M ratio was 4 : 5. The most common underlying disease was metastatic renal cell cancer. Pain to the mandible was the most common complaint (34.29%) and the majority of patients presented with bone exposure. The mean duration of intravenous and oral antiangiogenics before MRONJ development was 6.5 and 16.72 months, respectively. The most common additional risk factor was dental extraction (37.14%). Almost half of the MRONJ patients (48.57%) received surgical treatment. 18 patients (62.06%) were reported to have disease resolution within an average time of 6.75 months. Conclusion MRONJ associated with antiangiogenic therapy in antiresorptive-naïve patients is a rare but potentially serious adverse effect. Available data suggests that there might be notable differences between MRONJ associated with antiangiogenics and antiresorptives; however, further prospective well-designed studies are required.


Introduction
Medication-related osteonecrosis of the jaw (MRONJ) is an uncommon and potentially serious adverse side effect of antiresorptive and antiangiogenic agents [1]. It can cause chronic pain, infection, dysfunction, and disfigurement and can affect the quality of life of affected individuals [2,3]. The vast majority of cases of MRONJ are associated with antiresorptive agents including bisphosphonates, denosumab, and more recently romosozumab [4][5][6][7]. A notably smaller number of cases are associated with the use of antiangiogenic agents, both in individuals who also take antiresorptive drugs and in those who are antiresorptive drugs-naïve [8]. MRONJ can develop in approximately 7% of cancer patients taking high-potency bisphosphonates or high-dose denosumab and about 0.01-0.1% of those with osteoporosis using low-potency oral bisphosphonates or lowdosage denosumab [1,[9][10][11][12]. The use of antiangiogenic agents in combination with antiresorptive drugs is known to increase the risk of MRONJ development [13]; however, little is known regarding the incidence and prevalence of antiangiogenic-related MRONJ in antiresorptive drugs-naïve individuals. Antiangiogenic inhibitors have been increasingly used in the management of a range of malignancies including ovarian cancer, metastatic renal cell cancer, breast cancer, colorectal cancer, non-small-cell lung cancer (NSCLC), and glioblastoma multiforme [14]. Antiangiogenic inhibitors can be categorised into three major groups based on their mechanism of action: anti-VEGF monoclonal antibody (e.g., bevacizumab), VEGF decoy receptors or VEGF-Trap (e.g., 2 BioMed Research International aflibercept), and small molecule tyrosine kinase inhibitors (TKI) that block the VEGF receptors downstream signaling pathways (e.g., sunitinib, cabozantinib, and sorafenib) [15] ( Table 1). Additionally, the mammalian target of rapamycin (mTOR) inhibitors also seems to have antiangiogenic effects by inhibiting the production of VEGF and platelet-derived growth factors (PDGF) [16][17][18].
The number of patients developing MRONJ associated with antiangiogenic inhibitors or a combination of antiangiogenics and antiresorptive drugs has been growing over the last few years [8,13,19,20]. The purpose of the present study is to provide a comprehensive review of the published reports of MRONJ associated with antiangiogenic agents in patients with no history of antiresorptive therapies.

Literature Search Strategy.
A literature search was conducted to identify clinical trials, case reports, and case series on MRONJ associated with antiangiogenic treatment in antiresorptive drugs-naïve individuals using MEDLINE via PubMed (up to December 2017), EMBASE (from 1980 to December 2017), and Web of Science (from 1900 to December 2017). The search strategy used the following keywords: "osteonecrosis," "jaw osteonecrosis," "jaw bone necrosis," "oral osteonecrosis," "antiangiogenic," "angiogenesis inhibitors," "antineoplastic agents," "antiangiogenic activity," "antiangiogenic therapy," "chemotherapy," and "targeted therapy." The references of retrieved articles were manually searched in order to identify additional relevant articles and abstracts. The search included articles published in English and other languages. Inclusion criteria were patients developing MRONJ associated with antiangiogenic agents based on the definition of MRONJ proposed by the special committee on MRONJ of the American Association of Oral and Maxillofacial Surgeons (AAOMS) in 2014 [1]. Patients with history of radiotherapy involving the jaw bones and patients having previous history or concurrent use of antiresorptive therapy were excluded.
Regarding the management of MRONJ, seventeen patients (48.57%) were managed with surgical procedures alone or combined with medications (antibiotic therapy, antimicrobial mouthwash) and interruption of antiangiogenic agents. 16   one patient. Most surgical interventions (11 patients) were minimally invasive procedures including smoothening of exposed bone, local flap coverage, removal of superficial necrotic bone, soft tissue debridement, and bone curettage, whereas 6 patients underwent major surgery such as bone decortication, resection of necrotic bone with local flap coverage, segmental osteotomy, and block resection. The outcomes of therapy were reported for 29 patients (82.8%), whereas no information was provided for the other 6 cases. 18 patients out of these 29 (62%) were reported to have disease resolution, while 11 patients showed persistent bone exposure. Disease resolution was described as complete mucosal coverage/no evidence of exposed bone in 13 cases, whereas no clear description was provided for the remaining 5 cases. Of note, one patient who experienced disease resolution to the left side of the mandible eventually developed a new area of MRONJ to the right mandible. The mean time from MRONJ diagnosis to complete healing was 6.75 months (range months: 1.84-22; SD = 2.47; 95% CI: 0.90-12.59). The data of 35 reported cases with MRONJ associated with antiangiogenics are summarised in Table 3.

Discussion
The present study is the first comprehensive review upon MRONJ in patients treated with antiangiogenics in the absence of bone-modulating therapy. We present data from 35 patients with different metastatic cancers who developed MRONJ following antiangiogenic treatments. All individuals were antiresorptive drugs-naïve. We have identified a number of differences between MRONJ associated with antiangiogenic agents and MRONJ associated with antiresorptive drugs. Our data showed a sex ratio of 4 : 5 (F : M) and an age range of 33-80 years (mean: 59.06 years), compared to sex ratio of 3 : 2 and age range of 42-90 years (mean 66 years) reported for antiresorptive drugs-associated MRONJ [50][51][52]. There also seem to be differences in the prevalence of MRONJ in these two populations. The reported prevalence of MRONJ in patients who had been treated with intravenous bevacizumab alone for the treatment of advanced breast cancer was 0.2%, which was lower than that of MRONJ associated with intravenous antiresorptive agents (7%) [11,19]. However, it is important to note that the prevalence of MRONJ related to antiangiogenic agents may also depend on the epidemiology of underlying malignancies that require antiangiogenic therapy.
The clinical presentations of MRONJ associated with antiangiogenics also seem to be different from MRONJ due to antiresorptive agents. Approximately up to 25% of MRONJ cases related to antiresorptive medications can present without frank bone exposure [53], whereas most of the patients in the present review had clear evidence of bone exposure (91.43%). However, the number of patients with nonexposed MRONJ might be underestimated, since until 2014 MRONJ could only be diagnosed in individuals with clinical evidence of exposed bone as per AAOMS definition [53,54].
With respect to the presenting complaints and location, they appear to be similar in two populations. The majority of patients with antiangiogenic-related MRONJ in this study experienced pain to the jaw, which is also the most common complaint in patients with MRONJ associated with antiresorptive agents [51,55]. In the present study, most MRONJ cases associated with antiangiogenic therapy tended to occur in the mandible more frequently than in the maxilla, similar to those with antiresorptive drugs-induced MRONJ [50,52].
A number of additional risk factors were identified in the present review including dental extraction, the use of denture, periodontal infection, and dental implant. Almost 40% of reported cases in this study were predominantly preceded by tooth extraction, which is similar to those with antiresorptive drugs-related MRONJ [50,56].
There is a slight difference with respect to underlying malignancies between two populations. Patients with antiresorptive drugs-associated MRONJ showed the previous history of multiple myeloma, metastatic breast cancer, and metastatic prostate cancer, whereas those with MRONJ related to antiangiogenic medications in our review were mainly diagnosed with metastatic renal cell cancer, followed by metastatic colorectal cancer and metastatic breast cancer as demonstrated in Table 3 [52].
Although there was no consistent pattern in the time to MRONJ development in this review, the average time for developing MRONJ among patients with either intravenous or oral antiangiogenics was shorter than the average time to       MRONJ onset in those treated with antiresorptive drugs. The mean time to event for intravenous and oral antiangiogenic agents in this study was 6.5 and 16.71 months, respectively, while it was reported to be approximately 1.8 and 3 years for bisphosphonate therapy [12,57]. Patients with metastatic malignancy may receive a number of anticancer drugs simultaneously. In this review, we found that seven of the reported cases received more than one antiangiogenic agent in their treatment history [32,[44][45][46][47][48][49]. Of these patients, some were given different antiangiogenics at the same time, while others received these agents at different time points. The development of MRONJ is usually associated with the latest antiangiogenic agent used by the patient; however, one cannot exclude the fact that the antiangiogenic agents previously used by these patients might have contributed to it.
We included in this review two cases of MRONJ associated with new TKIs, namely, pazopanib in combination with axitinib ( = 1) and dasatinib ( = 1) [32,49]. According to the Food and Drug Administration's Adverse Event Reporting System (FAERS), pazopanib and axitinib have been associated with the development of MRONJ in 10 and 9 individuals, respectively; however, as data regarding concurrent or previous medication were not available in FAERS documentation [58], it is difficult to conclude whether these individuals were indeed antiresorptive drugs-naïve. Therefore, we decided not to include these 19 cases in our review.
With regard to the management of MRONJ, approximately half of the individuals with MRONJ associated with antiangiogenics (48.57%) were managed surgically, which is similar to those with bisphosphonate-related MRONJ [11,52].
However, the prognosis of antiangiogenic-related MRONJ appears to be better than that of individuals developing MRONJ associated with antiresorptive agents. We observed a 62% rate of disease resolution in those where outcomes were reported as opposed to approximately 50% reported in the literature for MRONJ associated with antiresorptive agents [56,59,60]. It is possible that the higher rate of disease resolution might be related to the shorter half-life of antiangiogenics [61,62], as well as the lower cumulative dosages [63]. Moreover, the average time to resolution for MRONJ associated with antiangiogenics appears to be shorter than antiresorptive drugs-induced MRONJ (6.75 months, range: 1.84-22 months versus 8.2 months, range: 0.2-25.6 months) [55].
In this comprehensive review, we excluded a number of potential antiangiogenic-related MRONJ cases due to a lack of adequate clinical information. For example, the 2012 report on aflibercept by the US Food and Drug Administration (FDA) described 3 aflibercept-treated bisphosphonate-naïve patients who developed MRONJ; however, none of these patients were added to the present review as one had jaw bone exposure for less than 8 weeks and no information was provided for the other two cases [64]. Furthermore, in a pivotal BOLERO-2 trial, MRONJ has been described in 2 patients in the experimental arm (everolimus-exemestane) and 1 patient in the control arm (exemestane), with one of three patients to receive bisphosphonate treatment [65]. However, there was no evidence to show whether the patient with a history of bisphosphonate treatment was in the experimental arm or control arm. More recently, Antonuzzo et al. [66] reported the first case with MRONJ potentially associated with regorafenib, one of the tyrosine kinase inhibitors, in an antiresorptive drugs-naïve individual. Although MRONJ appeared 22 months after regorafenib treatment, Fusco et al. [67] have noted that some details such as the use of other medications prior to regorafenib treatment, dosing, and the time on medication are still missing. This medication is usually used as a third or further line of treatment of metastatic colorectal cancer. Therefore, it is also worth knowing whether this patient has received other well-documented antiangiogenic medications such as bevacizumab and aflibercept prior to regorafenib. If this is the case, bevacizumab or aflibercept possibly might contribute to the development of MRONJ rather than regorafenib alone. Another patient with gastrointestinal stromal tumours (GISTs) receiving imatinib monotherapy presented with pain and exposed bone at lower right mandible after having the tooth removed for 5 weeks. The patient was treated with debridement of necrotic bone and antibiotic and then was discharged. Unfortunately, there was no further information about this patient [68]. The duration of persistent bone exposure in this case was not mentioned if it was longer than 8 weeks. Therefore, available data seems not to be enough to classify this case as MRONJ according to the definition of MRONJ [1] and to confirm the association between MRONJ and imatinib. In addition to the above reported cases, Hopp et al. [69] reported one patient with necrotic bone exposure after the 2-year intravitreal injections of bevacizumab for treatment of retinal vascular thrombosis without notable dental risk factors or use of bisphosphonates. After the patient experienced pain to the mandible, the lesion was completely healed by antibiotics treatment within 8 weeks. Therefore, this case seems not to be correlated with the definition of MRONJ formulated by the AAOMS in 2014 [1].
It is important to highlight that another case of oral soft tissues complication associated with bevacizumab was also reported by Magremanne et al. [70]. Although this case was included in previous reviews regarding cases of MRONJ associated with antiangiogenic agents, there was no evidence of osteonecrosis of the jaw and also the necrotic area seemed to be limited only to oral soft tissues. This reported case does not meet the definition of MRONJ and hence it was excluded from this review.

Conclusion
There remains incomplete information regarding the cases of antiangiogenic-related MRONJ in antiresorptive-naïve individuals reported in the literature. It is therefore difficult to draw any conclusion regarding the epidemiology and the characteristics of MRONJ in this patient population. Within the limitation of available data, we have identified a number of differences between MRONJ associated with antiangiogenics and MRONJ related to antiresorptive drugs including demographic characteristics, prevalence, the underlying malignant disease, time to the onset, and time to resolution. Considering that the list of antiangiogenic inhibitors that have potential to increase the risk of MRONJ development is increasing, further prospective and well-designed research is warranted to confirm our findings and increase knowledge and understanding of the disease.

Conflicts of Interest
The authors have no conflicts of interest to declare.