Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
J Neurooncol DOI 10.1007/s11060-016-2190-1 CLINIcAL STUDY Pembrolizumab: first experience with recurrent primary central nervous system (CNS) tumors Deborah T. Blumenthal1 · Michal Yalon2 · Gilad W. Vainer3 · Alexander Lossos4 · Shlomit Yust5 · Lior Tzach6 · Emanuela Cagnano3 · Dror Limon5 · Felix Bokstein1 Received: 4 May 2016 / Accepted: 23 June 2016 © Springer Science+Business Media New York 2016 Abstract Patients with progressive primary brain tumors (PBT) are attracted to promising new treatments, even prior to convincing data. Anti-PD1 immunotherapies have been in the spotlight since publication of groundbreaking results for metastatic melanoma with pembrolizumab (PBL). Our objective was to report on the response and toxicity of PBL in patients with advanced PBT. We retrospectively reviewed the charts of 22 patients (17 adults and 5 children) with recurrent central nervous system tumors treated with PBL. We analyzed prior antineoplastic therapies, steroid usage, and outcomes. Patients received a median of two neoplastic therapies prior to PBL, and a median of three infusions of PBL in adults and four in children. Twelve patients (9 adults and 3 children) started PBL on steroids (median dose in adults 4 mg; range 2–8, and in children 1.5 mg, range 0.5–4) and five patients received steroids later during PBL treatment. Twelve patients (10 adults and Deborah T. Blumenthal [email protected] 1 Neuro-Oncology Service, Division of Oncology, Tel Aviv Sourasky Medical Center, 64239 Tel Aviv, Israel 2 Pediatric Neuro-Oncology Service, Pediatric HematoOncology Department, Chaim Sheba Medical Center, 52621 Tel HaShomer, Israel 3 Molecular Pathology Service, Department of Pathology, Tel Aviv Sourasky Medical Center, 64239 Tel Aviv, Israel 4 Leslie and Michael Gaffin Center for Neuro-Oncology, Department of Oncology, Hadassah-Hebrew University Medical Center, 91120 Jerusalem, Israel 5 Oncology Institute, Davidoff Center, Rabin Medical Center, 49100 Petah Tikva, Israel 6 Oncology Institute, Chaim Sheba Medical Center, 52621 Tel HaShomer, Israel 2 children) received concomitant bevacizumab with PBL. Side effects were minimal. All patients showed progressive tumor growth during therapy. Median OS from the start of PBL was 2.6 months in adults and 3.2 months in children. Two GB patients underwent tumor resection following treatment with PBL. Tumor-lymphocytic response in these cases was unremarkable, and PD-L1 immunostaining was negative. In this series of 22 patients with recurrent primary brain tumors, PBL showed no clinical or histologic efficacy. We do not recommend further use of PBL for recurrent PBT unless convincing prospective clinical trial data are published. Keywords Primary brain tumor · Anti-PD1 · Pembrolizumab · Glioblastoma · Immunotherapy Introduction Glioblastoma (GB) is the highest grade primary brain tumor, with average median survival of 12 months [1]. Most patients receive standard treatment of radiation with concurrent chemotherapy of temozolomide, followed by adjuvant temozolomide. Approved second-line therapies for G B include intravenous bevacizumab [2] and carmustine wafers [3] (the latter, if surgical gross total resection of the recurrence is feasible). Re-irradiation may add benefit to selected cases [4, 5] and ongoing studies (RTOG 1205) continue to evaluate the benefit of this approach. However, the utility of these secondline regimens is limited and often the tumor progresses despite interventions, proving resistant to further salvage therapies. The treatment options for brain stem gliomas (BSG) are even more limited than those for GB. These precariously located intrinsic and infiltrative tumors are not amenable to resection although stereotactic biopsy may be feasible [6]. 13 2 Patients with progressive primary brain tumors (PBT) are in desperate need of effective therapies. Many of these patients continue to have independent level of function in the face of progressive tumor on imaging. Such patients turn in desperation to new “wonder drugs” despite lack of evidence of efficacy, due to lack of viable treatment options. Immunotherapy with anti-programmed cell death protein 1 (PD1) agents has resulted in remarkable responses in a significant percentage of cases of refractory metastatic melanoma [7]. This immune-modulating therapy has also shown a signal in certain head and neck, renal cell, and nonsmall cell lung cancer cases [8–10]. To date, there are no clinical studies published reporting the response of PBT to anti-PD1 therapy. We report our experience with a series of patients with progressive PBT who were treated with pembrolizumab (PBL), a monoclonal anti-PD1 antibody. Materials and methods We retrospectively collected data from 22 patients with progressive, advanced primary CNS tumors who were treated at four major Israeli brain tumor centers with PBL as a salvage agent. Identifying demographic data were blinded between the centers’ investigators, to protect the patients’ privacy. Response to treatment was assessed by clinical evaluation by the treating physicians, and by standard MRI examination at 2–3 month intervals or sooner if clinical deterioration warranted. Since the cohort was analyzed retrospectively, the imaging intervals were not uniform for all patients. Patients for whom there was a question of tumor progression versus inflammatory response on MRI, were continued on anti-PD1 therapy until either clinical deterioration; appearance of new lesions; or tumor growth of greater than 25 % on a subsequent scan. The study was approved by all centers’ institutional review boards. Two patients had tumors resected at the time of recurrence following treatment with PBL. These two samples were examined for lymphocyte staining and PD-L1. We utilized an FDA-approved 22C3 PD-L1 staining technique (DAKO) [11]. Five patients had tumors analyzed molecularly for microsatellite instability (MSI) and mutational load. Results Patient characteristics and results (Table 1) Our series of 22 patients included seventeen adults with 10 (G Bs), 2 anaplastic astrocytomas (AA), 2 anaplastic 13 J Neurooncol oligodendrogliomas, 2 patients with (transformed) low grade gliomas, and one brain stem glioma (BSG ); and 5 children: 2 BSG ; and 1 each of G B, atypical teratoid/ rhabdoid tumor, and medulloblastoma. Most adult patients were heavily pretreated: mean number of previous treatment lines in adults was 2 (range 1–6); 8/17 adults (47.1 %) received 2 treatment lines and 5/17 patients (29.4 %) received three and more treatment lines. In children, median number of treatment lines was 1 (range 1–2). The median dose of PBL was 150 mg administered every 3 weeks (range 100–200) in adults. All pediatric patients received PBL at the flat dose of 50 mg every 3 weeks. Twelve patients started PBL while on steroids (median dose in adults 4 mg [range 0.5–8 mg] and median dose in children 1.5 mg [range 0.5–4]), and 5 patients (3 adults and 2 children) received steroids later during PBL treatment due to clinical deterioration. Twelve patients (10 adults and 2 children) received concomitant bevacizumab (BVZ) with PBL. Patients who received therapy with BVZ and PBL were weaned off steroids or tapered to a minimal dose of 2 mg. All patients showed progressive tumor growth during therapy. Median number of PBL infusions was 3 (range 1–7) in adults and 4 (range 2–10) in children. Median overall survival (OS) from the initiation of treatment with PBL was 2.6 months (range 0.4–11.6) in adults and 3.2 months (range 2.3–7.9) in children. There were two cases of side effects which may have possibly been related to PBL treatment: one patient suffered a mild rash and increased liver enzymes; and a second patient experienced protracted diarrhea. Two G B patients underwent resection of recurrent tumor following treatment with four and two cycles, respectively, of PBL. Hematoxylin & eosin (H&E) and immunohistochemistry stains, including stains for lymphocytes and PD-L1 are shown in Fig. 1 for these two patients. Immunohistologic examination showed no evidence of local tumor-lymphocytic response in these two cases. Programmed death-ligand 1 (PD-L1) immune-staining following resection for recurrent tumor was as well negative in both cases. Five of our adult patients had sent tissue during the course of their treatment to FoundationOne (FoundationOne™, Foundation Medicine Inc., Cambridge, MA) for genomic sequencing of their tumor samples [12]. We retrospectively reviewed these cases for mutational load and microsatellite MSI. Analysis of these samples showed stable MSI with low mutation burden in four of the five cases (3–5 mutations per megabase [Mb]); these four cases involved tissue from the time of tumor diagnosis/initial surgery. However, one case (depicted in Fig. 1a–c) whose tumor sample was examined at the time of recurrence, showed stable MSI with a very high mutation rate of 75 per Mb. J Neurooncol 3 Table 1 Patient characteristics (n = 22) Total number Sex (f/m) Median age Pathology Glioblastoma Anaplastic astrocytoma Anaplastic oligodendroglioma Low grade glioma (probably transformed) Diffuse brainstem glioma Medulloblastoma Atypical rhabdoid-teratoid tumor No of previous treatment lines One Two Three and more Median Radiation treatment in the past Median dose of PBZ, mg Median number of PBZ cycles No of patients treated concomitantly with bevacizumab Steroids No of patients starting PBZ without steroids No of patients starting PBZ while on steroids No of patients starting steroids later on Median dose of dexamethasone (mg) Treatment response Median OS from start of PBZ, months [range] Discussion Immune-based, “checkpoint” therapies have created a revolution in the field of oncology over the recent years. With justification, these therapies have produced previously unbelievable results in patients with advanced metastatic melanoma and other cancers [7, 10, 11, 13–17]. The efficacy of anti-PD1 therapy in a variety of other tumor types is currently under investigation [18, 19]. Therapies targeting programmed cell death protein 1 (PD-1) work by facilitating the body’s innate immune defenses to work against the tumor. To evade immune recognition, tumors utilize various mechanisms. One of the key interactions between the tumor and the host immune system involves the signaling of PD-1 after binding to PD-L1. This happens when activated B and T-cells expressing the PD-1 receptor encounter tumor cells expressing PD-L1 on their surface [20]. The binding of PD-1 to PD-L1 generates an immunosuppressive cascade which allows the tumor cell to evade destruction by the innate immune system [21]. Anti-PD-1 and PD-L1 therapies prevent the tumor cell from Adults Children 17 3/14 38.5 [25–71] 5 All males 5 [3–7] 10 2 2 2 1 – – 1 – – – 2 1 1 4 8 5 2 [1–6] All 150 [100–200] 3 [1–7] 10 4 1 – 1 [1, 2] All 50 (all) 4 [2–10] 2 8 9 3 4 [2–8] PD—all 2.6 [0.4–11.6] 2 3 2 1.5 [0.5–4] PD—all 3.2 [2.3–7.9] masking itself from the immune system, and allow the lymphocytes to target the foreign cancer cell invader. Anti-PD1 and immunotherapy in primary brain tumors Pre-clinical data Pre-clinical data suggests a role for anti-PD1 therapy in glioma, specifically with synergistic radiation. Radiation has been shown to increase antigen presentation and promote a pro-inflammatory tumor microenvironment which could work synergistically to complement anti-PD1 therapy. A mouse glioma line was treated with anti-PD1 and stereotactic radiosurgery, as single or combined modalities, to test this theory. Results of this study showed a significant increase in survival in the combined modality group, with long term survivors seen in the combined modality cohort alone. Moreover, a correlating immune response was seen in the orthotopic tumor with cytotoxic T cell infiltration (increased CD8/CD4 ratio) in the combined therapy cohort [22]. 13 4 J Neurooncol Fig. 1 a–f Two patients with recurrent GB tumor samples following treatment with pembrolizumab (PBL). a–c CD8 (lymphocyte), Hematoxylin & Eosin (H&E), and PD-L1 immunohistochemistry stains respectively, in a patient treated with four cycles of PBL. HE shows an active glioblastoma with regions of pseudopallisading necrosis. Stains for CD8 lymphocytic infiltrate as well as PD-L1 are negative. This same patient showed a very high mutational load on genomic analysis. d–f Similar staining in a patient who underwent resection following treatment with two cycles of PBL. Both cases are notable for viable glioblastoma tumor on HE with scant to no lymphocytic infiltration; PD-L1 staining is negative. g Positive control slide (tonsil) for PD-L1 Clinical data in combination with ipilimumab, versus bevacizumab. The primary endpoint will be overall survival, with secondary endpoints of safety, PFS and ORR [24]. There is limited clinical data for anti-PD1 therapies in the context of PBT. Recent presentations from the American Society of Clinical Oncology (ASCO) 2015 annual meeting reported primarily safety data from an ongoing prospective study of nivolumab for recurrent GB patients. Twenty patients (ten per arm) were treated with either human IgG 4 PD1 antibody nivolumab alone or in combination with ipilimumab in a small randomized trial for GB at first recurrence. The study endpoints were tolerability and safety. Side effects were more severe and more common in the combination arm. Overall survival at 6 months was 75 % for the two groups, similar to historical standards. Final data are pending [23]. An ongoing phase IIB study for patients with first recurrence of GB will treat 240 patients with nivolumab alone or 13 Our cohort Responses in other settings of once thought “intractable disease” have proven the worth of anti-PD1 treatment. However, the efficacy of these therapies has not been proven in all oncologic settings, and many tumor types have not been shown to benefit from the anti-PD1 mechanism. The logic supporting PBT as an appropriate target for anti-PD1 therapy is theoretically sound for some cases of primary brain tumors; however, the clinical data for efficacy is lacking, and certainly not expected to be as dramatic as in the case of metastatic melanoma, if at all. J Neurooncol None of our 22 patients with recurrent PBT showed radiologic, clinical, or immune-histologic response to treatment with PBL. There are several possible explanations for the lack of response in our cohort. First, is the issue of molecule size and blood brain barrier (BBB) penetration. The BBB may mechanically prevent the large PBL molecule (146 kDa) from entering the CNS and approaching the target cells. Most molecules which permeate by transmembrane diffusion are close to 1000 times smaller than the PBL antibody [25]. Despite this theoretical concern of BBB penetration, anecdotal reports of patients with brain metastases report robust reactions of their central nervous system tumors to treatment with PBL, including histologic evidence of inflammation and microglial response [26]. As such, it is likely that the BBB concern is not the primary reason for lack of a clinical response to anti-PD-1 treatment. Secondly, the innate immune system and anti-tumor immune response which the anti-PD1 therapy potentiates, is not the same in the CNS as in the systemic periphery [22]. Systemic melanoma has long been addressed by immune therapies (alpha interferon, IL) and may be as such the ideal target for anti-PD1 therapy. The presence of CD8 + T-cells in tumor tissue has been associated with successful tumor response to anti-PD1 therapy in melanoma [27, 28]. We did not see evidence of T-cell infiltration in the two tumor cases in our cohort sampled after treatment with PBL. Furthermore, tumors which are more immunogenic (i.e., melanoma) have a higher MSI. Compared to other human cancers, the prevalence of somatic mutations in GB (and medulloblastoma) is on the lower end of the spectrum, certainly much lower than in tumors with high MSI such as melanoma or lung adenocarcinoma [29]. As such, these PBT would be expected to be less responsive to anti-PD1 therapies than tumors with higher scores. There may be selected subgroups of gliomas which may have higher MSI scores, perhaps due to mismatch repair mechanisms (MMR). A cohort of consanguineous pediatric gliomas has been reported, with an MMR mutation which may render them an ideal population for treatment with anti-PD1 therapy [30]. Mutational load has also been recognized as a predictor of response to anti-PD1 therapy. A recent report revealed a high mutational load in 37 pediatric patients (including GB) with a hereditary mismatch repair deficiency syndrome, who underwent exome sequencing; two siblings in this cohort were treated with anti-PD1 therapy with excellent clinical and imaging responses [31]. Mutations in the tumor can produce specific neoantigens which can elicit T cell responses to immunotherapy [32]. Despite a very high mutational burden (as might be expected in a recurrent tumor post-exposure to temozolomide) [33, 34] seen in one of the GBM tumor samples in our series that was examined at re-operation, this patient did 5 not respond clinically or radiographically to treatment with PBL. In the case of primary brain tumors, a “stable” MSI score may be a more significant negative predictive factor to anti-PD1 therapy than an increased mutational load of the tumor is a positive predictive factor. Thirdly, our patients were almost all heavily pre-treated, at an advanced and refractory stage of disease. Treating at an early stage of disease, and more specifically with synergistic radiation, has been shown to optimize effect of the anti-PD1 therapy in pre-clinical models [22, 35], and may be a superior strategy. Fourth, the use of corticosteroids, typically dexamethasone, is widespread in patients with advanced PBT for control of tumor-associated edema and preservation of neurologic functional status. Although there are reports of successful treatment with PBL during treatment of dexamethasone at a dose of 4 mg a day [36], it is still not known if steroid use may antagonize and diminish, if not nullify, the desired immune-enhancing effect of the anti-PD1 therapies. Some of our patients avoided or minimized steroid use by concurrent therapy with bevacizumab, to control tumor edema. The relatively high number of patients who were treated with dexamethasone during therapy with PBL may explain the low incidence of side effects we observed in this cohort; steroids may have prevented unwanted inflammatory responses in other organs. Lastly, we did not see expression of PD-L1 staining in the cases examined post-treatment, contrary to other reports. An analysis of PD-L1 gene expression in 446 GB cases from the Cancer G enome Atlas reported immunohistochemistry staining for PD-L1 expression in 70–90 % of GB cases, less in proneural and higher in mesenchymal subtypes. PD-L1 expression did not appear to correspond with age or prognosis, nor did it seem to change from initial diagnosis to recurrence [37]. A recent smaller study of 94 GB tumors showed PD-L1 positive staining in greater than 60 % of tumors, but most with a small percentage of cell surface expression. Expression of PD1 and PD-L1 in this study was found to be a negative prognostic factor for survival [28]. This discrepancy in the low incidence of PD-L1 expression in our cohort may be related to the small number of patients who underwent immunostaining, or it could be indicative of differences between our PD-L1 staining technique and others cited in the literature. Even if there had been PD-L1 expression in the tumors of our patients prior to treatment, we would expect to have seen some remnant of reactivity or inflammation at the time of reoperation following immunotherapy. Ethical viewpoint Patients with progressive GB who have exhausted the standard treatment options and for whom no feasible clinical 13 6 trials are available are desperate for a “life-line”. The ethics of providing unproven therapy to this population is wrought with controversy. Treatments may be potentially toxic, without the promise of proven benefit; possible toxic effects associated with treating with a developing drug off-label can deter the development of the product to make it available to the public who could benefit from it. The physician may find her/himself unduly pressured by the patient and family to prescribe an unproven therapy against better judgment, which begs the question of the need for increased institutional regulation. A modicum of skepticism regarding a new product is advised when data are sparse. However, physicians may differ in practice styles: while some practitioners do not treat in the absence of convincing evidence, certain physicians may feel uncomfortable denying a patient’s or family’s requests for off-label medications and are more influenced by the arguments of the patient or by unsubstantiated anecdotes, and acquiesce to these requests. Yet others seek to practice on the “cutting edge” and are more willing to pursue new therapies, even when data are lacking [38]. What is the accountability of the pharmaceutical companies to provide commercial medications if used off-label; is the physician, the hospital or health authorities or ultimately the “informed”, but far from objective patient culpable for the ultimate treatment decision? Some individuals turn to social media and groups of patients may petition for expanded access of new medications. It is not always clear if these requests come from adequately informed sources [39]. On the one hand, it is not always feasible to wait for level I evidence (phase III randomized trials) to gain access to new drugs for an “orphan” population; by waiting, lives may be compromised. However, on the other hand, there needs to be a mechanism to offer patients such vanguard treatments, requiring at least minimal clinical (even anecdotal) evidence for response, with acceptable toxicity data. Conclusion PBL is well-tolerated in patients with recurrent primary brain tumors, both in adults and in children. However, we did not observe a signal of efficacy for treatment with the anti-PD1 therapy in our patient population. High mutational load was not associated with response to PBL in our limited experience. We do not recommend further use of PBL for recurrent PBT unless convincing prospective clinical trial data is published. Acknowledgments We thank Phil Stephens, PhD and Joana Buxhaku from Foundation Medicine who assisted with the analysis of MSI and mutational number on available patient samples. Thank you to Jay A. Jacobson, M.D. of the University of Utah School of Medicine for medical ethics advice. We thank Anat Mordechai, RN from Hadassah Medical Organization for her assistance with clinical data retrieval. 13 J Neurooncol Funding This research was not funded by outside sources. Compliance with ethical standards Conflict of interest Gilad W. Vainer, MD, PhD served on advisory committees for Pfizer, Roche, and Merck (MSD) regarding immunologic and molecular testing. References 1. Stupp R, Mason WP, van den Bent MJ et al (2005) Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma. N Engl J Med 352:987–996 2. Friedman HS, Prados MD, Wen PY et al (2009) Bevacizumab alone and in combination with irinotecan in recurrent glioblastoma. J Clin Oncol 27:4733–4740 3. Westphal M, Hilt DC, Bortey E et al (2003) A phase 3 trial of local chemotherapy with biodegradable carmustine (BCNU) wafers (G liadel wafers) in patients with primary malignant glioma. Neuro Oncol 5:79–88 4. Bokstein F, Blumenthal DT, Corn BW et al (2016) Stereotactic radiosurgery (SRS) in high-grade glioma: judicious selection of small target volumes improves results. J Neurooncol 126:551–557 5. Fogh SE, Andrews DW, Glass J et al (2010) Hypofractionated stereotactic radiation therapy: an effective therapy for recurrent high-grade gliomas. J Clin Oncol 28:3048–3053 6. Cage TA, Samagh SP, Mueller S et al (2013) Feasibility, safety, and indications for surgical biopsy of intrinsic brainstem tumors in children. Childs Nerv Syst 29:1313–1319 7. Ribas A, Puzanov I, Dummer R et al (2015) Pembrolizumab versus investigator-choice chemotherapy for ipilimumab-refractory melanoma (KEYNOTE-002): a randomised, controlled, phase 2 trial. Lancet Oncol 16:908–918 8. Zandberg DP, Strome SE (2014) The role of the PD-L1:PD-1 pathway in squamous cell carcinoma of the head and neck. Oral Oncol 50:627–632 9. Tykodi SS (2014) PD-1 as an emerging therapeutic target in renal cell carcinoma: current evidence. Onco Targets Ther 7:1349–1359 10. Garon EB, Rizvi NA, Hui R et al (2015) Pembrolizumab for the treatment of non-small-cell lung cancer. N Engl J Med 372:2018–2028 11. Herbst RS, Baas P, Kim DW et al (2015) Pembrolizumab versus docetaxel for previously treated, PD-L1-positive, advanced non-small-cell lung cancer (KEYNOTE-010): a randomised controlled trial. Lancet 387:1540–1550 12. Frampton G M, Fichtenholtz A, Otto G A et al (2013) Development and validation of a clinical cancer genomic profiling test based on massively parallel DNA sequencing. Nat Biotechnol 31:1023–1031 13. Robert C, Schachter J, Long G V et al (2015) Pembrolizumab versus ipilimumab in advanced melanoma. N Engl J Med 372:2521–2532 14. Robert C, Long GV, Brady B et al (2015) Nivolumab in previously untreated melanoma without BRAF mutation. N Engl J Med 372:320–330 15. Weber JS, D’Angelo SP, Minor D et al (2015) Nivolumab versus chemotherapy in patients with advanced melanoma who progressed after anti-CTLA-4 treatment (CheckMate 037): a randomised, controlled, open-label, phase 3 trial. Lancet Oncol 16:375–384 16. Larkin J, Hodi FS, Wolchok JD (2015) Combined nivolumab and ipilimumab or monotherapy in untreated melanoma. N Engl J Med 373:1270–1271 17. Borghaei H, Paz-Ares L, Horn L et al (2015) Nivolumab versus docetaxel in advanced nonsquamous non-small-cell lung cancer. N Engl J Med 373:1627–1639 J Neurooncol 18. Seiwert TY, Burtness B, Mehra R et al (2016) Safety and clinical activity of pembrolizumab for treatment of recurrent or metastatic squamous cell carcinoma of the head and neck (KEYNOTE-012): an open-label, multicentre, phase 1b trial. Lancet Oncol. doi:10.1016/S1470-2045(16)30066-3 19. Moskowitz CH, Ribrag V, Michot JM et al (2014) PD-1 blockade with the monoclonal antibody pembrolizumab (MK-3475) in patients with classical Hodgkin lymphoma after brentuximab vedotin failure: preliminary results from a phase 1b study (KEYNOTE-013). In: 56th annual meeting of the American Society of Hematology, San Francisco, California 20. Freeman G J, Long AJ, Iwai Y et al (2000) Engagement of the PD-1 immunoinhibitory receptor by a novel B7 family member leads to negative regulation of lymphocyte activation. J Exp Med 192:1027–1034 21. Iwai Y, Ishida M, Tanaka Y, Okazaki T, Honjo T, Minato N (2002) Involvement of PD-L1 on tumor cells in the escape from host immune system and tumor immunotherapy by PD-L1 blockade. Proc Natl Acad Sci USA 99:12293–12297 22. Zeng J, See AP, Phallen J et al (2013) Anti-PD-1 blockade and stereotactic radiation produce long-term survival in mice with intracranial gliomas. Int J Radiat Oncol Biol Phys 86:343–349 23. Sampson JH, Vlahovic G , Solmaz SS et al (2015) Preliminary safety and activity of nivolumab and its combination with ipilimumab in recurrent glioblastoma (GBM): CHECKMATE-143. J Clin Oncol 2015:33 24. Sampson JH, Vlahovic G, Desjardins A et al (2014) Randomized phase IIb study of nivolumab (anti-PD-1; BMS-936558, ONO4538) alone or in combination with ipilimumab versus bevacizumab in patients (pts) with recurrent glioblastoma (G BM). J Clin Oncol 32:5 25. WM P (2001) Brain drug targeting: the future of brain drug development. Cambridge University Press, Cambridge 26. Cohen JV, Alomari AK, Vortmeyer AO et al (2016) Melanoma brain metastasis pseudoprogression after pembrolizumab treatment. Cancer Immunol Res 4:179–182 27. Tumeh PC, Harview CL, Yearley JH et al (2014) PD-1 blockade induces responses by inhibiting adaptive immune resistance. Nature 515:568–571 7 28. Nduom EK, Wei J, Yaghi NK et al (2016) PD-L1 expression and prognostic impact in glioblastoma. Neuro Oncol 18:195–205 29. Alexandrov LB, Nik-Zainal S, Wedge DC et al (2013) Signatures of mutational processes in human cancer. Nature 500:415–421 30. Amayiri N, Tabori U, Campbell B et al (2016) High frequency of mismatch repair deficiency among pediatric high grade gliomas in Jordan. Int J Cancer 138:380–385 31. Bouffet E, Larouche V, Campbell BB et al (2016) Immune checkpoint inhibition for hypermutant glioblastoma multiforme resulting from germline biallelic mismatch repair deficiency. J Clin Oncol 34:2206–2211 32. Schumacher TN, Schreiber RD (2015) Neoantigens in cancer immunotherapy. Science 348:69–74 33. van Thuijl HF, Mazor T, Johnson BE et al (2015) Evolution of DNA repair defects during malignant progression of lowgrade gliomas after temozolomide treatment. Acta Neuropathol 129:597–607 34. Cahill DP, Levine KK, Betensky RA et al (2007) Loss of the mismatch repair protein MSH6 in human glioblastomas is associated with tumor progression during temozolomide treatment. Clin Cancer Res 13:2038–2045 35. Ngiow SF, McArthur GA, Smyth MJ (2015) Radiotherapy complements immune checkpoint blockade. Cancer Cell 27:437–438 36. Rothermundt C, Hader C, G illessen S (2016) Successful treatment with an anti-PD-1 antibody for progressing brain metastases in renal cell cancer. Ann Oncol 27:544–545 37. Berghoff AS, Kiesel B, Widhalm G et al (2015) Programmed death ligand 1 expression and tumor-infiltrating lymphocytes in glioblastoma. Neuro Oncol 17:1064–1075 38. Schiff GD, Galanter WL, Duhig J, Lodolce AE, Koronkowski MJ, Lambert BL (2011) Principles of conservative prescribing. Arch Intern Med 171:1433–1440 39. Mackey TK, Schoenfeld VJ (2016) G oing “social” to access experimental and potentially life-saving treatment: an assessment of the policy and online patient advocacy environment for expanded access. BMC Med 14:17 13 8 13 J Neurooncol