Topical corticosteroid therapy for facial acneiform eruption due to EGFR inhibitors in metastatic colorectal cancer patients: a randomized controlled trial comparing starting with a very strong or a weak topical corticosteroid (FAEISS study, NCCH1512, colorectal part)

Although pre-emptive therapy with oral tetracycline, moisturizer, sunscreen, and topical corticosteroid is useful for preventing acneiform eruption (AfE) due to epidermal growth factor receptor (EGFR) inhibitors, no studies have examined the efficacy of topical corticosteroids themselves, or investigated the optimal potency of corticosteroid for treating facial AfE (FAfE). Screened patients with RAS wild-type colorectal cancer started pre-emptive therapy with oral minocycline and moisturizer on initiation of cetuximab or panitumumab therapy. Patients who developed grade 1 or 2 FAfE were randomly allocated to two groups: a ranking-down (RD) group that started with a very strong corticosteroid and serially ranked down every 2 weeks unless FAfE exacerbated; and a ranking-up (RU) group that started with a weak corticosteroid and serially ranked up at exacerbation. FAfE grade, patient quality of life, and adverse events (AEs) with topical corticosteroid were evaluated every 2 weeks. The primary endpoint was the total number of times grade 2 or higher FAfE was identified in the central review of the 8-week treatment period. No significant differences in total numbers of grade 2 or higher FAfE or in AEs caused by topical corticosteroids were observed between groups during the 8 weeks. Incidence of grade 2 or higher FAfE tended to be lower in the RD group during the first 2 weeks. Considering the long-term care of FAfE, the RU regimen appears suitable and should be considered the standard treatment for FAfE due to EGFR inhibitor therapy. UMIN Clinical Trials Registry (UMIN000024113).


Introduction
Epidermal growth factor receptor (EGFR) inhibitors including monoclonal antibodies (mAbs) for EGFR and tyrosine kinase inhibitors (TKIs) are effective therapeutic options for patients with colorectal cancer (CRC), non-small cell lung cancer (NSCLC), and squamous cell carcinoma of head and neck (SCCHN). Anti-EGFR mAbs such as cetuximab and panitumumab have been used for metastatic CRC as monotherapies or in combination with other anticancer agents. Patients receiving EGFR inhibitors commonly present with skin toxicities, including acneiform eruption (AfE), which sometimes cause discontinuation of the therapy. AfE develops as early as weeks 1-3 after initiating the therapy and reaches peak frequency by weeks 3-5 [1]. In particular, facial AfE (FAfE) can cause psychological distress and impair patient quality of life (QOL). In the STEPP [2] and J-STEPP [3] studies, pre-emptive therapy with oral doxycycline or minocycline, weak topical corticosteroid, and skin moisturizer were shown to exert favorable effects in preventing AfE with panitumumab therapy, although the effects of topical corticosteroid monotherapy on AfE remain unclear. Because of concerns regarding steroid-induced dermatitis (SID) [4] due to topical corticosteroid for facial lesions, weak-or medium-potency corticosteroids have been used for FAfE in clinical practice. However, the efficacy of weak topical corticosteroids controlling FAfE activity is sometimes unsatisfactory. Recently, rapid effects have been reported from starting topical therapy with a very strong corticosteroid [5]. In addition, distinguishing between SID and FAfE due to EGFR inhibitors is not easy for oncologists treating CRC patients without consulting dermatologists.
The present study was designed to compare two regimens of topical corticosteroid therapy for AfE due to EGFR inhibitors among patients with CRC or NSCLC, with a focus on facial lesions (a phase III, open-label, randomized trial evaluating topical corticosteroid for Facial Acneiform dermatitis by EGFR Inhibitors: Stepwise rank down from potent corticosteroid: FAEISS study; NCCH1512). The regimens comprised one with serial ranking down from a very strong corticosteroid, and one with serial ranking up from a weak corticosteroid. Here, we present the primary analysis of metastatic CRC patients who received cetuximab or panitumumab. The results in NSCLC patients have been published in a separate paper [6].

Patients
Eligible patients were those with RAS wild-type metastatic CRC who received anti-EGFR antibody drugs (cetuximab or panitumumab). Other eligibility criteria included age, 20-79 years; Eastern Cooperative Oncology Group Performance Status [7] score 0 or 1; absence of active bacterial, fungal, and viral infections; and absence of any facial lesions influencing the evaluation of facial lesions. Informed consent for participation in this prospective study was obtained prior to enrolment. Patients with a history of previous anti-EGFR therapy were excluded.

Study design
This study was an open-label, multicenter, randomized controlled trial. As the first registration for screening before randomization, eligible patients started pre-emptive therapy with oral minocycline at 100 mg or 200 mg daily, and application of skin moisturizer containing heparinoid to the face twice daily from the date of starting anti-EGFR antibody therapy. All screened patients were instructed to avoid solar ultraviolet radiation, and use of sunscreen was recommended. Patients were instructed to apply topical 0.5% prednisolone cream when they developed FAfE. Patients who developed FAfE grade 1 or 2 within 8 weeks were randomly assigned to one of two groups: a ranking-up (RU) group or a ranking-down (RD) group. The topical corticosteroids used in this study are listed in Supplementary Table S1. The RU group started topical therapy with 0.5% prednisolone cream. When the grade of FAfE was exacerbated as judged by the attending physician, the topical corticosteroid was ranked up to one class higher. In contrast, the RD regimen started with a topical very strong corticosteroid, and then corticosteroid was changed to one rank lower every 2 weeks. If the grade of FAfE exacerbated, the class of topical corticosteroid was maintained without ranking down (Fig. 1). In both groups, topical corticosteroids of strong or very strong class were used for AfE on non-face areas. If grade 3 or higher AfE affecting the whole body according to Common Terminology Criteria for Adverse Events (CTCAE) version 4.0-JCOG (Japan Clinical Oncology Group) was observed, protocol treatment was terminated.

Evaluations
Efficacy variables were FAfE grade and patient QOL. For the assessment of endpoints, FAfE was graded after the end of the 8-week treatment by three trained dermatologists (central review) based on digital photographs of the face taken under a pre-determined, standardized method on each evaluation date. According to the qualitative scheme described by Scope et al. [8], with the following minor modification, FAfE was classified as grade 0, no symptoms; grade 1, less than one-third of the face involved; grade 2, more than onethird but less than two-thirds of the face involved; or grade 3, more than two-thirds of the face involved. QOL was assessed using the Dermatology Life Quality Index (DLQI) questionnaire immediately before and after completing 8 weeks of topical corticosteroid therapy.
Safety variables were specified as topical corticosteroid-and EGFR inhibitor-induced adverse events (AEs). AEs due to topical corticosteroid were also assessed by 3 dermatologists based on the digital photographs of the face according to the following criteria for SID and topical (bacterial/fungal/viral) infection. SID was assessed as grade 0, no symptoms; grade 1, mild symptom, not interfering with daily life; grade 2, moderate symptom, partially interfering with daily life; or grade 3, severe symptom, severely interfering with daily life. Topical infection was assessed as grade 0, no infection; grade 1, involved lesion affects less than one-third of the area of face receiving topical treatment; grade 2, involved lesion affects more than one-third and less than two-thirds of the area of the face receiving topical treatment, and requires use of antibacterial medication; or grade 3, involved lesion affects more than two-thirds of the area of the face receiving topical treatment, and requires use of intravenous antimicrobial agents. The severity of EGFR inhibitor-induced AEs of the skin (i.e., pruritus, paronychia, and xerosis) and AfE of the whole body were assessed according to CTCAE version 4.0-JCOG by an attending physician.
Efficacy and safety variables were evaluated at the time of randomization (i.e., on the start date) and at 2, 4, 6, and 8 weeks after starting randomized topical corticosteroid therapy, with the exception of the aforementioned QOL assessments.

Study objectives (endpoints)
On central review, presence or absence of FAfE grade 2 or higher at every evaluation date during the 8-week study period was checked, and the total number of times FAfE grade 2 or higher occurred during the 8-week study period was determined as the primary endpoint. Secondary endpoints were the incidences of FAfE of grade 2 or higher, FAfE of grade 3 or higher, and AEs due to topical corticosteroid during the study period. The numbers of patients who continued EGFR therapy, with FAfE of grade 1 or higher and with non-lowered QOL scores at the end of the study period, were also evaluated.

Statistical analyses
A superiority test design was applied to determine whether the duration of grade 2 or higher FAfE during the study period was shorter in the RD group compared to the RU group. The total number of times FAfE grade 2 or higher was observed in the 8-week study (5 evaluation points) was compared between groups using the stratified Wilcoxon rank-sum test, with drug and sex as stratification factors. To test the null hypothesis (i.e., that there are no differences between groups in terms of the primary endpoint) using the stratified Wilcoxon rank-sum test with a one-sided alpha of 0.05 and 80% power, 94 patients were required for the primary analysis. Assuming the incident proportion of FAfE (grade 1 or higher) as 85%, a sample size of 120 patients was planned for the first registration. Although this study was initially planned to include NSCLC patients, due to the low incidence of FAfE from EGFR-TKI treatment, registration of NSCLC patients was closed early and the results of that subgroup were published independently [6]. The incidences of FAfE of grade 2 or higher, and of FAfE grade 3 or higher, and the proportions of continued EGFR therapy, FAfE grade 1 or higher, and non-lowered QOL scores at the end of the study were evaluated using Fisher's exact test. Incidences of AEs due to topical CS during the 8-week study period were compared between the RU and RD groups using Fisher's exact test. In addition, subgroup analyses for drug (cetuximab or panitumumab), additional anticancer drugs, sex, and age (less than vs. greater than or equal to the median) were performed. One-sided p < 0.05 was considered statistically significant for analysis of the primary endpoint. Statistical analyses were performed using SAS® version 9.4 software (SAS Institute, Cary, NC).

Patients
Between October 2016 and December 2018, a total of 172 patients with metastatic CRC were enrolled at the first  Oral minocycline 100 or 200 mg daily and skin moisturizer Pre-empƟve therapy registration and received pre-emptive therapy. All 106 patients who developed grade 1 or 2 facial AfE were randomized to the RU group (n = 53) or RD group (n = 53). Baseline characteristics were well balanced between groups (Table 1) and the proportion of grade 2 or higher FAfE was also similar between the RU group (11.3%) and RD group (9.8%). Two patients in the RD group terminated EGFR inhibitor therapy due to progression of the primary disease (untreated patients in Fig. 2). In the RD group, 1 patient judged as grade 3 by central review was ineligible. In the RU group, the potency of corticosteroid was ranked up to medium class in 2 patients before randomization within 2 weeks after initiating EGFR inhibitor therapy because of exacerbated FAfE.
In the RD group, 16 patients failed to achieve stepwise ranking down due to exacerbation of FAfE at the 8-week evaluation ( Table 2). The potency of corticosteroid was ranked up to medium in 10 patients and to strong in 2 patients at the 8-week evaluation in the RU group (Table 2).

Efficacy
As the primary endpoint, no significant difference in the total number of observation of grade 2 or higher FAfE was seen between groups, using anticancer agents and sex as stratification factors (p = 0.86) ( Table 3). The incidence of FAfE grade 2 or higher increased in a time-dependent manner. In the RU group, 11.3%, 26.1%, 32.6%, 35.0%, and 35.7% of patients had developed FAfE grade 2 or higher at weeks 0 (second registration), 2, 4, 6, and 8 of the study, respectively. In the RD group, 9.8%, 15.7%, 34.1%, 39%, and 53.8% developed FAfE grade 2 or higher at the corresponding weeks ( Table 4). Incidences of grade 2 or higher FAfE and grade 3 or higher FAfE during the 8 weeks, continuation of EGFR therapy, and grade 1 or higher FAfE at the end of protocol treatment showed no differences between groups (Supplementary Table S2).

QOL
The number of patients with no worsening of QOL score was low in both groups, at 2.0% in the RU group and 6.4% in the RD group, and no significant difference was apparent between groups (Supplementary Table S2).

Safety
Before randomization, 2 patients discontinued the preemptive therapy because of pruritus or liver dysfunction, probably caused by minocycline. After randomization, 2 patients in the RU group terminated protocol treatment due to grade 3 diarrhea and grade 3 anorexia, caused by oral minocycline. One patient in the RD group terminated treatment due to local infection of the face, which was suspected to be due to topical corticosteroid. However, no differences in incidences of AEs caused by topical corticosteroid (i.e., SID and local infection) were seen between groups (Supplementary Table S3).

Discussion
In the present study, pre-emptive therapy with oral minocycline and skin moisturizer on the face were started simultaneously with EGFR inhibitors, and topical 0.5% prednisolone cream, as a weak corticosteroid, was permitted for use against FAfE that developed before randomization. Grade 1 or 2 FAfE developed in 106 of 172 subjects (61.6%) before randomization. This incidence is lower than those of previous reports without pre-emptive therapy, in which 90% with panitumumab and 80-86% with cetuximab developed AfE at some site on the body [9]. Favorable effects from the preemptive therapy on FAfE were suggested.
The duration of grade 2 or higher FAfE during the 8-week study period was assessed as the number of observations among the 5 observation time points for the primary endpoint. The results showed no significant difference between RU and RD groups, and we could not identify any advantage of starting with a very strong corticosteroid. Moreover, no significant differences were seen between regimens in secondary endpoints, as the incidence of grade 2 or higher or grade 3 or higher FAfE during the study period, proportions with continuation of EGFR inhibitors, grade 1 or higher FAfE, or the percentage of patients with no worsening of QOL at the end of the study.
The RD regimen showed more favorable effects during the first 2 weeks but could not maintain control later and more than 50% of patients showed grade 2 or higher FAfE by the end of the study. The severity of FAfE could potentially have increased over time in some patients, probably because of the initial pre-emptive treatment with oral minocycline, and FAfE failed to maintain control in the RD group later in the treatment period, because the potency of topical corticosteroid was weakened in accordance with the protocol. This also suggested that topical corticosteroids offering strong potency or above are effective in controlling FAfE, whereas topical corticosteroids of medium potency or below might be inadequate to control FAfE in some severe cases.
Long-term use of stronger topical corticosteroids on the face carries a risk of AEs including SID and secondary  Primary efficacy analysis sets Primary safety analysis sets Table 2 Topical corticosteroid used during the course of protocol treatment One * out of 2 patients in the RU group and one ** in the RD group were assessed as ineligible.
Because of disappearance of AfE, some patients stopped using topical CS. infection, so we used a stepwise ranking down every 2 weeks from a very strong corticosteroid to a weak corticosteroid unless exacerbation developed. Grade 1 SID was observed in 1 patient in the RU regimen. Grade 1 SID in 2 patients and grade 2 local infection in 1 patient were observed in the RD regimen, but no marked differences were seen between regimens.

Conclusion
Pre-emptive therapy with oral minocycline and skin moisturizer showed some promise for reducing the risk of grade 2 FAfE. Stepwise down-ranking from a very strong topical corticosteroid for FAfE worked during the first 2 weeks, but failed to show significant advantages over the RU regimen from a weak corticosteroid through the 8 weeks of the study. Short-term use of a very strong corticosteroid may have some merit for suppressing initial FAfE, but a standard RU regimen from a weak corticosteroid appears suitable in longterm care for FAfE and should be considered the standard treatment for FAfE due to EGFR inhibitor therapy.

Limitations
The duration of protocol treatment was only 8 weeks and no follow-up period was provided afterward. Long-term observation of the duration and severity of FAfE, the potency of topical corticosteroid needed, and the AEs of topical corticosteroids on the face are necessary. We chose a central review using only photographs of the face to eliminate differences in evaluations by board-certified dermatologists and non-dermatologists, because not all hospitals have an attending dermatologist to evaluate skin symptoms.