Treatment holidays in patients with Erdheim-Chester disease receiving vemurafenib: a prospective pilot study

TO THE EDITOR:

Erdheim-Chester disease (ECD) is a rare non-Langerhans cell histiocytosis driven by mutations of MAPK pathway genes (eg, V-Raf Murine Sarcoma Viral Oncogene Homolog B [BRAF], MAP2K1).^1,2 Its clinical spectrum ranges from organ-limited to diffuse, life-threatening forms.^3-5 The prognosis of ECD was considered to be poor, but recent targeted treatments (ie, MAPK inhibitors, such as BRAF inhibitors [BRAFi] and mitogen-activated protein kinase kinase [MEK] inhibitors [MEKi]) greatly improved patients’ outcomes.^6-8 

BRAFi/MEKi have, however, significant toxicity, which sometimes limits their use.^9,10 Data from the long-term outcome after vemurafenib study indicate that definitive BRAFi withdrawal after achieving a metabolic response results in disease relapse in up to 75% of patients.⁵ Therefore, effective strategies to mitigate toxicity while maintaining disease control are needed. In this regard, different approaches have been proposed, including gradual dose reduction after achieving long-term remission, the addition of anticytokine drugs, or the use of better-tolerated alternative therapies (eg, mTOR inhibitors).¹¹ 

ECD is usually a slowly progressive disease. We hypothesized that short periods of therapy would not let ECD flare and designed a therapeutic protocol including “treatment holidays,” that is short and periodic intervals of BRAFi discontinuation, used for patients with sustained metabolic responses. A treatment holiday-based approach has already been used for other oncologic diseases such as colorectal and thyroid cancer with different drug types (eg, capecitabine/bevacizumab, lenvatinib).^12-14 Our protocol featured treatment pauses of 3 out of 12 weeks (9 weeks on, 3 weeks off; Figure 1A), which allowed a 25% dose reduction. Herein, we report the efficacy and safety of our schedule in 11 consecutive patients prospectively recruited at Meyer Hospital, Florence, Italy.

Patients with ECD were diagnosed according to the latest consensus recommendations.⁶ All had a biopsy-proven diagnosis and received vemurafenib as a single agent at a stable dose (480 mg twice daily in 8 cases and 240 mg twice daily in 3). They were started on the protocol only after achieving metabolic responses, defined either as complete, that is, the disappearance of fluorodeoxyglucose uptake on positron emission tomography–computed tomography (PET-CT) or partial, that is, a decline of ≥30% in 18F-fluorodeoxyglucose uptake of lesions on PET-CT and improvement of disease-related symptoms. These responses had to be sustained, that is confirmed on at least 2 occasions with a time interval of >6 months. Once the patients started the protocol, their daily vemurafenib dose had to be kept unchanged unless toxicity or lack of efficacy occurred.

To monitor disease activity during the treatment holiday protocol, all patients underwent routine clinical, laboratory, and imaging evaluations. PET-CT was performed before starting the protocol, 6 to 12 months after initiation, and every 12 months thereafter (Figure 1A). We also explored whether the protocol was associated with an improvement in quality of life, self-reported systemic symptoms, and drug-induced toxicity; thus, we administered the functional assessment of cancer therapy-general (FACT-G) Quality of Life questionnaire, ECD-Symptom Scale, and National Cancer Institute’s Patient-Reported Outcomes version of the Common Terminology Criteria for Adverse Events (NCI-PRO-CTCAE) to compare patient-reported outcomes (PROs) on-therapy vs off-therapy. The questionnaires were first administered at the end of the second "on" and at the end of the second "off" treatment cycle (Figure 1A). Afterward, they were administered twice yearly. The FACT-G Quality of Life is a 27-item questionnaire designed to measure 4 domains of Quality of Life (physical well-being, social/family well-being, emotional well-being, and functional well-being). The ECD-Symptom Scale measures symptom frequency, rated on a 5-point Likert scale (never, rarely, occasionally, frequently, and almost constantly), and severity, rated on a 0 to 10 scale.¹⁵ The NCI-PRO-CTCAE ITEMS is widely used to evaluate symptomatic toxicities by self-report in patients receiving cancer therapy. The questionnaires are available in the supplemental Material.

Supplemental Table 1 reports the main patient characteristics. All the patients had a BRAF^V600E mutation and long bone involvement. Central nervous system and peri-renal infiltration were detected in 8 patients (73%) each. Cardiac involvement was detected in 5 patients (46%), whereas 4 (36%) had coated the aorta. Other manifestations, including lung, endocrine, skin, and soft tissue infiltration, were less frequent.

All patients had been on vemurafenib monotherapy before treatment holiday initiation for a median of 34 months (interquartile range [IQR], 21-51) (supplemental Table 1). At the time of protocol initiation, 8 patients (73%) experienced a complete metabolic response, and 3 (27%) experienced a partial metabolic response. All patients completed at least 3 treatment holiday cycles; their median time on the protocol was 23 months (IQR, 11-33). None of the PET-CT evaluations performed during the protocol showed disease reactivation (from complete response) or metabolic progression (from partial response) (Figure 1B). The patients did not manifest any new disease-related symptoms during the treatment holiday protocol, and all patients remained on the protocol until the last follow-up (Figure 1C). C-reactive protein levels were undetectable before protocol initiation and remained normal thereafter. Furthermore, cell blood count, creatinine, and echocardiographic evaluations did not disclose signs of disease progression.

Based on PROs, the protocol led to a reduction in toxicity, without resulting in ECD-related symptom worsening. Although the improvement in toxicity was of borderline significance (median values of NCI-PRO-CTCAE scores on- and off-treatment 19 [IQR, 15.5-22.5] and 10 [IQR, 7.5-17]; P = .094), most patients (n = 8, 72%) reported an improvement in therapy-related side effects during off-therapy according to the NCI-PRO-CTCAE. Major improvements were observed in constitutional symptoms (nausea and fatigue) and skin toxicities, particularly photosensitivity. The total FACT-G score improved during treatment holidays in 4 patients (36%) and remained stable in 7 (64%); no significant differences emerged between on- and off-treatment (median values 56 [IQR, 52.5-57] and 56 [IQR, 53-57.5]; P = .947). Finally, the ECD-Symptom Scale did not show any significant change (the median score was 20 in both the on- and off-treatment periods; P = .921); in 1 case, an improvement in constitutional symptoms was reported, whereas the score was stable in the remaining 10 (91%). The results of the questionnaires administered beyond month 6 remained stable (data not shown). The changes in adverse events during the on- and off-treatment periods are detailed in supplemental Table 2.

This study represents, to our knowledge, the first systematic treatment strategy aimed at reducing the dose and, consequently, the toxicity of vemurafenib in patients with ECD while preserving its efficacy. Over the last decade, based on the results of the long-term outcome after vemurafenib study, which showed a 75% relapse rate at a median of 6 months after treatment withdrawal,⁹ there has been substantial reticence in the ECD community to reduce/discontinue vemurafenib. In our study, patients underwent periodic PET-CT evaluations, which showed that the metabolic response was maintained in all cases. We also evaluated the potential benefits of this approach using PRO measures and observed that our protocol was associated with an improvement in quality of life, symptoms, and drug toxicity.

Our protocol was deliberately designed to ensure patient safety, with a relatively short duration of the off-therapy periods. However, an extension of these periods could be considered for patients with long-lasting remission and less severe clinical manifestations. Of course, this is only one of the potential strategies, which, however, has the advantage of adapting the off-therapy periods to the patient's needs and guaranteeing an overall reduction of the medical expense.

This pilot study carries limitations, mainly due to the low number of patients and the relatively short follow-up period, which did not allow us to exclude late disease flares. However, our results are promising and merit validation in larger cohorts. The modulation of the frequency and duration of treatment holidays and the identification of the best candidate patients for this approach represent future goals.

In conclusion, we provide preliminary evidence that a treatment holiday protocol can reduce vemurafenib-associated toxicity and overall drug exposure without impairing efficacy.

This study was approved by the Meyer local ethics committee in accordance with the Declaration of Helsinki.

Contribution: F. Pegoraro and A.V. conceived the study; F. Pegoraro and F.C. collected the data and drafted the manuscript; A.V. critically revised the manuscript; F. Pegoraro, F.C., E.L.D., and A.V. contributed to designing the manuscript; F. Pegoraro, F.C., F.A., G.P., F. Peyronel, and A.V. followed the patients; and all authors read and approved the final version of the manuscript.

Conflict-of-interest disclosure: E.L.D. discloses unpaid editorial support from Pfizer, Inc and paid advisory board membership with Opna Bio, both outside of the submitted work. The remaining authors declare no competing financial interests.

Correspondence: Augusto Vaglio, Dipartimento di Scienze Biomediche, Sperimentali e Cliniche “Mario Serio”, Università di Firenze, Viale Pieraccini 6, 50139 Firenze, Italy; email: augusto.vaglio@unifi.it.