Avoiding Severe Toxicity From Combined BRAF Inhibitor and Radiation Treatment: Consensus Guidelines from the Eastern Cooperative Oncology Group (ECOG)

Abstract

BRAF kinase gene V600 point mutations drive approximately 40% to 50% of all melanomas, and BRAF inhibitors (BRAFi) have been found to significantly improve survival outcomes. Although radiation therapy (RT) provides effective symptom palliation, there is a lack of toxicity and efficacy data when RT is combined with BRAFi, including vemurafenib and dabrafenib. This literature review provides a detailed analysis of potential increased dermatologic, pulmonary, neurologic, hepatic, esophageal, and bowel toxicity from the combination of BRAFi and RT for melanoma patients described in 27 publications. Despite 7 publications noting potential intracranial neurotoxicity, the rates of radionecrosis and hemorrhage from whole brain RT (WBRT), stereotactic radiosurgery (SRS), or both do not appear increased with concurrent or sequential administration of BRAFis. Almost all grade 3 dermatitis reactions occurred when RT and BRAFi were administered concurrently. Painful, disfiguring nondermatitis cutaneous reactions have been described from concurrent or sequential RT and BRAFi administration, which improved with topical steroids and time. Visceral toxicity has been reported with RT and BRAFi, with deaths possibly related to bowel perforation and liver hemorrhage. Increased severity of radiation pneumonitis with BRAFi is rare, but more concerning was a potentially related fatal pulmonary hemorrhage. Conversely, encouraging reports have described patients with leptomeningeal spread and unresectable lymphadenopathy rendered disease free from combined RT and BRAFi. Based on our review, the authors recommend holding BRAFi and/or MEK inhibitors ≥3 days before and after fractionated RT and ≥1 day before and after SRS. No fatal reactions have been described with a dose <4 Gy per fraction, and time off systemic treatment should be minimized. Future prospective data will serve to refine these recommendations.

Introduction

The BRAF kinase gene V600 point mutations drive approximately 40% to 50% of all melanomas, with recent profiling of human tumors revealing a role in papillary thyroid cancer (30%-80%), anaplastic thyroid cancer (25%), pediatric astrocytoma (10%-20%), colon cancer (8%), and non–small cell lung cancer (5%) (1). This mutation is associated with decreased locoregional control and survival and with resistance to radiation therapy (RT) (1, 2). BRAF inhibitors (BRAFi) improve progression-free survival (PFS) and overall survival (OS) in patients with melanoma bearing either V600E and V600K mutations (3), and there is promise in other cancer histologies as well (2, 4). Although highly successful in achieving tumor responses in BRAF V600 mutant metastatic melanoma (approximately 50%), the PFS remains, on average, 6 to 7 months with BRAFi such as vemurafenib (5, 6) and dabrafenib (7). The MEK inhibitors (MEKi) trametinib and cobimetinib have further improved outcomes when added to dabrafenib and vemurafenib, respectively (7, 8), with median PFS extended to 10 to 11 months. RT may provide symptomatic relief in up to 84% of patients (9, 10). Approximately 50% to 97% of patients experience partial response (PR) or complete response (CR) of the radiated lesion, with CR rates ranging from 17% to 69%. Although many patients may discontinue their BRAFi at the time of disease progression, a significant minority (up to 45%) may experience progression in a few areas despite an overall significant response (11). For this scenario, often defined as oligoprogressive disease, a strategy may be to treat progressive or symptomatic areas with RT or surgery before resuming the systemic treatment that has provided overall clinical benefit. Preliminary data suggest improved outcomes with this approach, with OS increased in 1 series to more than 9.1 months from symptom onset for those resuming vemurafenib after a local therapy versus 3.4 months for those who did not (11). However, prospective trials leading to BRAFi and MEKi approval excluded RT, resulting in a lack of data on toxicity and efficacy when combined. There are data regarding dermatologic and visceral toxicity for both cytotoxic agents (eg, doxorubicin) and for targeted agents–for example, cetuximab (12), erlotinib, and sorafenib (13)–when used in combination with RT. It is unclear whether BRAFi should be held before, during, and after RT and, if so, how long. Even less is known about MEKi and RT interactions, although recent data suggest in vitro and in vivo radiosensitization from the combination (14, 15).

To identify publications describing outcomes from RT with BRAFi, MEKi, or both, PubMed.org was searched for all in vitro and in vivo data published in any language detailing any observed effect from the combination approach. Only primary publications were incorporated in this review. Three additional unpublished cases of toxicity encountered by the authors were also included. Clinically, there have been reports of increased dermatologic (16-33), lung (20), liver (16), esophageal (22, 34), brain (26, 35), and bowel toxicity (26) when RT has been given concurrently with or in proximity to BRAFi, including both vemurafenib and dabrafenib. Severe dermatitis has been reported during RT when given concurrently with a BRAFi, and it has also been described as an RT recall reaction despite starting a BRAFi many weeks after RT completion (18, 20, 21, 23, 24, 27, 28, 33, 36, 37). RT dermatitis can often be managed effectively with topical or systemic steroids and analgesics as needed, sometimes without BRAFi cessation. Organ damage has been reported months from RT completion (16, 20, 26). Multiple publications indicate that whole brain RT (WBRT) (21, 22, 24, 27, 31, 32, 38, 39) and stereotactic radiation surgery (SRS) (22, 26, 35, 38, 40, 41) are safe with BRAFis. However, various organ toxicities can be painful and disfiguring and have a significant impact on quality of life, and they may even be fatal. Until prospective data are available, we must rely on retrospective data to guide our practice. With the inclusion of BRAFis in clinical trials for malignancies that could eventually require palliative RT (eg, ClinicalTrials.gov identifiers: NCT02224781 & NCT02164916), the need for consistency and widespread knowledge regarding safe practice guidelines is essential.

Models Demonstrating BRAFi Radiosensitization

Inhibition of BRAF has been associated with radiosensitization in vitro (1, 42). Sambade et al (42) found that for V600E mutant melanoma cell lines, PLX4032 (vemurafenib) increased cell cycle arrest in G1 through inhibition of the MAPK/Erk signal transduction pathway. This is a relatively radiosensitive portion of the cell cycle, and this may have decreased the repair of potentially lethal DNA damage and increased RT induced apoptosis. This effect was present only for BRAF V600E-mutated cells with an RT enhancement ratio of 10, compared with a ratio of 1 (ie, no enhancement) for BRAF wild-type cells. Dasgupta and colleagues (1) found an additive relationship between RT and the BRAFi PLX4720 (a specific BRAF V600E inhibitor) for BRAF V600E-mutated melanoma and colon cancer cell lines. Similar to Sambade et al (42), no increased effect was found for BRAF WT cells. Both of these studies hypothesized that combined BRAFi and RT could provide a therapeutic advantage against BRAF-mutated cancers.

Hecht et al (22) quantified radiosensitization by assaying peripheral lymphocytes from 35 patients treated with dabrafenib and/or vemurafenib. Chromosomal aberrations were quantified as mean breaks per metaphase (B/M value). Scores indicating dramatically increased radiosensitivity were present in 4 of 8 vemurafenib patients, 1 of 9 dabrafenib patients, and none of the control group, with a significant difference found between the vemurafenib and dabrafenib groups (P=.004). Patients with average B/M values had no skin toxicities, whereas patients with increased B/M values had more frequent acute and late ≥2 dermatitis. Of note, the only grade 3 dermatitis occurred in a patient receiving dabrafenib who had previously been treated with vemurafenib. Although this is intriguing, the assayed lymphocytes are expected to be BRAF WT, and the in vitro data from both Dasgupta et al (1) and Sambade et al (42) found that only BRAF mutated cells were radiosensitized by BRAFi. Therefore, these data regarding vemurafenib as being more radiosensitizing are hypothesis generating, and they need validation in alternative, larger patient cohorts.

Dermatologic Toxicity

Various systemic agents have been implicated in increasing the dermatologic toxicity from RT, including cytotoxic agents such as actinomycin D, anthracyclines, taxanes, 5-FU, hydroxyurea, methotrexate, and cisplatin (13), and the targeted therapies cetuximab (12), sorafenib (43, 44), and erlotinib (45). RT enhancement is defined as dermatitis exacerbation resulting from a systemic agent started during or within 7 days of RT completion, whereas for RT recall the agent is started more than 7 days after RT (46). Table 1 details 57 cases of dermatologic RT enhancement (16, 17, 22-32, 34, 47, 48) and 6 cases of skin RT recall (18, 21, 33, 36, 37) associated with BRAFi. Typically, RT-induced dermatitis occurs at least 10 to 14 days after the start of RT. Severe RT enhancement reactions occurred as quickly as 3 days after the beginning of RT (31). For the recall reactions, vemurafenib was started a median of 3.5 weeks after RT (range, 3 weeks to 3 months), with dermatitis noted a median of 2 weeks (range, 1-4 weeks) after BRAFi initiation.

Table 1. Extracranial toxicity with combined BRAF inhibitors and radiation therapy.

Row #	Study	Toxicity	N	CTCAE grade	BRAFi	BRAFi → RT	Concurrent BRAFi
1	Anker et al (16)	Hepatic hemorrhage & dermatitis	1	5 & 3	Vem	1 mo	N (4 day washout)
2	Baroudjian et al (17)	Hemothorax & dermatitis	1	5 & 2	Vem	3 mo	Y
3	Forschner et al (20)	Pneumonitis	1	3	Vem	NA	N
4	Forschner et al (20)	Pneumonitis	1	2	Vem	NA	N
5	Current study	Bowel perforation	1	4	Dab/Tram	NA	N
6	Peuvrel et al (26)	Anorectitis/diarrhea & dermatitis	1	3 & 2	Vem	NS	Y
7	Merten et al (34)║	Esophagitis & dermatitis	1	3 & 2	Vem	NS	Y
8	Hecht et al (22)	Esophagitis	1	NS	Vem	NS	Y
9	Hecht et al (22)	Dermatitis	8	3	Vem or Dab	NS	Y
			23	2
		FCP/CVG*	11	NA
		Other dermatologic†	3	NS
		Miscellaneous nondermatologic/non-CNS toxicities‡	5	NS
10	Wallach et al (32)	Mucositis/dermatitis	1	3	Vem	Same d	Y
11	Schulze et al (31)	Dermatitis	1	3	Vem	1 wk	Y
12	Saco and Mitchell (29)	Dermatitis	1	3	Vem	3 wk	Y
13	Pulvirenti et al (27)	Dermatitis	1	3	Vem	Same d	Y
14	Churilla et al (19)	Dermatitis	1	3	Vem	2 wk	Y
15	Satzger et al (30)	Dermatitis	1	3	Dab	4 mo	Y
16	Satzger et al (30)	Dermatitis	1	3	Dab	7 mo	Y
17	Satzger et al (30)	Dermatitis	1	2	Dab	8 mo	Y
18	Satzger et al (30)	Dermatitis	1	2	Vem	5 mo	Y
19	Ducassou et al (47)	Dermatitis	1	2	Vem	6 mo	Y
20	Conen et al (37)	Dermatitis	1	2	Vem	NA	N
21	Pulvirenti et al (27)	Dermatitis	1	2	Dab	NS	Y
22	Pulvirenti et al (27)	Dermatitis	1	2	Dab	2 mo	Y
23	Pulvirenti et al (27)	Dermatitis	1	2	Vem	Same d	Y
24	Hurmuz et al (48)	Dermatitis	1	2	Vem	1 day	Y
25	Houriet et al (23)	Dermatitis	1	2	Vem	NA	N
26	Forschner et al (20)	Dermatitis	1	2	Vem	NA	N
27	Boussemart et al (18)	Dermatitis	1	2	Vem	NA	N
28	Boussemart et al (18)	Dermatitis	1	2	Vem	NA	N
29	Levy et al (25)	Dermatitis	1	NS	NS	NA	N
30	Current study	Dermatitis with CVG	1	2	Vem	NA	N
31	Lang et al (24)	Dermatitis with CVG	1	2	Vem	NA	N
32	Harding et al (21)	Dermatitis with CVG	1	2	Vem	Same d	Y
33	Harding et al (21)	Dermatitis with CVG	1	2	Vem	NA	N
34	Current study	Dermatitis with in-field FCP	1	2	Vem	NA	Y
35	Pulvirenti et al (27)§	Dermatitis with in-field FCP	1	2	Vem	NA	Y (last 5 fractions)
36	Reigneau et al (28)	Dermatitis with in-field FCP	1	2	Vem	2 wk	Y
37	Braunstein et al (36)	Dermatitis with in-field & out-field FCP	1	3	Vem	NA	N
38	Wang et al (33)	Dermatitis with in-field & out-field FCP	1	2	Vem	NA	N
39	Houriet et al (23)	FCP	1	NA	Vem	NA	N
40	Schulze et al (31)	FCP	1	NA	Vem	1 mo	Y
41	Total¶		89
Abbreviations: BRAFi = BRAF inhibitor; RT = radiation therapy; BRAFi→RT = start of BRAF inhibitor to start of RT; CTCAE = Common Toxicity Criteria for Adverse Events version 4; CVG = cutaneous verrucous gyrata; FCP = follicular cystic proliferation; L = lumbar vertebral body; T = thoracic vertebral body; Mo = months; n = number of cases; NA = not applicable; NS = not specified; RT → BRAFi = time from RT completion to start of BRAF inhibitor; SCV = supraclavicular; Dab = dabrafenib; Tram = trametinib; Vem = vemurafenib; Washout = time from BRAFi end to RT start.
*Including one case of CVG.
† 1 case of hand-foot syndrome in radiated area, 1 case of impaired wound healing, and 1 case of hyperpigmentation.
‡ 3 cases of hearing difficulty related to external auditory canal toxicity, 1 case of polyneuropathy, 1 case of taste disorder.
§ Only toxicity with a V600K mutation; all others involved V600E or were not specified.
¶ The 89 cases of toxicity occurred in 78 patients.
║ Based on personal communication with Dr Hecht, the esophagitis case described in Merten et al (34) is one of the 2 cases of dysphagia noted in Hecht et al (22). Personal communication also revealed unpublished details of the second esophagitis/dysphagia case in Hecht et al (22) as well as information that the 52 cases of extracranial toxicity in Hecht et al (22) and Merten et al (34) occurred in 41 patients.

Row #	RT target	RT dose (Gy/number of fractions)	Dose at toxicity (Gy)	RT → BRAFi	RT completion to toxicity	BRAFi/RT held?
1	Bone (T10-L1)	20/5	After RT	2 days	Liver: 4 mo; skin: 2 wk	BRAF
2	Nodes (axilla)	20/4	After RT	None	NS	BRAF
3	Nodes (mediastinum)	40/16 → boost 10/5	After RT	3 wk	7 wk	Neither
4	Nodes (axilla)	50/20	After RT	4 wk	7 wk	Neither
5	Ilium/femurs/L spine	20/5	After RT	10 days	2 mo	BRAF
6	Rectal 1° adenocarcinoma	30/10→ 4 wk break → 25/10	30	No break	During	RT
7	Bone (T4-T7)	T4-T7: 36/12; 8th rib: 30/10	NS	No break	Eso: NS; Skin: During RT	Neither
8	Mediastinum	59.4/33	NS	NS	NS	NS
9	Numerous sites	Brain mean: 33.6/NS; other site mean: 36/12	After RT	No break	NS	BRAF in 9% RT in 4%
10	Brain & nodes (SCV)	Neck: 50/20 & brain: 30/10	After RT	No break	1 mo	Neither
11	Brain & leptomeninges	30/10	9	No break	During	BRAF
12	Bone (femur)	30/10	15	No break	During	Neither
13	Brain	30/10	After RT	No break	1 wk	Neither
14	Bone (T10-L5)	30/10	After RT	No break	1 wk	Neither
15	Bone (T12)	36/18	After RT	No break	During	Neither
16	In-transit (leg)	60/30	34	No break	During	Neither
17	Nodes (infraclavicular)	60/30	12	No break	NA	RT
18	Nodes (inguinal)	20/10 (60/30 4.5 y prior)	After RT	No break	NS	Neither
19	Bone (T12)	30/10	After RT	No break	4d	Neither
20	Skin & Nodes (axilla)	Skin: 30/6; axilla: 35/10	After RT	3 mo	3 mo	Neither
21	Nodes (axilla)	36/12; no bolus	21	No break	During	Neither
22	Bone (2nd rib; iliac; pubic)	8/1	After RT	No break	1 mo	Neither
23	Brain	30/10	After RT	No break	1 wk	Neither
24	Bone (shoulder; iliac/femur)	8/1	After RT	No break	5 d	Neither
25	Neck/face	30/6	After RT	3 days	3 mo	Neither
26	Brain	30/10	After RT	3 wk	5 wk	Neither
27	Subcutaneous & bone (T8)	18/3	After RT	1 day	10 d	Neither
28	Bone (iliac)	20/5	After RT	23 days	7 d	Neither
29	Skin	18/3	After RT	1 wk	1 wk	Neither
30	Brain	37.5/15	After RT	4 days	4 d	BRAF
31	Brain	30/10	After RT	1 wk	4 wk	BRAF
32	Brain	30/10	21	No break	During	RT
33	Brain	35/14	After RT	3 wk	6 wk	Neither
34	Nodal basin (axilla)	22.5/5	After RT	No break	1 mo	Neither
35	Brain	30/10	After RT	No break	1 wk	Neither
36	In-transit (scalp)	30/10	After RT	No break	1 mo	BRAF
37	Nodal basin (neck)	71/38	After RT	6 wk	2 mo	BRAF
38	Nodal basin (axilla)	NS	After RT	4 wk	5 wk	Neither
39	Neck/face	30/6	After RT	3 days	3 mo	Neither
40	Brain & leptomeninges	30/10	After RT	No break	1 mo	Neither
41

Timing of RT and BRAFi

Dermatitis severity appears related to RT and BRAFi timing. All grade 3 dermatologic reactions have been described with concurrent RT and BRAFi administration (19, 22, 27, 29-32) or when starting BRAFi within 2 days of RT (16) except for 1 patient who received 71 Gy in 38 fractions to the neck nodal basin followed by vemurafenib 6 weeks later (36). This patient received the highest dose of all patients reviewed, and a grade 3 dermatitis requiring a 1-week break from RT occurred before vemurafenib was started 6 weeks later. Two weeks after starting vemurafenib, wound dehiscence occurred, prompting drug cessation and calcium alginate dressing changes. The frequency of grade 2 and 3 reactions from RT enhancement is best estimated from Hecht et al (22). For the 86 courses of RT delivered to 70 patients concurrently with BRAFi, 27% and 9% of irradiated sites had grade 2 and 3 adverse reactions, respectively. The authors found that BRAFi given concurrently with WBRT increased dermatitis rate significantly (P<.001) compared with those not receiving BRAFi, with grade 2 rates of 34% versus 7% and grade 3 of 6% versus 0%. Of note, intracranial SRS did not cause any ≥grade 2 dermatitis. With skin dose reported at <2 Gy (38), the lack of dermatitis from SRS with or without BRAFi is not surprising (26, 35, 38, 41, 49).

Dose of RT and BRAFi

Dermatitis severity may be related to total RT dose to the skin, as opposed to dose per fraction. In the report by Ducassou et al (47), adjuvant RT alone to 30 Gy in 5 fractions did not cause dermatitis. However, concurrent vemurafenib and 30 Gy in 10 fractions to spine metastases caused grade 2 dermatitis, with the skin surface dose estimated at 20 Gy. Vemurafenib was discontinued before WBRT to 30 Gy in 10 fractions, and again no dermatitis was noted. Merten et al (34) found twice the incidence of pigmentation and erythema from concurrent BRAFi and RT versus RT alone using a computer-assisted digital image evaluation for 2 separate courses of 36 Gy in 12 fractions in the same patient. For another patient treated with anteroposterior:posteroranterior (AP:PA) RT fields to spine metastases given concurrently with BRAFi, the estimated skin dose was 23 to 31 Gy on the back (grade 3 dermatitis) but only 16 to 20 Gy on the chest and abdomen (grade 1 dermatitis) (19).

In contrast to RT dose, dermatitis does not appear to be dependent on BRAFi dose, at least when given concurrently. Hecht and colleagues (22) reported nonsignificant differences in dermatitis for full-dose and reduced-dose BRAFi at 10% and 7% for grade 3 and 26% and 27% for grade 2, respectively. BRAFi doses were reduced either because of adverse events related to BRAFi therapy that occurred before RT or (in 1 center) during RT aimed at reducing concomitant treatment toxicity (personal communication with Dr Hecht, May 2015). Overall, both vemurafenib (6 of 63) and dabrafenib (2 of 23) resulted in grade 3 skin toxicity in 10% of patients, although grade 2 toxicity was insignificantly lower for dabrafenib (20 vs 29%).

Summary

BRAFi increase the risk of grade 2 and 3 dermatitis with RT. The severity of the reaction appears dependent on the dose of RT but not BRAFi, and all but 1 grade 3 dermatitis incident was reported in the setting of concurrent RT and BRAFi administration.

Nondermatitis skin toxicities

Dermatologic toxicities with both vemurafenib and dabrafenib in the absence of RT are frequently reported, including photosensitivity reactions and alopecia. Hyperproliferative skin lesions including hyperkeratosis, keratoacanthomas, and squamous cell carcinomas are thought to be the result of a paradoxical activation of the extracellular signal-regulated kinases (ERK) in BRAF wild-type cells in response to RAF kinase inhibition. It is proposed that vemurafenib activates wild-type RAS-RAF and epidermal growth factor pathways in keratinocytes (50). Thus, BRAFis induce activation, repopulation, and proliferation of keratinocytes. A hypothesis for normal tissue toxicity from RT is that RT preferentially affects proliferating and dividing cells, so more keratinocytes are likely to be killed by RT, therefore facilitating more intense skin reactions (27). BRAFi can cause alterations in keratinocyte proliferation, leading to the development of benign epithelial tumors (warty papilloma) or malignant lesions (squamous cell carcinoma or keratoacanthoma) in 25% of patients. Photosensitivity is predominantly induced by ultraviolet A rays, manifesting in 26% to 52% of patients receiving vemurafenib and only about 3% of patients receiving dabrafenib (47, 51). The vast majority of patients who experienced dermatologic and visceral toxicities did not have photosensitivity, indicating likely different mechanisms for these processes. The molecular structure of vemurafenib is different from other ultraviolet A photosensitizing drugs such as fluoroquinolones and 6-TG, and although the mechanism is only partially understood based on in vivo data, it appears less dependent on singlet oxygen than the other agents (52).

Toxicities involving follicular cystic proliferation (FCP), including keratosis pilaris syndrome, have been noted by multiple investigators (Table 1). Follicular cystic lesions have been described by multiple authors (18, 23, 28, 47). Schulze and colleagues (31) reported on a patient in whom miliary cystic lesions developed 1 month after concurrent vemurafenib and WBRT only in the RT field. Biopsy showed numerous actively proliferating epidermal cysts, considered a reactive response to vemurafenib and RT. This was still present 3 months later, with vemurafenib continued throughout. In this current review article, the authors report an unpublished FCP and underlying brisk dermatitis that occurred 1 month after RT given concurrently with vemurafenib (Fig. 1, Table 1). Dysfunctional keratinocyte proliferation was noted by Wang et al (33) in a patient who received vemurafenib 1 month after chest wall RT. Diffuse skin thickening with goosebump-like papules was noted resembling keratosis pilaris, with an asymptomatic erythematous background in the radiated area. The patient was given ammonium lactate, with no mention of its efficacy. Houriet et al (23) noted epidermal cysts not associated with dermatitis 3 months after RT that persisted at 1 year. Despite reasonable cause for concern, thus far there have been no reports of increased malignancies in the RT field.

Fig. 1.

Cystic proliferation and underlying brisk dermatitis noted over right flank and axilla 1 month after palliative radiation therapy given concurrently with vemurafenib.

Multiple authors have noted tortuous folding and convoluted furrowing of the scalp after RT and BRAFi resembling cerebral gyri and sulci, descriptively termed cutis verticis gyrata (CVG) (Table 1). The dramatic cerebriform appearance of the scalp is due to galea aponeurotica restriction on soft tissue expansion, overgrowth of the scalp, or both, and on histologic examination the appearance is variable and depends on the underlying cause. Lang et al (24) reported a patient experiencing this effect 4 weeks after starting vemurafenib, which was initiated 1 week after WBRT completion to 30 Gy in 10 fractions. This reaction is the most severe CVG case reported; significant thickening of the scalp and face resulted in hearing impairment. Vemurafenib was stopped, but intravenous (IV) methylprednisone resolved only the dermatitis. Examination of a biopsy specimen revealed only a small amount of inflammatory infiltrate, but multiple follicular retention cysts and milia were present. Owing to the lack of inflammation, IV steroids were stopped. Treatment was changed to ipilimumab, and the gyrate-like skin appearance resolved over several weeks. Of note, these findings extended adjacent to the RT field over the cheeks and chin. Harding et al (21) reported on 2 patients with painful CVG. After only 21 Gy, 1 patient experienced a severe confluent erythematous and hyperkeratotic plaque with tortuous skin folding and thickening limited to the RT field, prompting an RT break. Vemurafenib was not stopped, and he was treated symptomatically with topical emollients. Subsequent topical corticosteroids, retinoids, and antibiotics afforded slow improvement. Two months after RT, aclometasone was started, and the lesions resolved 1 month later. The second patient had a similar but less pronounced reaction despite the passage of 3 weeks between WBRT and vemurafenib initiation. Three weeks after starting the BRAFi, a grade 1 RT recall dermatitis included furrowing and ridging of the scalp. Vemurafenib was continued, but it was not mentioned whether the administered salicylic acid was beneficial. In a previously unpublished case of the current authors (Fig. 2, Table 1), vemurafenib was started 4 days after WBRT to 37.5 Gy in 15 fractions; 4 days after beginning the BRAFi a grade 2 dermatitis occurred concurrently with CVG changes. Examination of a biopsy specimen showed follicular hyperkeratosis and syringosquamous metaplasia. Topical clobetasol was started and the BRAFi continued. Significant improvement was noted after 2 months, but complete resolution required almost 5 months.

Fig. 2.

Images of cutis verticis gyrata appearance obtained 2 weeks after starting vemurafenib; the BRAFi was started 4 days after whole brain radiation therapy. (A) Lateral view with biopsy site circled, showing follicular hyperkeratosis and syringosquamous metaplasia (B) Posterior view.

Summary

RT and BRAFi may cause painful and disfiguring non-dermatitis reactions, the most common of which are FCP and CVG.

Treatment

Treatment of dermatologic toxicity, including dermatitis, hyperproliferative cystic reaction, or both, involves predominately topical agents (Table 2). The mainstay of treatment involves barrier creams and topical corticosteroids, although silvadene may be considered for moist desquamation (19). In addition to discomfort, the disfiguration associated with proliferative skin reactions (eg, CVG) can cause psychological stress, and in some cases the lesions may not resolve (23, 31). In severe cases, either RT (21) or vemurafenib (31) have been held (Table 2), although most reports involve continuation of both therapies (22). Salicylic acid (21) and ammonium lactate (33) have been mentioned in the treatment for hyperkeratosis reactions, but no benefit for either was described. Many skin reactions are painful, so analgesics may be required.

Table 2. Treatment of BRAFi-related and radiation therapy—related toxicities.

Toxicity	Treatment*	Expected outcome
Dermatitis	Dry desquamation: barrier creams (eg, Aquaphor, Calmoseptine, Desitin, Balneol); topical steroids† (20, 29)	Resolution in wk (18) to 1-2 mo (27, 37)
	Consider urea cream (31)
	Moist desquamation: silvadene cream (19)
CVG	Recommended: topical steroids† (21, 24)	Resolution in wk (24) to 5 mo (Current)
	(IV steroids unnecessary) (24)
	Consider: retinoids & antibiotics (21)
FCP	Ulcers: calcium alginate dressings (36)	Resolution in wk (28) to mo (31); with some cases
	Folliculocentric eruptions: topical steroids† (36)	unresolved for beyond 1 y (23)
Pneumonitis Anorectitis‡	Prednisolone with or without IV antibiotics (20)	Prompt improvement in symptoms (20)
	Oral prednisone (26)	Slow improvement over mo (26)
	Discontinue BRAFi (26)
	If refractory: consider colostomy (26)
Mucositis/esophagitis‡	Supportive care & TPN if needed (34)	Slow improvement over mo (34)
	Discontinue BRAFi (34)

Abbreviations: BRAFi = BRAF inhibitor; CVG = cutaneous verrucous gyrata; FCP = follicular cystic proliferation; IV = intravenous; TPN = total parenteral nutrition.

Numbers in parentheses indicate references.

^*Analgesics should be considered for all patients as needed.

^†Examples of topical steroids include betamethasone (18), aclometasone (21), and triamcinolone (36).

^‡All reported grade 3 gastrointestinal tract toxicity including mucositis, esophagitis, and anorectitis occurred with concurrent BRAFi and radiation therapy.

Mucosal Toxicity

Mucosal toxicity, occasionally severe, has been reported along the length of the gastrointestinal (GI) tract involving the pharynx (32), esophagus (22, 34), and rectum (26) (Table 1). In 1 case, after receiving vemurafenib concurrently with RT to the neck and whole brain (32), grade 3 dermatitis developed 1 month later, involving yellow-white papules and crusting lesions over all irradiated skin extending into the oral cavity. The oral cavity received at most 12 Gy, a dose not expected to cause such severe toxicity (personal communication with Dr Fox, January 2016). Despite continuation of vemurafenib, over the next several weeks the patient's oral mucosal and skin toxicities improved significantly, with no specific treatment described. Merten et al (34) reported a patient who received RT for spine metastases concurrently with vemurafenib, resulting in grade 2 dermatitis and grade 3 esophagitis. Hospitalization for total parenteral nutrition was required, and solid foods were not tolerated for 2 months (Tables 1 & 2). The retrospectively determined mean esophageal dose of 31 Gy was lower than a commonly accepted dose limit of 34 Gy for RT courses given over about 6 weeks (53). However, because the patient's RT was hypofractionated over 2 weeks, the corresponding delivered biological equivalent dose (BED) of 39 Gy₁₀ was actually higher than the BED of 38 Gy₁₀ that corresponds to the 34 Gy physical dose limit. Therefore, RT alone, not vemurafenib, may be the cause of this toxicity. Hecht et al (22) reported a dysphagia rate of only 2% (2 patients) in their series that included RT to the spine (22), 1 of which was the case reported by Merten et al (personal communication with Dr Hecht, May 2015). Although it is not specified how many people were at risk for this toxicity, to decrease the chance of GI toxicity no patient was treated with an AP:PA technique. Peuvrel et al (26) described a patient treated with hypofractionated palliative RT to a primary rectal adenocarcinoma concurrently with vemurafenib given for metastatic melanoma. The resulting toxicity included grade 3 anorectitis and diarrhea, with severe pain recalcitrant to morphine and IV steroids. Because of persistent problems, a colostomy was required 10 months after RT, at which time vemurafenib was held, allowing gradual recovery (Table 2).

In this review we also present a previously unpublished case in which 20 Gy in 5 fractions was delivered to painful pelvic bone melanoma metastases, followed 10 days later by dabrafenib and trametinib. One month later, CT imaging to evaluate severe abdominal pain showed free air but no clear perforation. Intraoperatively, significant inflammation and a pencil-sized perforation was found, and pathologic examination showed no association with a diverticulum or tumor. The surgeon noted the sigmoid colon draped in the pelvis corresponding to the region of the patient's RT fields (Figs. 3A, 3B, and 3C). Colonic perforations have been reported 1 to 2 months after palliative RT (30 Gy in 10 fractions) given concurrently with sorafenib (43, 44). The similar timing of the toxicity in the present case raises further concern that vemurafenib after RT was implicated. However, a causal relationship between RT and BRAFi/MEKi cannot be confirmed, and although the bowel dose was within tolerance (54) spontaneous perforations can occur even without RT.

Fig. 3.

(A) Axial, (B) sagittal, and (C) coronal views of radiation plan for patient with bowel perforation 1 month after dabrafenib and trametinib were started. These systemic agents were started 10 days after 20 Gy in 5 fractions using a 4 field technique with 10 MV photons to pelvic bone metastases (shaded red). The V_{18 Gy} and V_{20 Gy} of the large bowel loops (green), small bowel loops (orange), and bowel bag volume (not shown) were 180 and 90 cc, 460 and 230 cc, and 1450 and 750 cc, respectively. (A color version of this figure is available at www.redjournal.org.)

Summary

The increased risk of mucosal toxicity from BRAFi with RT beyond that expected with RT alone appears low, but the GI organs should not intentionally be targeted with RT during BRAFi therapy; rather, their concurrent use should be minimized.

Hearing Loss

After WBRT, a patient described by Wallach et al (32) had hearing loss presumptively resulting from erythematous scaly plaques extending into the ear canals. Hecht et al (22) described ototoxicity in 3 of the 32 patients receiving WBRT, with hearing reduced from middle ear effusion in 2 patients and the third with complete hearing loss of unknown cause (personal communication with Dr. Hecht, May 2015).

Summary

These data do not raise increased concern specifically for ototoxicity with concurrent BRAFi administration, but there may be an increased risk for obstructive-type hearing disturbances from dermatologic toxicity within the auditory canal.

Lung Toxicity

RT-recall pneumonitis (RRP) has been reported with the use of cytotoxic therapies such as methotrexate and paclitaxel (55, 56). This phenomenon may occur from RT and BRAFi interactions according to data from Forschner et al (20), who noted pneumonitis in 2 patients treated with vemurafenib after RT (Table 1). The presenting symptoms were cough for 1 patient and shortness of breath leading to hospitalization for the second, with prompt symptomatic improvement in both patients after oral steroids were given. The risk of RRP based on published data was <5% for patient 1 (V₂₀ = 12.4%, D_mean=6.9 Gy) and 15% for patient 2 (V₂₀ = 32.9%, D_mean=17.4) (57), so it is possible these cases were unrelated to vemurafenib. Of note, the authors saw no pulmonary symptoms in 5 other patients treated with axillary RT followed by vemurafenib. Of concern is a report of severe pleural toxicity after 20 Gy in 4 consecutive fractions was given concurrently with vemurafenib for right axillary lymphadenopathy (17). The patient experienced grade 3 dermatitis followed by CR at 1 month, but a hemothorax leading to death 1 month later raises suspicion of a severe hemorrhagic pulmonary/pleural toxicity. Although a second patient had no toxicity despite a higher dose of 30 Gy in 6 daily fractions to the pleural surface, the risk of hemorrhage should be noted.

Summary

Although the likelihood of RRP, pleural hemorrhage, or both is low, vigilance in detecting symptoms of RRP (cough, fever, shortness of breath, and chest pain) is recommended for at-risk patients. Prompt administration of corticosteroids may prevent the need for cessation or dose reduction of BRAFi (Table 2) (20).

Hepatic Toxicity

Vemurafenib alone has been associated with hepatic toxicity involving transaminase elevations (3). A fatal hepatic hemorrhage was reported by Anker and colleagues (16) for a patient who received 20 Gy in 5 fractions through AP:PA fields to T10-L1 for spine metastases (Table 1). The retrospectively determined mean liver dose was only 2.7 Gy, far below the typical dose limit of 31 Gy (58). Vemurafenib was held 4 days before and 2 days after RT. Computed tomographic (CT) imaging showed hepatic injury in the radiated field 4 months after treatment. The liver is typically located along an axial plane coinciding with the T8-L2 vertebral bodies, and other reports involving AP:PA RT with concurrent vemurafenib or dabrafenib to that region have not shown evidence of hepatoxicity (19, 30, 47). There have likely been many additional similarly treated patients without consequence, indicating a probable undetermined idiosyncratic feature for the patient who experienced the fatal liver toxicity. Hecht et al noted no hepatotoxicity despite RT given concurrently with vemurafenib or dabrafenib for multiple patients receiving RT to the spine, although liver dose was minimized through the use of posterior oblique RT fields (personal communication with Dr. Hecht, May 2015).

Summary

Although the probability of hepatotoxicity with RT and BRAFi appears very low because only 1 case has been reported, the consequences may be severe, and care to minimize liver dose is recommended.

Intracranial Neurologic Toxicity

Although there is evidence of intracranial antitumor activity for BRAFi alone, median CNS progression-free survival is only 3 to 6 months (59, 60). The risk of neurologic symptomatic intracranial toxicity from SRS and BRAFi appears low in the 111 patients who received the combination (22, 26, 35, 38-41, 49), with an incidence of hemorrhage and necrosis similar to patients who had SRS without BRAFi (61) (Table 3). SRS is effective with high rates of local control (35, 39-41, 61, 62), and in many cases it obviates the need for WBRT (35, 63). The main concern for SRS toxicity involves necrosis and hemorrhage. Although hemorrhage is frequently associated with melanoma metastases, Ghia et al (64) found no influence of SRS on the risk of hemorrhage. In a report of 52 patients with known BRAF mutation status, Ly et al (35) identified 17 patients treated with BRAFi with a washout period initiated before and after SRS (median, 7 days; range, 1-20 days). At a median follow-up time of 10.5 months, no patient had radionecrosis. Of note, BRAFi treatment for patients with BRAF mutant melanoma was associated with a decreased rate of freedom from hemorrhage at 1 year: 77.0% versus 39.3% (P=.0003). However, despite this difference, OS was not significantly different between patients who did and did not receive BRAFi. The leading cause of neurologic death was hemorrhage, but of the 15 patients who died of hemorrhage, the majority (60%) never received a BRAFi (personal communication with Dr David Ly, May 2015). These data indicate that the increased risk of hemorrhage noted did not significantly correlate with an increased mortality risk. Likewise, other studies involving SRS and BRAFi did not find an increase in clinically significant hemorrhage. Out of 80 lesions treated in 24 patients with BRAFi held 2 to 3 days before and after BRAFi, Ahmed et al (40) reported only 1 episode of hemorrhage that led to a craniotomy 2 months after SRS. In another series, of the 26 patients who received SRS concurrently with a BRAFi, only 1 had a symptomatic hemorrhagic intracranial event. Gaudy-Marqueste et al reported no toxicities from the combination despite most patients (20 of 30) receiving concurrent BRAFi and SRS (41). In a separate study, 2 patients who received SRS to 24 Gy concurrently with BRAFi (1 vemurafenib, 1 dabrafenib) did not show evidence of necrosis or hemorrhage at magnetic resonance imaging (MRI) 3 months after SRS (38).

Table 3. Central nervous system toxicity with combined BRAF inhibitors and sterotactic radiosurgery.

Row #	Study	N (total no. of patients/total no. of lesions)	Follow-up (mo)	Median size (cc)	Median dose (Gy)/no. of fractions	WBRT before/after SRS	BRAFi & SRS timing
1	Liew et al (61)	333/1570	3.8	1.4	18/1	33%/NS	NA
2	Ahmed et al (40)	24/80	5.1	0.28	24/1	None	Held 2-3 d before & after
3	Gaudy-Marqueste et al (41)	30/53	4.8	NS	20-28/1	NS	Concurrent = 20; break = 6; after RT = 4
4	Ly et al (35)	17/96	10.5	NS	20/1	0/33%	Held median 7 d; range 1-20 before & after
5	Narayana et al (39)	SRS alone: 6/14 SRS + PBRT: 1/3 PBRT: 2/2 WBRT: 3/29	12.2	1.7	SRS: 20/1 PBRT/WBRT: 30/15	None as salvage	7 concurrent; 5 BRAF started median 8.7 wks after
6	Peuvrel et al (26)	1/2	9	NS	20/1	NA	Concurrent; Vem started 3 mo before
7	Liebner et al (49)	1/3	NS	2.2	22/1 & 24/4	NA	Vem started 1 wk after
8	Liebner et al (49)	1/1	NS	2.7	Surgery → 30/5	NA	Vem 3 mo before; 2 wk after
9	Rompoti et al (38)	2/2	≥3	NS	24/1	NA	Concurrent
10	Hecht et al (22)	27/NS	NS	NS	NS	NA	Concurrent
Abbreviations: BRAFi = BRAF inhibitor; CNS = central nervous system; CTCAE = Common Toxicity Criteria Adverse Events version 4; Dab = dabrafenib; DMFS = distant brain metastasis-free survival; FFH = freedom from hemorrhage; Hem = hemorrhage; Gr = grade; LC = local control; NA = not applicable; NS = not specified; OS = overall survival; PBRT = partial brain radiation therapy; RT = radiation therapy; SRS = stereotactic radiosurgery; Unk = Unknown; Vem = vemurafenib; WBRT = whole brain radiation therapy.

Row #	BRAFi	LC (%, 6/12 mo)	DMFS (%, 6/12 mo)	OS (%, 6/12 mo)	Toxicity	CTCAE grade
1	None given	81/63	57/33	47/25	Hem: 5% Gr 4/1% Gr 5; 25% radiologically	4/5
2	Vem	92/75	45/23	61/38	Hem: 4% patients	4
3	Vem (87%)/Dab (13%)	NS	Median 12.9 wks	NS	Hem: 20% lesions	3
4	Vem (41%)/Dab (53%)/ Unk (6%)	90/85	NS/33	NS/50.2	1-yr FFH: 39.3%; deaths: 4 (25%)	5
5	Vem	75/60	57/NS (median 14.5 mo)	92/55	Hem: 0% after SRS; necrosis vs tumor progression 1 mo after RT for 1 patient	3
6	Vem	NS	NS	NS	Radionecrosis; dexamethasone dependent	3
7	Vem	NS	NS	NS	Radionecrosis 7 wks after RT	4
8	Vem	NS	NS	NS	Radionecrosis 4 mo after RT	4
9	Vem	NS	NS	NS	No CNS toxicity	NA
10	Vem or Dab	NS	NS	NS	No CNS toxicity	NA

There have been several reports of toxicity with BRAFi and SRS, but none of them definitively attributes toxicity to the combination. Narayana et al (39) reported on 6 patients who received vemurafenib before or after SRS alone. One patient receiving SRS followed by vemurafenib 1 month later experienced symptomatic weakness requiring steroids 4 months after SRS. MRI was suggestive of radionecrosis, although tumor progression could not be excluded. Peuvrel et al (26) described SRS to 20 Gy for 2 asymptomatic brain metastases given concurrently with vemurafenib started 3 months previously. Two months afterward, MRI revealed perilesional edema and radionecrosis, at which time the patient experienced severe headaches and was dependent on steroids until death. Liebner et al (49) reported on 2 patients with radionecrosis who received vemurafenib and SRS. One episode occurred after reirradiation, and the other involved a relatively high dose of 30 Gy in 5 fractions; the authors acknowledged that radionecrosis could have occurred independently of BRAFi.

Systemic agents such as methotrexate (65), gemcitabine (66), and interferon-α (67) have been associated with neurotoxicity when given in proximity to WBRT. Of 47 patients who received BRAFi with WBRT, many of whom experienced dermatitis with or without associated hyperkeratotic changes (21, 22, 24, 27, 31, 32, 38, 39), only 1 had a hemorrhagic event but it was not clearly related to combination treatment (personal communication with Dr Hecht, January 2016).

On the basis of available data, the risk of brain radionecrosis does not appear increased with BRAFi. Only 1 study suggested an increased risk of intracranial hemorrhage with BRAFi, but this did not correlate with increased mortality (35). Lower CNS neurotoxicity may in part be due to the blood-brain barrier, inasmuch as lower CNS concentrations of dabrafenib and trametinib have been noted to result from active efflux through P-glycoprotein and the breast cancer resistance protein (Brcp1) (68).

Summary

There is no conclusive evidence linking BRAFi and RT with intracranial neurotoxicity with either fractionated RT or SRS.

Enhanced RT Efficacy

Radiosensitization from RT and BRAFi has been associated with remarkable increases in treatment efficacy, not just toxicity. The most dramatic evidence involves 6 patients with unresectable disease, all treated with vemurafenib (median duration, 5.8 months) for induction of response, followed by consolidative RT (median dose 57 Gy, conventional fractionation), with 3 patients receiving debulking interval surgery (69). With 29 months' minimum follow-up, crude local control was 100%. The 3 patients who experienced relapse received successful salvage therapy to become free of disease at latest follow-up. Baroudjian et al (17) reported on a patient who had single-site progression in the axilla after a CR in in-transit metastases while taking vemurafenib. He achieved a CR based on positron emission tomography/CT with no toxicity after 30 Gy in 6 consecutive daily fractions given concurrently with vemurafenib, and this patient was still disease free over 3 months afterward.

Lee et al (70) reported on a patient with positive cerebral spinal fluid (CSF) cytology that developed after 4 months of vemurafenib started for progression of metastatic disease. This patient received WBRT to 30 Gy in 10 fractions, with vemurafenib held 7 days before and after RT. There was no reported dermatitis, and the patient had resolution of CSF involvement and associated neurologic symptoms, allowing for return to work. At latest follow-up 18 months after RT, CSF was still negative, and the patient had stable extracranial disease. Given that leptomeningeal spread is almost uniformly rapidly fatal in melanoma, this case indicates a potentially synergistic effect between RT and vemurafenib. The authors hypothesized that RT could have disrupted the blood-brain barrier, allowing greater distribution of vemurafenib in the CSF (71). Inasmuch as dabrafenib and vemurafenib can lead to intracranial response rates of approximately 40% in patients with treated or untreated brain metastases (60, 72) despite apparent limitations in penetration of the blood-brain barrier (68), it is reasonable to expect improved treatment efficacy with the combination.

However, Satzger et al (30) described 4 patients with infield progression while taking a BRAFi (3 dabrafenib and 1 vemurafenib) who experienced RT-enhanced dermatitis, all of whose tumors progressed despite concurrent RT. Inasmuch as the majority of palliative RT for melanoma induces a PR or better, it is not clear why these sites responded poorly. It is possible these tumor clones were already selected to be resistant, resulting in ineffective treatment. This underscores the need to investigate safety and efficacy of the combination within a clinical trial.

Recommendations

Most authors advise caution with the combination of BRAFi and RT, including some who recommend holding BRAFi for 5 to 14 days before and after RT (17, 21, 35, 47). Some recommend concurrent treatment only within a clinical trial (70), whereas others advise caution but believe concurrent treatment is safe, noting that dermatologic reactions respond well to corticosteroids (29, 32, 34, 37). However, the severity of grade 3 reactions should not be dismissed, and of greater concern are potentially fatal toxicities related to the combination. It is of interest that toxic patient deaths associated with the use of BRAFi involved hemorrhage in the liver, lung, and brain. Because untreated melanoma is marked by neoangiogenesis and is therefore predisposed to hemorrhage, there is difficulty in assigning these events to the combined modality therapy. Of note, in vitro studies have found BRAFi resistance to be associated with reactivation of the mitogen-activated protein kinase (MAPK) pathway, which can lead to vascular endothelial growth factor (VEGF) production (73). Therefore, vascularity and bleeding may be of greater concern when RT is given to BRAFi-resistant lesions. However, there may be a decreased risk for this toxicity when BRAFi and RT are used for other tumor histologies that are less prone to hemorrhage. The vast majority of toxicities have involved RT toxicity enhancement, occurring when RT was administered concurrently or within 7 days of BRAFi use. Only 1 dermatologic grade 3 toxicity occurred when a BRAFi was used over 2 days from RT, but this was in a patient who had a grade 3 reaction before the BRAFi was even started because of the very high RT dose administered. Given that causing severe skin toxicity is inconsistent with the underlying goal of palliation from RT, our recommendation is to hold BRAFi during RT. However, given that melanoma is often widespread and may grow rapidly, interruptions of systemic therapy should be minimized. An analysis of doubling times showed melanoma to have a doubling time of 48 days, compared with 80 and 110 days for adenocarcinomas of breast and colon, respectively (74). Although retrospective data exist suggesting a potential benefit from biologically effective doses >39 Gy₁₀ corresponding to courses stronger than 30 Gy in 10 fractions (75), this improvement may be from selection bias, with patients having better performance status being prescribed longer courses of RT. Potential gains from longer RT courses also must be weighed against the risk of systemic progression while patients are not taking BRAFi. Randomized trials have not shown any difference in response between longer versus more hypofractionated, shorter courses of RT (76). However, in this review, visceral toxicity that resulted in or may have resulted in death occurred only in patients for whom fraction size was ≥4 Gy (16, 17). Although there were multiple cases where this fraction size did not result in significant toxicity, caution must be used in the choice of a nonstereotactic, hypofractionated RT regimen.

Directions for the future hold promise for increased efficacy of treatment from novel combinations of RT and BRAFi, MEKi, or both based on preclinical and clinical data for melanoma. In murine orthotopic xenografts of high-grade glioma, treatment with BRAFi and RT resulted in a survival advantage compared with either modality alone (77). BRAFi alone is being investigated in other histologies; a prospective SWOG study is investigating BRAFi for patients with mutated stage IV colorectal cancer (ClinicalTrials.gov identifier: NCT02164916). The anticipated emergence of numerous trials involving BRAFi underscores the importance of carefully gathered prospective safety data on the use of these various agents with RT. Prospective studies are expected to shed light on the potential interactions between RT and vemurafenib. Those of particular interest include a recently completed phase 2 study of vemurafenib administered as neoadjuvant therapy for patients with CNS metastases before surgery, ablation, or RT (ClinicalTrials.gov identifier: NCT01781026) and an upcoming phase 1 study involving RT with MAPK pathway blockade for patients with oligoprogressive extracranial disease (ClinicalTrials.gov identifier: NCT01843738).

Of great concern is avoiding severe (≥grade 3) dermatologic toxicities, and the risk of this appears negligible with BRAFi held ≥3 days before and after RT. Although Hecht et al (22) noted significantly higher rates of dermatitis for vemurafenib compared with dabrafenib, grade 3 toxicities were similar, and we therefore do not have a recommendation regarding BRAFi preference aside from the general recommendation to enroll patients on a trial. Although there are several case reports of radionecrosis and hemorrhage reported with BRAFi and SRS, these toxicities do not appear increased compared with patients who did not receive a BRAFi. This review contains the first original report of a potential toxicity with RT and a BRAFi with MEKi combination, involving a sigmoid colon perforation that led to the patient's death. Although it is not possible to separate the MEKi contribution to this event, caution with MEKi in the setting of RT is warranted. Although we recognize the difference in half-lives between dabrafenib (8 hours) and vemurafenib (57 hours), for the sake of consistency and simplicity our recommendations pertain to all BRAFi and MEKi agents. Additional limitations of our analysis include the potential for overestimating toxicities from publication bias, because the most severe cases of toxicity will be overrepresented, and we do not know the total number at risk for these toxicities. Therefore, with so few cases of toxicity reported, there is much uncertainty when recommendations are made regarding the safe timing of RT and BRAFi. Until more prospective data are available, the consensus recommendations of the Eastern Cooperative Oncology Group (ECOG) include the following for all patients receiving a BRAFi, MEKi, or both:

• BRAFi and MEKi recommendations (eg, vemurafenib/dabrafenib and trametinib/cobimetinib)

  • Hold ≥3 days before and after fractionated RT.
  • Hold ≥1 day before and after SRS.
• RT recommendations

  • Consider dose per fraction <4 Gy unless using a stereotactic approach or the patient has very poor prognosis/performance status.
  • For adjuvant nodal basin RT, consider a dose ≤48 to 50 Gy in 20 fractions.
  • For spine metastases, consider posterior oblique RT fields when feasible and safe to minimize exit dose through visceral organs.