Final analysis of a post‐marketing surveillance study of dabrafenib and trametinib combination therapy in Japanese patients with unresectable malignant melanoma with BRAF V600 mutation

Targeted therapy with a combination of dabrafenib and trametinib has been developed and widely used for treatment of melanoma. However, data regarding its safety and efficacy in Japanese patients with malignant melanoma are limited. A post‐marketing surveillance (PMS) study was conducted to investigate the safety and efficacy of combination therapy in a Japanese clinical setting with a surveillance period of June 2016 to March 2022; 326 patients with unresectable malignant melanoma with BRAF mutation were enrolled. The interim results were published in July 2020. Herein, we report the results of the final analysis based on the data collected until the completion of the PMS study. The safety analysis population included 326 patients, the majority of whom had stage IV disease (79.14%) and Eastern Cooperative Oncology Group performance status 0 or 1 (85.28%). All patients were treated with the approved dose of dabrafenib, while 99.08% were treated with the approved dose of trametinib. Adverse events (AEs) occurred in 282 patients (86.50%) and the major AEs (incidence ≥5%) were pyrexia (47.85%), malignant melanoma (33.44%), hepatic function abnormal (9.82%), rash and blood creatine phosphokinase increased (8.59% each), malaise (6.44%), nausea (5.52%), and diarrhea and rhabdomyolysis (5.21% each). The incidences rates of adverse drug reactions of safety specifications were 45.71% for pyrexia, 15.95% for hepatic impairment, 12.58% for rhabdomyolysis, 4.60% for cardiac disorders, and 3.07% for eye disorders. In the efficacy analysis population of 318 patients, the objective response rate was 58.18% (95% confidence interval [CI] 52.54%–63.66%). The progression‐free survival rates at 90, 180, and 360 days were 88.14% (95% CI 84.00%–91.26%), 69.53% (63.85%–74.50%), and 52.07% (45.71%–58.03%), respectively. Consistent with previous interim results, no new safety or efficacy concerns were observed in this final analysis of a PMS study conducted in a Japanese real‐world clinical setting.


| INTRODUC TI ON
Patients with malignant melanoma had a poor prognosis, and none of the available agents improved the overall survival of patients with malignant melanoma in a phase 3 trial conducted before 2010. 1 Thereafter, major developments in targeted therapies and immunomodulators have provided robust improvements in the response and survival of patients with malignant melanoma. [2][3][4] A targeted therapy for BRAF kinase and mitogen-activated protein kinase pathway, which includes MEK with a combination of dabrafenib and trametinib, has been developed and widely used for the treatment of melanoma. 2,5 In pivotal global phase 3 studies, the combination of dabrafenib and trametinib demonstrated statistically significant and clinically relevant improvements in the progressionfree survival (PFS) and overall survival of patients with BRAF V600 mutation-positive melanoma, which established this combination therapy as a standard treatment option for BRAF V600 mutationpositive melanoma. 6,7 A phase 1/2 study conducted in Japanese patients with BRAF V600 mutation-positive advanced cutaneous melanoma demonstrated that the efficacy and pharmacokinetic properties of dabrafenib and trametinib combination in Japanese patients were comparable to those observed in global studies. 8 In Japan, combination therapy with dabrafenib and trametinib was approved for the treatment of BRAF mutation-positive unresectable malignant melanoma in 2016. Thereafter, an additional indication for the postoperative adjuvant therapy for malignant melanoma was approved in 2018.
Although the results of some retrospective investigations on the real-world efficacy and safety in Japanese patients have been recently reported, 9,10  Interim results were published in July 2020 based on the data collected until November 2018, and no new safety or efficacy concerns were reported. 11 Here, the results of the final analysis of the PMS were based on the data collected from over 300 patients until the completion of the PMS in March 2022.

| Study design and patients
This was the final analysis of a PMS study conducted in Japan to collect information on the safety and efficacy of dabrafenib and trametinib in clinical practice. All patients with BRAF V600 mutationpositive melanoma who were administered with dabrafenib and/or trametinib in Japan between June 2016 and September 2017 were included in the study (all-case investigation). Patients with melanoma who were administered with dabrafenib and trametinib as adjuvant therapy after surgery were excluded. Patients treated with dabrafenib or trametinib monotherapy were enrolled but excluded from the safety and efficacy analysis sets in the PMS. The observation period was 1 year.
The PMS was an observational survey conducted by Novartis Pharma K.K. in accordance with the ministerial ordinance on Good Post-Marketing Study Practice in Japan, which determines that informed consent is unnecessarily required for conducting the PMS in Japan. However, Novartis Pharma K.K. provided the investigator with an information sheet describing the objective and content of this surveillance, protection of personal information, and publication of surveillance data, and the investigator explained the surveillance to patients using the information sheet before the start of treatment. The protocol of this surveillance was approved by the institutional review board of each medical site.

| Study assessments and statistical analyses
The demographic and baseline data, status of dabrafenib and/or trametinib administration, safety endpoints, and efficacy endpoints were evaluated using the method described in the interim analysis report for the PMS. 11 Adverse drug reactions (ADRs) were defined as adverse events (AEs) in which a causal relationship to dabrafenib and/or trametinib was not denied by the investigator. ADRs of safety specifications in this PMS included the following events in advance: cardiac disorders, hepatic impairment, pyrexia, eye disorders, squamous cell carcinoma, secondary malignancies other than squamous cell carcinoma, and rhabdomyolysis. The statistical analyses were also performed using the method described in the interim analysis report for the PMS. 11

| Patients' disposition
The patients' disposition is shown in Figure 1. A total of 707 patients from 166 sites in Japan were registered in the PMS study. The safety analysis population comprised 326 patients after excluding two patients who were transferred to another hospital and 29 patients of dabrafenib or trametinib monotherapy, off-label use, prior use of dabrafenib or trametinib, or participants in a clinical trial of an unapproved drug. Of the 326 patients in the safety analysis population, 318 were included in the efficacy analysis population after excluding eight patients who were not evaluable for efficacy.

| Status of dabrafenib and trametinib therapy
Of the 326 patients in the safety analysis set, 205 (62.88%) discontinued both dabrafenib and trametinib therapy, and the most frequent reasons for treatment discontinuation were occurrence of AEs in 113 patients (55.12%) and an insufficient therapeutic effect in 101 patients (49.27%) (which included some overlap) (Supporting Information Table S1).
All patients were treated with the approved dose (≤300 mg/day) of dabrafenib. Most patients (323 patients, 99.08%) were treated with the approved dose (≤2 mg/day) of trametinib. The mean duration of the combination therapy was 214.5 ± 123.31 days (Supporting Information Table S2).
Serious ADRs were reported in 67 patients (20.55%), and the major serious ADRs (incidence ≥ 2%) were pyrexia in 24 patients (7.36%), malignant melanoma in eight patients (2.45%), and rhabdomyolysis in seven patients (2.15%) ( Table 2). The outcomes of serious ADRs were death in five patients (all malignant melanoma), sequelae in one patient (thalamus hemorrhage), not resolved in four patients (of whom two had malignant melanoma, one had putamen hemorrhage, one had blood lactate dehydrogenase increased, and one had C-reactive protein increased), and unknown/not specified in one patient (uveitis). The outcomes of the other events were either resolved or resolving. Of these patients, one patient with malignant melanoma with an outcome of death and one patient with increased blood lactate dehydrogenase with an outcome of not resolved were the same patient (data not shown).
A total of 98 AEs leading to death occurred in 79 patients.
Of these, five events in five patients were ADRs (all malignant melanoma) which were related to both dabrafenib and trametinib (data not shown).

| Safety analysis by patient factor
To examine the contributing factors that may affect safety, the incidence of ADRs was analyzed based on patient factors (Supporting Information Table S4). The contributing factors for which the 95% confidence interval (CI) of the odds ratio did not include 1 were "sex" and "immediate prior therapy for underlying disease using nivolumab" ( Table 4). The adjusted analysis using these two factors as explanatory variables of the logistic regression model showed that the 95% CI of the adjusted odds ratio did not include 1 for either "sex" or "immediate prior therapy for underlying disease using nivolumab" (data not shown).
The incidence of ADRs of pyrexia was also analyzed based on patient factors. The contributing factors for which the 95% CI of the odds ratio did not include 1 (lower incidence of ADR of pyrexia) were "elderly (≥65 years)," "elderly (≥75 years)," and "immediate prior therapy for underlying disease using nivolumab: No" (data not shown). An adjusted analysis was performed using two contributing factors, "age" and "immediate prior therapy for underlying disease using nivolumab," as explanatory variables of the logistic regression model. The results showed that the adjusted odds ratio (95% CI) for "age (elderly ≥65 years)" was 0.59 (0.38-0.92), indicating a higher incidence of ADRs of pyrexia in patients aged <65 years. For the other two factors, 95% CIs of the adjusted odds ratio included 1 (data not shown).

Patient background factors
Category/summary statistics n (%) Abbreviations: ADR, adverse drug reaction; AE, adverse event; PT, preferred term. a One event was excluded from the analysis, since its seriousness was unknown.

TA B L E 3
Incidence of ADRs of safety specifications (safety analysis population).  elderly patients (aged <75 years) (Supporting Information   Table S4).

| Antitumor effect
In the efficacy analysis set of 318 patients, the response rate based on the physician's overall assessment was evaluated in accordance with RECIST version 1.  (Figure 2b-d).

| Efficacy analysis by patient factor
To examine the contributing factors that may affect the treatment efficacy, the ORR was analyzed based on patient factors (Supporting Information Table S5). The results showed that the 95% CI of the odds ratio in the categories did not include one in patients with "age" "prior therapy for underlying disease using nimustine," "prior therapy for underlying disease using vemurafenib," and "immediate prior therapy for underlying disease using vemurafenib" (Supporting Information Table S6). An adjusted analysis was performed using the following three contributing factors as explanatory variables of the logistic regression model: "age", "prior therapy for underlying disease using nimustine", and "immediate prior therapy for underlying disease using vemurafenib".
The results showed that the 95% CI of the adjusted odds ratio did not include one in patients with "age" or "immediate prior therapy for underlying disease using vemurafenib", but included one in patients with "prior therapy for underlying disease using nimustine" (data not shown).

| Efficacy analysis in patients with special characteristics
The response rate of patients with special characteristics (elderly patients and patients complicated with renal impairment, hepatic impairment, or cardiac disease) was investigated. Although the efficacy in elderly patients (aged ≥65 years; 51.03%, 74/145 patients) was slightly lower than that in nonelderly patients (aged <65 years; 64.16%, 111/173 patients), none of the results raised concerns about the impact of age on efficacy (data not shown). The denominator of incidence of ADRs is the safety analysis population for each category. b Indicates the reference categories.

| Positioning of this surveillance
Previous investigations on the safety and efficacy of dabrafenib and trametinib combination therapy in Japanese melanoma patients included only 12 patients in a Japanese phase 1/2 clinical trial 8 and 112 patients in a retrospective observational study. 10 This PMS study included over 300 Japanese patients with melanoma treated with dabrafenib and trametinib. The results of this surveillance will provide useful information for physicians in Japanese clinical practice.

| Status of dabrafenib and trametinib therapy
All patients were treated with dabrafenib (≤300 mg/day) and most patients (99.08%) were treated with trametinib (≤2 mg/day) in this surveillance, therefore in Japanese clinical practice, dabrafenib and trametinib are suggested to be generally administered within the dosage range specified in the package insert for each drug.

| Safety results
The incidence rates of AEs were 95% (199/209 patients) and 98% (343/350 patients) in the pivotal phase 3 COMBI-d 6 and the COMBI-v studies 7 evaluating the safety and efficacy of dabrafenib and trametinib as combination therapy, respectively, which did not include Japanese patients. The incidence rates were 100% (12/12 patients) in the Japanese phase 1/2 study 8 and 88% (44/50 patients) in a Japanese observational study. 9 In this surveillance, the incidence of AEs was 86.50% (282/326 patients), which was not markedly higher than that reported in previous studies. In addi- a combination of grade 3/4 pyrexia, hospitalization due to pyrexia, and/or permanent treatment discontinuation due to pyrexia. This algorithm was suggested to reduce the incidence of composite pyrexia events to 8.0% compared with a historical control (20.0%) in the COMBI-AD study. 12 The results of the COMBI-APlus study were  these AEs were 3.68%, 47.85%, and 6.13% for eye disorders, pyrexia, and cardiac disorders, respectively. These results suggest that the combination therapy of dabrafenib and trametinib may result in a higher incidence of pyrexia and a lower incidence of eye disorders compared with that of encorafenib and binimetinib. Clinicians should consider the differences in the incidence of AE types for each combination therapy of BRAF and MEK inhibitors, and use each combination therapy to achieve better outcomes in melanoma patients.
In the safety analysis by patient factor, the factors for which the 95% CI of odds ratio did not include 1 were "sex" and "immediate prior therapy for underlying disease using nivolumab". The incidence of ADRs was higher in females in "sex" and in patients with "immediate prior therapy for underlying disease using nivolumab: Yes". These were explanatory values obtained first in our surveillance. Based on spontaneous reports and the results of research reports collected to date, no information showing a similar tendency has been obtained, hence collection of relevant information will be continued in pharmacovigilance activities.

| Efficacy results
The The PFS rate was lower in patients with elevated LDH level before the start of treatment than in those with normal LDH levels. In the COMBI-d and COMBI-v studies, the PFS rate tended to be lower in patients with an LDH level above the upper limit of the normal range at baseline than in those with an LDH level at or below the upper limit of normal. 6,7 In a Japanese observational study, a lower PFS rate tendency was reported in patients with an elevated LDH level at baseline compared with that in patients with a normal LDH at baseline. 9 In the present surveillance, the PFS rate tended to be lower in patients with ECOG PS 2-4 compared with that in patients with ECOG PS 0-1. Although patients with ECOG PS 2-4 were not included in the COMBI-d and COMBI-v studies, subgroup analyses showed a lower trend of PFS in patients with ECOG PS 1 than in those with ECOG PS 0. 6,7 Therefore, in terms of PFS according to the LDH level or ECOG PS, a similar tendency was observed among the present surveillance, clinical trials, and previous observational study. In the present surveillance, the PFS rate tended to be lower in patients with prior IO drugs than in those without prior IO drugs, although comparable data from previous studies were not available.
In the present survey, the median PFS was not reached (95% CI 329.0 days-not reached). The median PFS was 9.3 months (95% CI not available) in the COMBI-d study, 6 11.4 months (95% CI not available) in the COMBI-v trial, 7 and 12 months (95% CI 0-27.2 months) in a Japanese observational study, 9 therefore the median PFS in the present survey was the same or longer than that reported in the previous clinical trials and observational study. Of note, interstitial lung disease (ILD) was more commonly reported in patients with immediate prior therapy with nivolumab compared to patients without nivolumab therapy (incidence 6.15% vs 0.77%). Furthermore, most of the patients with ILD were elderly, aged ≥65 years, except for one patient aged 51 years. Additional safety measurements are considered unnecessary because of the small amount of collected data for this concern and the outcomes of all reported ILDs were "resolved" or "resolving". Since a relationship of ILD incidence after dabrafenib plus trametinib combination therapy with the nivolumab therapy cannot be ruled out, the relationship needs to be investigated in the ongoing pharmacovigilance activities.

| Limitations
This was a single-arm surveillance study with no control group (patients not treated with dabrafenib and trametinib), therefore there is a limitation in interpreting the causality between exposure to the drugs and the obtained results.

| CON CLUS IONS
No new safety or efficacy concerns were observed in this final analysis of PMS in Japanese patients with unresectable malignant melanoma with BRAF mutation who received dabrafenib and trametinib in a real-world clinical setting. In addition, there were no major differences compared with the incidence or type of AEs in clinical trials.
We believe that it is important to obtain more detailed information on the safety and efficacy of these drugs in clinical practice.

ACK N OWLED G M ENTS
We would like to thank Akihira Mukaiyama of Novartis Pharma K.K.
(Japan) for his help in planning this PMS study.

CO N FLI C T O F I NTER E S T S TATEM ENT
This PMS was sponsored by Novartis Pharma K.K. (Japan) and all authors are employees of Novartis Pharma K.K. Medical writing support was provided by Havas Health & I Co., Ltd. and was funded by Novartis Pharma K.K.