Phosphoinositide 3-kinase (PI3K) inhibitors are a new class of cancer therapeutics that inhibits one or more enzymes in the PI3K/AKT/mTOR tumor growth pathway. As compared to other tyrosine kinase inhibitors, there is evidence that PI3K inhibitors have a higher incidence of severe cutaneous adverse events (CAEs) ranging from 2-21%. There is a lack of further characterization of clinical trials and management options for these CAEs.
A retrospective chart review of our institution's records between January 2015 and May 2019 was conducted; electronic medical records were queried by using a pharmacy database and ICD-10 codes for patients receiving PI3K inhibitor who experienced CAEs during therapy. These CAEs were characterized by two board-certified dermatologists at a major cancer center.
Eleven patients were identified as having 12 cumulative CAEs. Average time to rash onset was 4 weeks, and the most common identified rashes were eczematous (25%) and morbilliform (17%). Four patients experienced a dose delay, and one patient immediately discontinued their PI3K inhibitor.
Although most CAEs caused by PI3K inhibitors in this study were limited to grade 1–2 and were controlled with topical corticosteroids and oral antihistamines, a number of patients experienced dose impact. This highlights the dermatologist's role in managing and minimizing interruption of therapy while maintaining quality of life.
Phosphoinositide 3-kinase (PI3K) inhibitors are a relatively new class of therapeutics that targets the PI3K/AKT/mTOR pathway, leading to tumor growth suppression.[1–3] The first PI3K inhibitor, idelalisib, was approved by the US Food and Drug Administration (FDA) in July 2014 for chronic lymphocytic leukemia. Since then, many PI3K inhibitors with differing molecular targets are undergoing clinical trials to test their efficacy and safety. Common adverse events (AEs) of PI3K inhibitors include diarrhea, fatigue, and cutaneous adverse events (CAEs).
CAEs related to cancer therapies have become an increasingly important issue across targeted therapies with incidence rates as high as 50% in combination therapies. CAEs are some of the most common AEs related to PI3K inhibitors, and the incidences from trial literature are significant, from 10–48%.[6,7] CAEs can range from painful acneiform eruptions to life-threatening Stevens-Johnson syndrome and lead to dose reduction or discontinuation of cancer therapy. They can be difficult to manage without interfering with the tumor response.
The severity and incidence of CAEs may vary by the type of PI3K inhibitor and its manipulated isoform. Idelalisib (delta-isoform specific) has an all grade rash incidence of 10–13%, whereas pilaralisib (pan-isoform specific)–associated rashes have a 32% incidence.[6,9,10] This increase is thought to be due to PI3K alpha-isoform inhibition by pilaralisib—the alpha isoform is universally expressed in mammalian cells including skin cells, and the inhibition of this pathway could account for the higher incidence. High-grade rashes associated with idelalisib have an incidence of 2%, whereas high-grade rashes associated with duvelisib, a delta gamma dual-isoform––specific PI3K inhibitor, have an incidence of 21%.[6,11] AEs (31%) and disease progression (24%) were the main reasons for treatment discontinuation of duvelisib. CAE was the most common AE that led to drug dose modification in patients with pictilisib therapy, another class-1 pan-isoform PI3K inhibitor that had a high rash incidence of 48%. CAEs related to PI3K inhibitors have significant impact on dose continuity, and from a dermatology standpoint, need to be probed to provide holistic supportive care for cancer patients.
Trial data show that PI3K inhibitors may cause more severe grade reactions than other targeted therapies, necessitating a study of the diagnosis and management of CAEs. For example, alpelisib (PI3K alpha-specific inhibitor) has a grade 3 and above rash and/or permanent discontinuation incidence of 4.2%, whereas the incidence for grade 3 and above rash for acalabrutinib, a Bruton tyrosine kinase inhibitor, is 0.8%, a stark difference. Additionally, Stevens-Johnson syndrome and erythema multiforme were reported in 0.4% and 1.1% of patients undergoing treatment with alpelisib.
With increased FDA approval and clinical trial use of PI3K inhibitors for refractory chronic lymphocytic leukemia, relapsed follicular lymphoma, and breast, lung, and renal cancers, there is a need to characterize CAEs to help oncologists, dermatologists, and the support team anticipate and manage this impactful side effect.[6,14,15] This study aims to describe CAEs due to PI3K inhibitors and their management within a cohort of patients at a major cancer center.
The institutional pharmacy database was used to identify patients who were receiving PI3K inhibitor monotherapy or combination therapy from January 1, 2015, to May 19, 2019, and 135 patients were identified. Institutional review board approval for this study was granted under PA15-0959. Using International Statistical Classification of Diseases and Related Health Problems (ICD) 10 diagnosis codes for eczema, dry skin, pruritus, rash, and cellulitis, a list of 59 patients with possible cutaneous reactions to PI3K inhibitors was generated. This computer-generated database was used to perform a retrospective chart review of each patient to evaluate the association between the dermatologic diagnosis and PI3K therapy. Patients were included in the study if a study note or a provider attributed the rash to the PI3K inhibitor. Of the 59 patients, 11 had cutaneous toxicities correlated to treatment with a PI3K inhibitor. Each patient's CAEs were reviewed by two board-certified dermatologists (OP and ABP) to confirm association and identify the dermatologic diagnosis from the clinical presentation, explicit documentation, and histologic analysis when available. Unfortunately, photos of the rashes were not available to which to refer; however, good-quality pathology was available for three patients.
The institutional pharmacy database identified 135 patients who were treated with PI3K inhibitor monotherapy or combination therapy from January 2015 to May 2019. Of these 135 patients, 11 (8%) were identified as having cutaneous reactions to treatment with a PI3K inhibitor. Demographic characteristics of patients who did and did not experience a CAE are compared in Table 1. Although there was not a significant difference in sex between the two groups (p = 0.8), the difference in patients' ages was significant (p = 0.04). The 11 subjects had 12 cumulative CAEs with a median rash onset of 28 days after initiation of PI3K therapy. The median age of the subjects with a CAE was 61 years and the range, 32–81 years. The group was split, with four males and seven females. Eight of 11 subjects were Caucasian, and nine had diagnosis of stage 4 cancer. Characteristics of patients who had a CAE due to PI3K inhibitors are summarized in Table 2.
The median cumulative exposure to the inciting drug was 124 days. To be eligible for most PI3K inhibitors at our institution, patients must have experienced failure of first-line and second-line treatments. The group of patients receiving PI3K monotherapy versus combination therapy was split in half. The most common combination for polytherapy was gedatolisib (pan-class I isoform) plus palbociclib (CDK4/6 inhibitor) plus letrozole or fulvestrant (36%). More than half of the CAEs associated with PI3K inhibitors were due to idelalisib (27%) and gedatolisib (27%).
The most common CAEs were eczematous dermatitis (3), uncharacterizable (3), and morbilliform eruption (2). Other CAEs included erythroderma (1), cellulitis (1), psoriasiform dermatitis (1), and granulomatous dermatitis (1). Two patients (18%) had an inflammatory skin condition as part of their past medical history but the patients' CAEs were not related to a past/ongoing skin condition. CAE characteristics are summarized in Table 3.
For 11 patients, a dermatologist was consulted during 6 instances. Standard-of-care biopsies were obtained for 3 of those 6 visits, and pathology results were incorporated into the clinical diagnosis. A summary of the data from these 11 patients is provided in Table 4.
There was relatively high treatment impact, with one patient terminating their treatment due to severe CAE and 4 patients delaying their treatment. Of the 4 patients, 3 patients permanently decreased PI3K inhibitor dose, and 1 patient eventually discontinued PI3K inhibitor therapy. Seven of the 11 CAEs were successfully treated with triamcinolone 0.1% (58%), and three patients were supplemented with oral antihistamines for pruritus. Two patients were given intravenous antibiotics for cellulitis and erythroderma, and one was given oral antibiotics. Median time to improvement was 27 days after initiation of treatment intervention. None of the CAEs worsened after dermatologic intervention and all stabilized (8%), improved (25%), or resolved (67%). This signifies that CAEs due to PI3K inhibitors can be effectively treated, commonly with topical steroids and antihistamines, and for more severe CAEs, dose reduction or delay.
The most common extradermatologic reactions were diarrhea, nausea, and fatigue; all reactions are recorded in Table 2.
The incidence of CAEs in patients undergoing therapy with PI3K inhibitors at this institution was 8%, but this may be an underestimate as it is possible for lower-grade rashes to be underreported in the real-world setting. The median age difference between patients with (61 years) and without (70 years) CAEs was significant and should be further studied. It is possible that younger patients may have a more reactive immune system or may be more concerned with the cosmesis of CAEs than older patients. Although most patients' CAEs were easily managed with mild intervention, they also had a significant impact on dose compliance—33% of the CAEs caused a dose delay and 25% caused a dose delay and reduction. One subject (8%) had to immediately discontinue their PI3K inhibitor due to severe CAEs characterized as erythroderma, and one patient discontinued after a delay and dose reduction due to an eczematous eruption. Erythroderma is a severe and potentially life-threatening inflammatory condition whereby more than 90% of the body surface area is affected by scaling and redness. Eczematous rashes can vary in severity but are clinically characterized as intensely pruritic erythematous rough plaques. There were no occurrences of erythema multiforme or Stevens-Johnson syndrome in the study.
Of the four patients who had a delay in treatment, three consulted with dermatologists. Three patients who experienced a delay and decrease in dose were prescribed only topical steroids, usually medium potency. Further, the patient who immediately stopped treatment did not see a dermatologist. One could argue that more aggressive therapy for the CAEs should have been used in an attempt to mitigate dose impact. Perhaps earlier collaboration with dermatology or collaboration with dermatologists who have expertise in this area could have prevented some of these dose impacts. The incongruence between the high-dose impact and low potency of treatment suggests that patients were not treated aggressively enough to minimize CAEs and continue PI3K treatment.
The median time of 28 days to rash onset is consistent with previous studies that state CAEs usually appear within the first 2 months of starting treatment. A previous study of rashes associated with mTOR inhibitors stated most patients developed CAEs within the first month. All patients who continued therapy in this study had resolution, improvement, or stabilization of their rash. Patients who were rechallenged with the same or a lower dose did not have a relapse of the rash. This could be due to a dose-dependent effect of the PI3K inhibitor and/or to the continuation of treatments for the rash, which could have prevented a second flare.
Management recommendations for most patients, based on this study, include medium-potency topical steroids and nondrowsy oral antihistamines. For more severe CAEs that may require dose impact, earlier and more aggressive treatment may prevent this occurrence. The most common treatment, similar to that of other cancer therapies, was topical corticosteroids. If pruritus was present, a nondrowsy antihistamine was prescribed. This treatment protocol is consistent with available literature. Although oncologists and the cancer care team are well advised to treat mild to moderate CAEs with topical steroids, consultation with a dermatologist can be helpful in narrowing and tailoring treatment options best suited for the more severe CAEs. Acneiform rashes are often treated with topical or systemic antibiotics but over time can develop resistance and should be adequately monitored. Treating preexisting skin conditions such as eczema and psoriasis may prove to be difficult without interfering with the mechanism of the cancer therapy, and a dermatologist can be helpful in treating CAEs and preventing PI3K inhibitor interruptions before CAEs worsen and cause impairment.
Study limitations include this being a retrospective study using ICD-10 codes for rashes, which skews to more severe CAEs in appearance to the oncology team, but may not be reflected in higher Common Terminology Criteria for Adverse Events (CTCAE) scores. It is possible to have missed CAEs because ICD-10 codes were used to identify study patients. Additionally, the lack of a control group prevented characterization of nonrash patients and inability to identify risk factors for developing CAEs. Chart diagnosis was the primary mode of data extraction, as photographs were not available. Larger prospective studies have the ability to address these limitations and provide a more comprehensive analysis of CAEs due to PI3K inhibitors and their management. This study is however a composition of standard-of-care and trial patients and reflects a real-world population; it highlights the variety of CAEs and their significant dose impact. Additionally, the results have a sizable percentage of uncharacterized CAEs owing to a low number of dermatology consults and minimal documentation. Based on the low-medium potency therapies used for the CAEs and insufficient dermatology consults but high-dose impact, this study suggests the need for dermatologic involvement in management and possibly more aggressive CAE treatment to decrease dose impact, and future prospective studies.
CAEs associated with PI3K inhibitors have the potential to cause significant dose delay, reduction, or discontinuation. This study expands on trial literature and further characterizes CAEs to aid dermatologists and oncologists in their treatment plans and patient counseling. Managing CAEs associated with PI3K inhibitors is an interdisciplinary effort and an important part of helping patients continue their lifesaving treatment.
We would like to thank the dermatology department at University of Texas MD Anderson Cancer Center for their collaboration.
Source of Support: None. Conflict of Interest: Anisha B. Patel reports advisory board work with Novartis regarding guideline development for rash management from their PI3K inhibitor. She does not prescribe this medication or any other PI3K inhibitor. The other authors have no conflicts of interest.