Can Rash Associated With HER1/EGFR Inhibition Be Used as a Marker of Treatment Outcome?

Erlotinib (Tarceva) is an orally available, potent, highly specific, reversible inhibitor of HER1/epidermal growth factor receptor (EGFR) tyrosine kinase. Phase II trials have shown clinical activity of erlotinib in non-small-cell lung cancer, head and neck cancer, and ovarian cancer, with a common toxicity being a skin rash that typically is acneiform in character and mild to moderate in severity. This skin rash is a class effect of EGFR inhibitors. Accumulating data with erlotinib and other EGFR inhibitors suggest that rash may be predictive of objective tumor response and prolonged survival in cancer patients. Skin Rash as a Toxicity of HER1/EGFR Inhibition HER1/EGFR is expressed in the epidermis, sebaceous glands, and hair follicle epithelium, and is recognized to play a role in the normal differentiation and development of skin follicles and keratinocytes. Studies in mouse models including HER1/EGFRknockout studies, expression of dominant negative HER1/EGFR mutants, and partial HER1/EGFR inhibition showed that absence or blocking of the receptor produced hair, skin, and eye abnormalities, including necrosis and disappearance of hair follicles and presence of inflammatory elements in skin.[1-4] Administration of erlotinib in the nude mouse model produced skin lesions consisting of diffuse, mild-to-moderate thickening of the epidermis, infiltration of mostly acute inflammatory cells in the dermis, and formation of dry scabs on the eyes.[5] The current prevailing hypothesis for the mechanism of skin toxicities as- sociated with HER1/EGFR inhibition is that this inhibition results in follicular occlusion due to lack of differentiation in the epithelium and affects sebaceous glands to produce a rosacea- like reaction, resulting in increased production of inflammatory mediators and inflammation. The cutaneous toxicity associated with HER1/EGFR inhibition in preclinical studies is manifest primarily as an acneiform skin rash in the clinical setting, with other skin disorders including pruritus, dry skin, and erythema. The skin rash is a class-related effect, having been observed with all HER1/EGFR tyrosine kinase inhibitors (eg, erlotinib, gefitinib(Drug information on gefitinib) [Iressa]), dual HER inhibitors (eg, GW2016, and pan-HER inhibitors (eg, CI-1033), as well as with HER1/ EGFR-specific monoclonal antibodies (eg, cetuximab(Drug information on cetuximab) [Erbitux], ABX-EGF, EMD72000). Table 1 provides an example of the consistently high rash/ cutaneous toxicity frequency observed with erlotinib, gefitinib, and cetuximab in phase II trials in non- small-cell lung cancer and with ABXEGF in a phase II trial in colorectal cancer[6-9]; it should be noted that cutaneous toxicity is fairly infrequent in cancer chemotherapy, and thus categorization and grading of the particular cutaneous toxicities have likely been somewhat inconsistent among clinical trials. Available evidence indicates that the frequency of rash is dose-dependent. The rash generally occurs above the waist, with involvement of the nose area being most common, and is characterized by clusters of monomorphic pustular lesions.[10] The most common histopathologic finding is neutrophilic infiltration of dermal tissues. There is anecdotal evidence that rash spontaneously resolves in some patients. Currently, there are no clear guidelines for rash management. Treatment options include topical/systemic antibacterials (which may help due to superinfection of affected areas), retinoids, short-term systemic corticosteroids (for inflammatory reaction), but such measures have generally had limited efficacy in resolving rash. Further studies are required to improve understanding of rash etiology and management. Association of Rash With Tumor Response and Survival Observations in studies with HER1/ EGFR inhibitors have suggested that rash may be predictive of tumor response and prolonged survival and have prompted analysis of these potential associations in study populations. The incidence and grade of rash in phase II studies of erlotinib in refractory non-small-cell lung cancer, head and neck cancer, and ovarian cancer are shown in Figure 1. A significant relationship between rash severity and objective tumor response was observed in the trial in non-smallcell lung cancer and for the three studies combined, with these data also showing that objective responses were not observed in any patients without rash (Figure 1).[11] Analysis of survival in the trial in non-small-cell lung cancer shows a strong association of rash with prolonged survival (Figure 2). Median survival among patients without rash was 46.5 days, compared with 257 days in those with grade 1 rash (P < .0001) and 597 days in those with grade 2/3 rash (P < .0001). As shown in Table 2, there were trends toward significantly increased survival in patients with grade 1 rash in the erlotinib phase II trials in head and neck cancer and ovarian cancer, and the occurrence of grade 2 rash was associated with a significant increase in survival duration in both studies. Pooling of the three phase II studies shows a strong association of rash with prolonged survival (Figure 2); median survival in patients with no rash was 103 days, compared with 191 days in those with grade 1 rash (P = .0001) and 266 days in those with grade 2/3/4 rash (P = .0001). With regard to other EGFR-targeted agents, there is clear evidence of an association of rash and survival with cetuximab. As shown in Figure 3, a significant relationship between rash grade and survival has been observed in four studies of cetuximab alone or in combination in nearly 300 patients with colorectal cancer, head and neck cancer, or pancreatic cancer.[ 12] Evidence of such an association currently is somewhat less strong for gefitinib, although limited data in this regard have been reported; as yet, no data on the potential association of rash with response/survival have been reported from the IDEAL trials of this agent in non-small-cell lung cancer. However, in the IDEAL trials, the incidence of rash was slightly higher in patients who received the higher dose of gefitinib (500 mg) as shown in Table 1, and there was no difference in survival in the patients who received 500 or 250 mg. That said, an analysis of the relationship between rash and survival within each group of patients has not been presented. In a phase II study of gefitinib in 52 patients with head and neck cancer, skin toxicity was significantly associated with objective response (P = .004), progression-free survival (P = .0002), and overall survival (P = .001).[13] More recently, skin rash was reported to be associated with enhanced survial in a phase II study of gefitinib in patients with bronchioalveolar carcinoma.[14] However, no correlations between skin toxicity and response/survival were observed in a phase II study of gefitinib in combination with carboplatin(Drug information on carboplatin) (Paraplatin)/ paclitaxel(Drug information on paclitaxel) in 24 patients with non- small-cell lung cancer or in a phase II study of gefitinib combined with FOLFOX 4 in 43 patients with colorectal cancer.[15,16] It bears noting that gefitinib added no benefit to carboplatin/paclitaxel in the former study and exhibited no antitumor activity in colorectal cancer in the latter. A relationship between rash and response/ survival has been observed with ABX-EGF, although data in this regard are limited. No data on these potential associations have been reported for CI-1033 or EMD72000. There are important potential implications for the association of rash and rash severity with response and survival. It is possible that rash can be used as a marker for clinical activity, and it may be that rash should be used as a tool to maximize clinical response. The association of rash and survival observed with erlotinib, for example, suggests that this agent should be given at its maximum tolerated dose, at which rash is more frequent and severe, in order to achieve greater response rates and survival durations. These important implications are being assessed in a phase II erlotinib dose-to-rash trial. In this trial, patients with stage IIIB/IV non-small-cell lung cancer who have received at least one prior chemotherapy regimen and who have Eastern Cooperative Oncology Group performance status of 0 or 1 are receiving a starting dose of 150 mg daily, with the dose being escalated in 25- or 50-mg increments in the absence of grade 2 or worse toxicity. The primary study outcome measures are objective response rate and duration of response analyzed by grade of rash. Correlation of Rash With Pharmacokinetic Parameters Limited available data suggest a potential correlation between rash and increased erlotinib exposure. Data from a phase II trial of erlotinib monotherapy in patients with breast cancer suggest a correlation between erlotinib area under the concentration:time curve and maximum severity of rash experienced (Figure 4).[16] There was a strong trend toward reduced time to most severe rash experienced in patients with drug exposure above the 25th percentile compared with those with lower drug exposure (P = .071) (Figure 4); overall, the median time to rash onset was 8 days in the higher-exposure group compared with 14.5 days in the lower-exposure group and the median times to highest grade of rash were 12 days and 28 days, respectively. These pharmacodynamics observations could be the result of pharmacokinetic variation or pharmacogenomic heterogeneity among patients or both, but do not rule out the ability to increase rash frequency or severity by increasing the drug dose. These observations provide some support for the notion that increasing drug exposure may result in more severe rash and reduced time to rash onset as an indicator of improved clinical response and a predictor of improved outcome. Conclusion Rash appears in most cases to be a necessary, but not sufficient, condition for response to erlotinib treatment. The consistent relationship between higher grades of rash and increased response rates or survival durations with erlotinib treatment suggests that rash may be an important tool to optimizing treatment outcomes. Further studies and retrospective analyses of prior studies are needed to improve understanding of this relationship.