PD1 monoclonal antibodies, pembrolizumab (Keytruda^®, Merck, Kenilworth, NJ, USA), and nivolumab (Opdivo^®, Bristol-Myers Squibb, New York, NY, USA) are now approved for metastatic melanoma, with superior response rates and lower adverse event rates compared to ipilimumab [73, 74]. As a consequence, PD1 inhibitors may supplant CTLA-4 inhibitors as monotherapy for first-line treatment for metastatic melanoma. Indeed, responses rates (30–35% vs. 12%), progression-free survival, and overall survival favored pembrolizumab over ipilimumab, when compared head-to-head in a randomized phase III trial, using two different schedules of pembrolizumab compared to ipilimumab (given at standard dose and schedule of 3 mg/kg every 3 weeks for 4 doses) (Keynote-006) [74]. The toxicity profile of pembrolizumab was also favorable compared to ipilimumab with nearly half the number of grade 3–5 adverse events (10–13% vs. 19.2%). In a phase 3 study comparing nivolumab with dacarbazine in previously untreated melanoma, for instance, rare neurologic complications included headache (4.4%), paresthesias (1%), vertigo (1%), and hypophysitis (0.5%) [74]. Case reports highlight rare irNAEs including vasculitis [100], and encephalitis [101].

Notwithstanding the somewhat more favorable irAE profile, response rates to monotherapy with either PD-1 or CTLA-4 inhibitors remain <50%, prompting evaluation of dual immune checkpoint blockade and several combined approaches with targeted therapies, radiation, and chemotherapy. Two recent studies evaluating concurrent immunotherapy with CTLA-4 and PD1 inhibition reported improved response rates compared to monotherapy with either ipilimumab or nivolumab [73, 102]. While overall survival data are not yet available, the rate of severe (grade 3 or 4) adverse events was >50% leading to a 30–40% discontinuation rate in the combination setting, although irNAEs were not addressed. Future studies will address the issue of whether sequential immune checkpoint inhibition is comparable to concurrent therapy, but these studies underscore the importance of evaluation of adverse event monitoring with single-agent immunotherapies, combinations of immune checkpoint inhibitors, and the emerging combinations of immune checkpoint inhibitors with targeted agents or traditional chemotherapy. The adverse event rate in patients treated with nivolumab following prior treatment with ipilimumab was 18–33%, which is lower than seen with combined therapy but higher than reported rates of irAEs with nivolumab in the first line [103, 104]. There is a paucity of data regarding the side effect profile of patients receiving ipilimumab following progression on nivolumab or pembrolizumab, although there are some recent reports of early and severe toxicities following sequential immune checkpoint inhibition [105].

Advances in our understanding of cancer immunology are leading to rapid FDA approval of novel cancer treatments, harnessing the power of the immune system to fight cancer. While most of these immune therapies are well tolerated due to the exquisite self-regulation of the immune system, immune-related adverse events, including neurological side effects, can occur. Of the agents discussed here, vaccines cause the least neurologic complications, since they are both the most specific and allow for immune self-regulation. With global activation of the immune system, such as using IL-2 or checkpoint inhibition, rare autoimmune neurological complications can be seen such as myasthenia gravis, neuropathies, plexopathies, cranial nerve involvement, to name but a few. While irNAEs are rare with checkpoint inhibitors, they can cause neuroendocrine complications in the form of autoimmune hypophysitis, and rare but dangerous conditions such as Guillain–Barre syndrome (GBS). Cytokines such as IL-2 and especially IFN cross the blood–brain barrier and have CNS effects including neuropsychiatric changes ranging from depression to confusion. Some immunotherapies such as the anti-angiogenic and immune-active bevacizumab (Avastin) have neurotoxicities directly related to their mechanism of action, such that bevacizumab presents with increased risk of bleeding (intracranial hemorrhage), clotting (stroke), as well as complications related to microvascular dysfunction (PRES). CD19 targeting methods such as BiTEs or CD19 CAR T cells are the most neurologically active and can cause confusion, ataxia, encephalitis, aphasia, and neuropsychiatric complications; some neurotoxicities can be severe enough to force medication discontinuation. Many in the field are looking toward combination therapies either combining multiple immunotherapies, or combining immunotherapy with targeted, biological, or cytotoxic treatments with the hope of increased responses and durability. These trials have just begun, and it will be important to see how this changes the landscape of neurotoxicity in cancer treatment.