The Severe Immunotherapy Complications (SIC) Service and Clinical-Translational Research Effort at Mass General Cancer Center is working to address the urgent need to understand how and why immune-related adverse events (irAEs) occur. SIC's multidisciplinary team comprises more than 50 individuals across 19 areas of the hospital, and this unique group of clinicians and researchers are involved in the care of every patient admitted with a suspected irAE after checkpoint blockade immunotherapy.

In this interview series, Kerry Reynolds, MD, director of the SIC service, sits down with a SIC team member for a deeper-dive into their work. This month we meet Amanda Guidon, MD, a neuromuscular neurologist at Massachusetts General Hospital, who has a clinical and research interest in the neuromuscular complications of cancer and its therapies, particularly the neurologic immune-related adverse events of cancer immunotherapies.

Kerry Reynolds, MD
Kerry Reynolds, MD

Kerry: Amanda, thank you so much for joining us today. We enjoyed your talk this month in the Severe Immunotherapy Complications Service seminar series. You and your research team have been doing very interesting work. The recent paper in JITC defining the spectrum of neurological immune-related adverse events (irAE-N) was really a tour de force. It is so critical to have accurate phenotyping when studying these events. Can you tell us more about it, and can you tell us about the neurology team that drafted the original definitions?

Amanda: Absolutely. You and I began thinking about this project about 3 years ago. Our goal was to improve diagnosis and characterization of neurologic immune-related adverse events from immune checkpoint inhibitor therapy. We wanted to develop a common language and accessible framework which everyone could use – neurologists, oncologists, advanced practice providers on the front lines, community and academic providers, and researchers. It became clear that to provide the most appropriate patient care and to advance the field, we needed a more robust classification system with less heterogeneity and less potential for misclassification.

Amanda Guidon, MD
Amanda Guidon, MD

At MGH, we have a great neurology irAE team working in this area. In the past year, we’ve expanded to include two neuromuscular neurologists (Dr. Leeann Burton and myself) and two neurologists with autoimmune neurology training (Dr. James Hillis and Dr. Bart Chwalisz, who also has neuro-ophthalmology training). The group met iteratively to write a complete draft of the definitions. Your input, Kerry, was also key in this process as the oncologist in the core group.

Once we had these draft definitions, we invited members of a larger multi-institutional and multi-specialty panel to participate. It was a modified Delphi consensus process, with two rounds of anonymous ratings by panelists and two meetings to discuss areas of controversy. We had representation from 27 clinicians and researchers from 15 major centers involved in this work. The non-profit Project Data Sphere provided support for this process.

The panelists voted on a total of 53 rating scales. They rated content for usability, appropriateness, and accuracy on 9-point scales in electronic surveys, and provided free text comments. Aggregated survey responses were incorporated into revised definitions. Consensus was based on numeric ratings using the RAND/University of California Los Angeles (UCLA) Appropriateness Method with prespecified definitions. Seventy-seven percent (41/53) received first round consensus. After revisions, all items received second round consensus. The consensus definitions and severity criteria were then aggregated and published. We have already started using them clinically and in our research!

Kerry: How do you think this work impacts the field of immune-related adverse events? How do you see the definitions evolving over time?

Amanda: Let me describe the definitions a little more. With the aim of simplifying and reducing unnecessary heterogeneity, the definitions start very broad and get more and more specific. The first distinction is whether the patient is having central of peripheral nervous system irAE. There is guidance in the paper about how to work through the framework for the non-neurologist. Patients can then be identified as having 1 of 7 core syndromes – the 4 central nervous system syndromes are irMeningitis, irEncphalitis, irDemyelinating disease and irVasculitis. The 3 peripheral nervous system syndromes are irNeuropathy, irNeuromuscular junction disorder or irMyopathy in the PNS. For each core syndrome, we identified 6 descriptors if this information is available: syndrome subtype, antibody association, level of certainty, severity, de novo presentation or pre-existing disease exacerbation and concurrent or overlapping irAEs. I think this paper helps clinicians understand and work through these complex toxicities, and hopefully better understand and treat what they are seeing at the bedside. We hope the tables and accompanying text is used as part of a clinical toolbox.

We’re learning more and more that early diagnosis and treatment of Neuro irAEs leads to better neurologic outcomes and maybe allows for more anti-cancer treatment options for patients. The heterogeneity now in in the literature currently regarding neuro irAE reporting and diagnosis makes data aggregation impossible. We hope to gain insight into true incidence and prevalence as well as natural history of irAEs with this common framework. It lays the foundation needed for prospective biomarker studies and treatment clinical trials.

This is a living document. We will study the performance of the definitions as well as plan for a revision in a few years with updates.

Kerry: What is the most pressing need, in your opinion, in the field of immune-related adverse events. In short, where do we go from here?

Amanda: I think there are two pressing needs. First, we need to develop better tools to predict, at symptom onset, who will have refractory/severe neurotoxicity and then develop better, targeted treatments for those patients. In the most severe cases, when corticosteroids aren’t effective acutely, we really don’t have good options for second line therapy. Mortality rates in this patient subset is high. The second critical need is understanding the burden of disease of neuro irAEs for individual patients and on our healthcare system. We don’t know how many patients are seeking care for neuro irAEs and their healthcare utilization.

Kerry: In your 7 years of seeing this patient population, what is the biggest change that you have noticed?

Amanda: At first, it was really like we were sailing without a compass. Every patient was a first or an “N of 1.” Perhaps there had been 1 or 2 similar cases reported in the literature. That was challenging both for patients and for me as a clinician. I’ve learned so much from the collective experience of the oncologists and other members of the SIC service. As we have built our experience here at MGH and the body of literature has grown, it has been so incredibly rewarding to see patients now and be able to offer them data to support our decision-making and to speak with more confidence about the neuro irAEs and what response we might see to treatment. We still have so far to go. Let’s meet back in another decade and ask that same question again!


Guidon AC, Burton LB, Chwalisz BK, Hillis J, Schaller TH, Amato AA, Betof Warner A, Brastianos PK, Cho TA, Clardy SL, Cohen JV, Dietrich J, Dougan M, Doughty CT, Dubey D, Gelfand JM, Guptill JT, Johnson DB, Juel VC, Kadish R, Kolb N, LeBoeuf NR, Linnoila J, Mammen AL, Martinez-Lage M, Mooradian MJ, Naidoo J, Neilan TG, Reardon DA, Rubin KM, Santomasso BD, Sullivan RJ, Wang N, Woodman K, Zubiri L, Louv WC, Reynolds KL. Consensus disease definitions for neurologic immune-related adverse events of immune checkpoint inhibitors. J Immunother Cancer. 2021 Jul;9(7):e002890. doi: 10.1136/jitc-2021-002890.

Video: Framework for Defining ICI-induced Neurotoxicities