Updated Brazilian guidelines for the management of immune-related adverse events associated with checkpoint inhibitors - edition 2.0

General aspects and pathophysiological mechanisms of irAEs were previously published in the first edition of the “Brazilian guidelines for the management of immune-related toxicities associated with checkpoint inhibitors”.6 Although the specific mechanisms underlying irAEs are not yet fully understood, accumulated evidence suggests that the development of irAES is associated with the disruption of immune coreceptors' role in the maintenance of immune homeostasis and selftolerance. Different manifestations of irAEs reflect different pathophysiological mechanisms including: 1. exacerbation of a subclinical autoimmune condition or upregulation of pre-existing autoantibodies; 2. triggering of a new inflammatory or immune-related condition resulting from the aberrant activation of autoreactive lymphocytes; 3. imbalance in the local and systemic levels of cytokines, leading to an inflammatory and selfreactive state; and 4. cytotoxic responses mediated by the interaction of monoclonal antibodies with their target receptors (e.g., cytotoxic T-lymphocyteassociated protein 4 [CTLA-4] expressed by healthy tissues).7 Data from translational studies suggest that irAEs usually develop through a combination of these factors involving autoreactive T-lymphocytes, autoantibodies, and cytokine production.8 However, the magnitude to which each of these components contribute to the development of irAEs remains poorly understood and may vary depending on the severity and profile of the irAEs.

The incidence and kinetics of onset of irAE may be influenced by the type of monoclonal antibodies used and regimen (combination vs monotherapy), the underlying malignancy and patient's intrinsic risk factors. Immune-related AEs may occur at any time during treatment, including the period after treatment discontinuation, and may range in severity from mild to severe, and even life-threatening events. Available data on irAEs are based primarily on toxicities documented in clinical trials and, more recently, on the results of several meta-analyses dedicated to the study of such events and reallife cohorts. Considering all grades and agents, the most common irAEs were fatigue (18.26%), pruritus (10.61%), and diarrhea (9.47%). The most common grade ≥3 irAEs are fatigue (0.89%), anemia (0.78%), and elevated aspartate aminotransferase (AST) (0.75%). No statistically significant relationship between irAE rates and primary cancer type has been identified to date, although the highest incidence of irAEs has been described in patients with melanoma and the lowest in patients with lung cancer.9 However, it is important acknowledge that the available evidence on the incidence of such events may be limited by variations in the screening protocols used to detect irAEs in each study and also in clinical practice, as such protocols dynamically evolve with increasing knowledge about irAEs. In general, the rates of any grade and grade ≥3 irAEs associated with monotherapy with anti-PD-1/ PD-L1 agents (60-75% and 10-15%, respectively) are lower than those associated with ipilimumab monotherapy, whose expected rates of any grade and grade ≥3 irAEs, usually reported in the ranges of 75-80% and 20-25%, depending on the dose of anti-CTLA-4 used.9 10

Possible exceptions include specific endocrinopathies, such as thyroiditis/thyroid dysfunction, which have been more frequently reported with anti-PD-1/PD-L1 administration. A meta-analysis including patients from 19 randomized clinical trials compared the rates of AEs related to anti-PD-1 versus anti-PD-L1 monoclonal antibodies. The incidence of irAEs was similar between with anti-PD-1 and anti-PD-L1 agents (relative risk = 1.24; 95% confidence interval [CI] = 0.79-1.93), as was the incidence of irAEs leading to death (relative risk = 1.38; 95% CI = 0.11-16.89).11 On the other hand, combinatorial strategies such as nivolumab/ipilimumab are associated with a significant increase in the incidence of toxicities (any grade: 95%; grade 3 or higher: 55%) and higher treatment discontinuation rates.12-15 Combinations of ICI with other classes of antineoplastic agents, such as anti-angiogenic drugs and/or chemotherapy, do not significantly increase the incidence of irAEs but are associated with increased overall toxicity due to the addition of AEs from each treatment class. Chemoimmunotherapy combinations, when compared to immunotherapy alone, have a higher incidence of grade 3-4 AEs (relative risk = 1.32; 95% CI = 1.12-1.55), with no significant impact on mortality rates.16

Regarding the time to onset of irAEs, patients treated with ipilimumab usually experience a new AE within the first 12 weeks of treatment and resolution in 6-8 weeks (approximately 7 weeks for grade 3-4 AEs), but 5-15% of patients may experience persistent irAEs after 24 months, most of which grades 1-2.17 Similarly, most irAEs associated with the use of antiPD-1/PD-L1 agents occur within the first 4 months after initiating therapy; however, the interval to onset tends to be shorter for combinations of anti-CTLA-4 and antiPD-1/PD-L1 drugs.18 Combination treatments with chemotherapy or anti-angiogenic agents do not appear to impact the onset, duration, or resolution of irAEs.16 19

The frequency of fatal irAEs was evaluated in a meta-analysis including more than 19,000 patients, which demonstrated a mortality rate of 0.36% with anti-PD-1 therapy, 0.38% with anti-PD-L1, 1.08% with anti-CTLA-4, and 1.23% with the combination of anti-PD-1/PD-L1 and anti- CTLA-4. Colitis is the most frequent cause of death following anti-CTLA-4 therapy. Pneumonitis, hepatitis, and neurological events are the main causes of death associated with anti-PD-1/PD-L1 therapy. Colitis and myocarditis are the most frequent causes of death due to irAEs in combination regimens, although we must acknowledge that there may be an overlap of severe and potentially fatal AE from both types of agents.20 The onset of fatal AEs also appears to be early in most cases, with a median for onset of 40 days with anti-PD-1/PD-L1 agents in monotherapy and less than 15 days for those treated with the combination of ipilimumab and nivolumab.

A possible association between the development of irAEs and better treatment outcomes in patients treated with ICI has been documented in different studies. Two recent meta-analyses demonstrated a positive association between the development of low-grade irAEs and response rates (RRs), progression-free survival (PFS), and overall survival (OS) in patients treated with immunotherapy regardless of the primary site, immunotherapy used, and irAEs. High-grade irAEs (grade 3 or higher) were associated with higher RR but lower OS.21 22 One of the challenges in confirming this association is the fact that patients who achieve greater benefit from immunotherapy may eventually be exposed to therapies for a longer period, inducing a bias related to the duration of exposure. However, evidence of a true association between irAEs and better treatment outcomes is suggested by the fact that, as previously mentioned, irAEs tend to occur in the first few months of treatment and that this association persists after applying statistical methods to minimize this bias.

Before initiating therapy with ICIs, all patients should be assessed for susceptibility to developing irAEs, as toxicity can potentially involve any organ or tissue. Subclinical manifestations can make the diagnosis of irAE especially challenging for physicians.23 24 Therefore, both patients and the multidisciplinary team must be aware of the potential risks, thus allowing early identification of irAEs (Figure 1).25 26

It is recommended that all patients receiving ICIs be followed closely and carefully, and undergo periodic evaluations, including detailed anamnesis and a physical examination performed before each treatment administration. It is also recommended that laboratory tests be performed before the initiation of treatment and repeated throughout treatment as summarized in Table 1.27 The frequency of tests can be modified based on individual assessments, the occurrence of AEs, and clinical suspicion. Nonetheless, no ideal screening strategy is unanimously accepted as a standard of care and, consequently, many of the current routines have been adapted from the clinical research protocols that led to the approval of these agents.

Patients with a history of autoimmune disease are at increased risk of developing an irAE or a “flare” of the underlying autoimmune condition, therefore requiring greater attention and closer monitoring. As this is a very heterogeneous group, encompassing a wide range of pathologies and severity of manifestations, individualizing treatment is essential in the therapeutic decision-making process, and the benefit of each intervention should be considered along with the risk of toxicity and exacerbation of the autoimmune condition. Data showing the safety and efficacy of ICIs in patients with autoimmune diseases are limited. Retrospective data suggest that patients with stable autoimmune diseases may receive antiPD-1 therapy. Although the risk of irAEs is higher in this population and that patients with preexisting autoimmune diseases may experience flares when treated with both anti-CTLA-4 and anti-PD-1 agents, such exacerbations occur at a rate of approximately 20 to 40%, and tend to be manageable, provided they are immediately recognized and managed appropriately. In addition, the RRs in this subgroup appear to be similar to those achieved in the general population.28-31 However, the use of ICIs should be avoided in patients with severe active autoimmune disease or those using high-dose corticosteroids or immunosuppressants since any additional immune activation can be potentially life-threatening. Two ongoing clinical trials, one on lung cancer (NCT03656627) and the other on several tumor types (NCT03816345), are prospectively evaluating the use of ICIs in patients with underlying autoimmune disease.

Regarding the use of corticosteroids, retrospective studies suggest that the use of prednisone ≥10mg/ day or equivalent when initiating therapy with antiPD-1/PD-L1 is associated with lower RR, PFS, and OS. A multivariate analysis adjusted for Eastern Cooperative Oncology Group performance status, smoking, and presence of brain metastases showed that corticosteroids were associated with worse survival.32 33 Thus, caution is recommended when using corticosteroid therapy before the initiation of ICIs until more robust evidence is available. Importantly, the use of corticosteroids or immunosuppressants for the management of irAEs after the beginning of treatment does not seem to affect the outcomes of patients receiving ICIs, although this is still controversial. In an analysis of efficacy and safety of patients with advanced melanoma and lung cancer who discontinued immunotherapy due to irAEs, the use of corticosteroids was not associated with worse RR, PFS, or OS.34-37

Patients with chronic viral infections were excluded from most clinical trials to date. However, the safety of ICI in patients with hepatitis B or C has been suggested by an increasing number of case series demonstrating hepatotoxicity rates similar to those seen in the general population.38 Similarly, a phase I/II study evaluating the safety and efficacy of nivolumab in patients with hepatocellular carcinoma showed an acceptable safety profile in patients with chronic viral infection.39 Regarding patients with HIV infection, a systematic review identified 73 patients treated with ICIs. In patients with melanoma and lung cancer the effectiveness was similar between patients with or without HIV infection, with RRs of 27% and 30%, respectively. Patients living with HIV did not show an increased incidence of AEs, and the viral load remained undetectable in 93% (26 of 28) of the patients who did not have a detectable viral load prior to treatment initiation.40 Thus, treatment of patients with chronic viral infection with ICI appears to be safe, although close and multidisciplinary follow-up are of paramount importance.

Multidisciplinary management is also important in the complex setting of patients with a history of solid organ transplantation being considered for treatment with ICI. Safety and efficacy data are even more scarce in this population and limited to a small number of case reports or case series. Despite reports about safe administration of these agents in selected cases, recent case series suggested that the risk of rejection can exceed 40%, leading to mortality rates as high as 40-50%, depending on the organ transplanted.41 42 Thus, therapeutic decisions involving the use of ICIs in solid organ transplant recipients should always be shared with the patients and other teams involved in their care and take into consideration the risk of transplant rejection, potential benefits from ICI therapy, and available therapeutic alternatives.

Regarding the indication of ICI for elderly patients, age alone should not be a contraindication to this therapeutic approach. Despite being underrepresented in randomized controlled trials, subgroups analyses in prospective and retrospective studies suggested that the efficacy and safety of immunotherapy in the older adults are similar to that of general population.43 44 Importantly, factors such as comorbidities and potential drug interactions should be carefully considered in this population. Geriatric and frailty assessments are important predictors of irAEs and worsening quality of life, and both are recommended in this subgroup.45

Early symptom recognition and rapid intervention remain critical factors in the management irAEs. In the initial evaluation, the exclusion of differential diagnoses (such as infections) is essential. Invasive procedures for obtaining tissue samples (bronchoscopy, endoscopy, skin biopsies, etc.) are frequently indicated. The temporary discontinuation of ICIs is often necessary, and the use of corticosteroids and symptomatic therapy remain the pillars of irAEs management. The use of corticosteroids to control irAEs does not seem to affect the efficacy of cancer treatment,13 24 34 and the definitive discontinuation of immunotherapy is still indicated in severe cases.

Permanent discontinuation is usually not required for endocrine toxicities such as glandular dysfunction (even grade 4) provided that adequate hormonal replacement is initiated and the patient's clinical condition is stable. For patients with severe AEs that are refractory to oral or intravenous corticosteroid therapy, subsequent lines of treatment include monoclonal antibodies such as infliximab, rituximab, or tocilizumab (among others), mycophenolate mofetil, or other immunosuppressants (such as azathioprine and cyclosporine), particularly if no improvement is seen after 3-5 days of intravenous corticosteroid therapy. In selected cases, particularly those with neurological or hematological complications, plasmapheresis or intravenous immunoglobulin are therapeutic alternatives. It should be noted that the early recognition of refractory cases and the timely introduction of the appropriate therapy (corticosteroids, usually in addition to other treatments, when indicated) are crucial to the management of irAEs. In these situations, or even in cases of lower grade events, an evaluation by a specialist (such as endocrinologist or gastroenterologist) is encouraged, preferably a specialist who is familiar with the management of irAEs. Corticosteroids should be slowly tapered after adequate improvement of the irAE is achieved, usually over 4 weeks, with the possibility of extension for 6-8 weeks or longer in cases of pulmonary or hepatic irAEs.

Patients treated with corticosteroids and/or immunosuppressive agents are at risk for opportunistic infections and tuberculosis reactivation, and, when appropriate, should receive antimicrobial prophylaxis as well as be monitored for signs of infectious conditions. The initial approach and management of irAEs are summarized in Figure 2. The doses of the drugs more commonly used in this clinical setting are described in Appendix 1.

With greater access to immunotherapy and better understanding and management of irAEs, the possibility of rechallenging with new exposure to ICI after temporary discontinuation is increasingly frequent. Prospective data from randomized controlled trials supporting this strategy are scarce because most prospective studies used management algorithms that established permanent discontinuation of treatment for severe irAEs. In a cohort of 93 patients treated with anti-PD-1/PD-L1 who had grade 2 or higher irAEs, 40 were re-exposed to the same agent, 55% of which had a recurrence of one or more irAEs. There was no increase in irAE severity on re-exposure, and approximately 20% of patients developed different types of irAE from the previous ones.46 These data are in line with those of other previously published cohorts47 and suggest that rechallenge with anti-PD-1/PD-L1 agents can be considered with proper monitoring.

Recommendations regarding the discontinuation of immunotherapy and rechallenging are summarized in Table 2. It should be noted that irAEs may have late onset, even after the discontinuation of the immunotherapy; therefore, monitoring for these adverse effects should be continuous,48 and re-evaluations of toxicity (anamnesis, physical examination, and laboratory tests) after the discontinuation of treatment are suggested every 3-4 months in the first year and every 6 months thereafter.12 48 49

The estimated incidence of hepatic irAEs is low, occurring in about 2-7% of patients treated with anti-PD-1/PD-L1 or anti-CTLA-4 alone, and the most common presentation is asymptomatic elevation of liver tests, such as aspartate aminotransferase, alanine aminotransferase, gamma- glutamyl transferase, or bilirubin.14 15 17 57 However, there is a significant increase in the risk of hepatic AEs when dual immune checkpoint blockade is used (any grade, 15-30%; grades 3/4, 6.0-18.8%).15 57 The initial presentation with acute liver failure is rare and the onset of symptoms typically occurs between 4-12 weeks.58 Treatmentrelated liver toxicity is usually the diagnosis of exclusion, and viral and cancer- related causes for liver abnormalities should be ruled out. Unlike autoimmune hepatitis, hepatotoxicity caused by immunotherapy is not predominant in women. Antinuclear antibody (ANA) may be present in up to 50% of patients, usually at low titers (1:80), whereas anti-smooth muscle antibody (SMA) test results are rarely positive. Liver biopsy shows no or few plasma cells and fewer CD20+ and CD4+ lymphocytes when compared with autoimmune hepatitis.58

The management of hepatotoxicity includes the use of systemic corticosteroids and, in more severe cases, other immunosuppressants such as mycophenolate mofetil (Appendix 2, Supplementary Material).

Infliximab is contraindicated because of the intrinsic potential for hepatic toxicity. The added value of performing liver biopsy remains controversial, and is based on expert recommendations. Liver biopsy can be considered in patients with low risk of bleeding and in corticosteroid-refractory cases, concomitant use of other hepatotoxic medications or suspected liver metastases. The main limitation of liver biopsy, in addition to its cost and being an invasive procedure, is the fact that there are no pathognomonic histological findings and that, in most cases, the biopsy does not change the management of this irAE.59

Although pulmonary toxicity is relatively uncommon, it is a potentially life-threatening event and should be considered in patients with new onset of respiratory symptoms.12 49 Of note, real-world data suggest that the incidence of pneumonitis is higher than that reported in clinical trials, and as high as 8-19% in patients with non-small cell lung cancer (NSCLC) treated with anti-PD-1/PD-L1 agents.60 61 Also, pneumonitis is the most common fatal irAE, representing about 35% of the causes of death due to the use of anti-PD-1/PD-L1.20

Retrospective and post hoc analyses suggest that patients previously exposed to chest radiotherapy may be at increased risk for pneumonitis and radiation recall and, therefore, require closer surveillance. In KEYNOTE-001 trial evaluating pembrolizumab in patients with NSCLC, the incidence of pneumonitis was numerically higher in the patients who had received previous thoracic radiotherapy versus those with no previous thoracic radiotherapy (8% vs. 1%; p=0.15).62 However, in the PACIFIC study, which evaluated the addition of durvalumab after chemoradiotherapy in patients with stage III NSCLC, the use of immunotherapy numerically increased the risk of pneumonitis, radiotherapy- induced pneumonitis, or any grade pneumonia but did not increase the risk of grade 3-4 pneumonitis (3.4% vs. 2.6%, respectively).63

Pneumonitis treatment must be adapted to the severity of the condition (Appendix 3, Supplementary Material). In this population, the concomitant use of empirical antibiotic therapy is frequent. Nevertheless, the exclusion of differential diagnoses using tests such as bronchoscopy or lung biopsy should be considered when the diagnosis is uncertain.

Dermatologic toxicities, including pruritus, xeroderma, rash, psoriasiform lesions, and lichenoids, are the most common irAEs in patients treated with ICIs. They are more common in patients treated with anti-CTLA-4 agents and combination regimens. Although most cases are grade 1 or 2, serious events may occur, such as toxic epidermal necrolysis or Stevens-Johnson syndrome. Psoriasiform lesions and maculopapular eruptions are often the irAEs of earliest onset, with a median manifestation of 2-5 weeks after the initiation of therapy.64 On the other hand, lichenoid dermatoses, bullous (bullous pemphigoid), and hypopigmentation/ depigmentation vitiligo-like lesions tend to have later onset.64 A higher incidence of cutaneous irAEs is observed in patients with advanced melanoma than other neoplasms, particularly vitiligo-like changes, which are strongly associated with higher PFS and OS.13 48

The management of cutaneous irAEs follows the same principles recommended for other irAEs. Consultation with a dermatologist is often recommended, with a low threshold for skin biopsies (Appendix 4, Supplementary Material). Most cases can be managed using topical steroids combined or not with antihistamines, and ICI treatment usually does not need to be discontinued. For patients with pruritus, the use of GABA receptor agonists, such as pregabalin or gabapentin, or the NK-1 receptor antagonist aprepitant may be considered. In refractory or severe cases, a skin biopsy is crucial, and strategies such as systemic corticosteroid therapy, immunosuppressants, or monoclonal antibodies may also be necessary for therapeutic management.65

Endocrine irAEs are diagnosed in up to 10% of patients treated with ICIs. A high level of suspicion is recommended, since the symptoms may be nonspecific.66 The onset of endocrine irAEs usually occurs at between the 4^th and the 18^th weeks of treatment (median: 11 weeks), but late manifestations can also occur.67 A meta-analysis of 38 randomized controlled trials compared the incidence of endocrine irAEs resulting from different immunotherapies. Patients who received a combination of anti-PD-1/ PD-L1 and anti-CTLA-4 had a higher incidence of thyroid dysfunction than those who treated only with ipilimumab. Comparisons between anti-PD-1 and ipilimumab as monotherapies demonstrated that the use of anti-PD-1 was more likely to induce hypothyroidism, while the use of ipilimumab was associated with a higher incidence of hypophysitis. Other endocrine irAEs, including primary adrenal insufficiency and insulin-dependent diabetes mellitus, were uncommon and occurred in 0.7% and 0.2% of patients, respectively.66 In patients receiving ICIs and presenting with hypotension/shock, nausea, vomiting, and mental confusion, the possibility of an adrenal crisis must be ruled out, although it is an extremely unusual event. The approach and management of the most common endocrine irAEs are summarized in Appendix 5 (Supplementary Material). It is worth mentioning that the use of corticosteroids, except for hormone replacement, is controversial in cases of endocrine toxicity and usually not recommended (except for the control of local symptoms/mass effects).

Conditions affecting the thyroid, such as hypothyroidism, hyperthyroidism, or symptomatic thyroiditis can occur, although the latter presentation is uncommon. Hypothyroidism with the use of ICIs occurs in approximately 6.6% of patients, with the lowest incidence (3.8%) reported with ipilimumab and the highest incidence (13.2%) reported with combination therapies.

The risk of hyperthyroidism, but not hypothyroidism, appears to be greater with the use of anti-PD-1 than with anti-PD-L1 (odds ratio [OR] = 5.36; p=0.002), and it is often temporary.66 Thyroid function tests should be monitored before each dose or monthly and then every 6-12 weeks for 6 months after the completion of treatment. The median onset of thyroid dysfunction was 4 weeks after the initiation of immunotherapy.68 The recommended treatment is the standard: hormone replacement or antithyroid drugs in selected cases.

Pituitary dysfunction is among the most commonly reported endocrine irAEs. The greatest incidence occurs with anti-CTLA-4 agents and combined regimens. The incidence is dose-dependent (1.0-4.0% with ipilimumab 3mg/kg and 16% with ipilimumab 10mg/kg). It is a rare event with the use of anti-PD-1 monotherapy (0.4%).66

Adrenal insufficiency is usually permanent and requires continuous hormone replacement. It presents as high levels of adrenocorticotropin hormone in the presence of low cortisol, differentiating it from hypophysitis. If adrenal insufficiency and hypothyroidism are both present, corticosteroids should be started before thyroid hormone replacement is administered due to the risk of adrenal crisis. Sepsis, for which broad- spectrum empirical antibiotic therapy is usually required, should always be investigated.69

Autoimmune (type 1) diabetes is a rare event reported in approximately 0.4-0.9% of patients using anti-PD-1 agents. Fasting blood glucose is the preferred test for patient follow-up and surveillance. Importantly, the use of high-dose corticosteroids for the treatment of other irAEs can induce or exacerbate hyperglycemia. Diabetic ketoacidosis may occur, even in patients with a previous diagnosis of type 2 diabetes and may present with symptoms such as increased thirst, pollakiuria, nausea, vomiting, mental confusion, abdominal pain, and dehydration. Corticosteroid therapy is not recommended for the treatment of type 1 diabetes since there is no evidence to support its use and it may further worsen glycemic control.69 70

Regardless of the presentation of endocrine irAEs, the permanent discontinuation of treatment is rarely recommended if hormone replacement therapy is initiated and the symptoms resolve. Follow-up by an endocrinologist is recommended.

As previously discussed, irAEs presentations are extremely heterogeneous and any organ/tissue could theoretically be the target of an immunemediated injury. With the expansion of the use of immune coreceptor blockers, greater knowledge about rare irAEs has been acquired.

Cardiac irAEs, including myocarditis, pericarditis, cardiac fibrosis, arrhythmias, and heart failure, can occur even in patients without significant risk factors, and are associated a high mortality rate.

The time to onset of these events is variable, and may occur after a single dose of immunotherapy. Patients receiving combination therapies appear to be at an increased risk for these AEs, followed by patients receiving anti-PD- 1 monotherapy.71 The role of cardiac enzymes monitoring remains uncertain, despite being included in more recent research protocols in addition to periodic echocardiography or other imaging tests. The treatment of cardiac irAEs should be adapted according to the severity of the clinical presentation (Appendix 6, Supplementary Material). Permanent discontinuation of therapy is recommended, with the initiation of high-dose corticosteroids (prednisone 1-2mg/kg). The timing of the initiation of corticosteroid therapy in mild cases should be individualized considering the lack of data in this setting. In refractory patients, doses of corticosteroids similar to those used in cases of heart transplant rejection are usually indicated (1g of methylprednisolone per day, for 3 to 5 days) in addition to the addition of mycophenolate or infliximab.72

Hematological irAEs are rare, although a myriad of manifestations has been described. An analysis of the World Health Organization pharmacovigilance database identified 168 cases of hematological irAEs. The most common cases were immune thrombocytopenic purpura (68 cases) and hemolytic anemia (57 cases), including four cases in which both conditions occurred concurrently.73 The incidence of anemia and thrombocytopenia appears to be higher in patients treated with ICIs for refractory Hodgkin's disease, which may be partly related to the previous use of myelotoxic therapies.74 The treatment of hematological irAEs must be adapted to the condition's severity (Appendix 7, Supplementary Material). In steroid-refractory presentations, a bone marrow biopsy should be considered. In addition to the usual approaches, management of hematological irAEs may also include transfusion support, rituximab, and plasmapheresis. Ruling out alternative diagnoses for hematological conditions is of paramount importance, and possible causes include disease progression, bone marrow infiltration, bleeding, and AEs of other drugs.

Approximately 14% of patients using ipilimumab plus nivolumab may develop neurological irAEs, a significantly higher incidence than with anti-CTLA-4 or anti-PD-1 monotherapy, which have an incidence of 1% and 3%, respectively. Guillain-Barré syndrome, myasthenia gravis, aseptic meningitis, limbic encephalitis, transverse myelitis, reversible posterior encephalopathy syndrome, enteric neuropathy, granulomatous inflammation of the central nervous system, and Tolosa-Hunt syndrome have been reported. Severe cases of limbic encephalitis were reported in patients with small cell lung cancer.

Neurological paraneoplastic syndromes or autoimmune encephalitis can be exacerbated or revealed by the immune stimuli triggered by immunotherapy. Thus, tumors with a greater propensity to develop neurological paraneoplastic syndromes, such as small cell lung carcinoma and Merkel carcinoma, must be monitored with special attention to the possibility of neurological irAEs.75-77

The management of neurologic irAEs is similar to that for other irAEs (Appendix 8, Supplementary Material). In patients whose symptoms are suggestive of neurological irAEs, corticosteroid therapy and discontinuation of immunotherapy should be considered, even in mild presentations. Intravenous immunoglobulin and plasmapheresis may be considered in addition to corticosteroids and immunosuppressants. Pyridostigmine may be considered for patients with myasthenia gravis. Approximately one-third of patients with neurological irAEs have residual sequelae.77 Pre-existing neurological conditions should not contraindicate treatment; however, individual risks should be evaluated before the use of ICIs.78

The most common musculoskeletal and rheumatological symptoms are arthritis and polymyalgia, but their incidence has not been precisely established, which reflects the difficulty distinguishing between irAEs and musculoskeletal symptoms caused by other factors. Arthralgias without clear inflammatory signs can also significantly impact quality of life and require specific procedures. A meta-analysis that included clinical trials, observational studies, and case reports or case series reported an incidence of arthralgia of 1.0-43% and myalgia of 2.0-20%, showing a significant variability between reported symptoms and diagnosed rheumatological irAEs. Anti-PD-1 therapy is more commonly associated with rheumatologic irAEs. Myositis, although rare, can be fatal and present as a reactivation of preexisting paraneoplastic polymyositis or dermatomyositis.

Depending upon the clinical manifestation, mild cases can be managed with anti-inflammatory drugs, and in case of treatment failure, corticosteroids should be initiated. Disease-modifying antirheumatic drugs (DMARDs) and even plasmapheresis may be used in steroid-refractory patients, in addition to other rheumatological monoclonal antibodies commonly used for autoimmune conditions.79-82 The approach to rheumatological irAEs must be adapted according with the event's severity (Appendix 9, Supplementary Material). In a cohort of 65 patients, predominantly composed of patients with lung cancer and melanoma who developed musculoskeletal symptoms, an algorithm for the management of irAEs led to symptom control in 54.4% of patients using prednisone 10 mg associated or not with non-steroidal anti-inflammatories (NSAIDs), and 12.3% required the use of DMARDs.

Treatment was discontinued in 8.8% of cases and, as demonstrated in other immune-mediated toxicities, complete disease remission in patients with melanoma was observed in 39% of patients with rheumatologic irAEs, compared to 4% of patients without irAEs.83

Although initially characterized as rare (3.0-5.0% of patients), any renal grade irAEs may occur in 1520% of patients treated with ICIs. Acute interstitial nephritis is the most common finding in renal biopsies. Renal toxicities have an earlier onset with combined regimens (2-3 months) than with antiPD-1 monotherapy (3-10 months). In particular, patients being treated with platinum, pemetrexed, and pembrolizumab combination regimen should be closely monitored due to an increased risk of renal toxicity. The treatment of renal AEs should be tailored to the severity of the irAE (Appendix 10, Supplementary Material).

Ocular irAEs include uveitis, episcleritis, and conjunctivitis.84 In a pharmacovigilance database, ophthalmic events were rare, representing up to 3% of all irAEs reported, with visual disorders (30.8%) and uveitis (15%) being the most common events. The combination of anti-CTLA-4 and anti-PD-1 results in the greatest risk for the development of uveitis (OR=4.77; 95%CI=3.835.94). Regarding tumor types, patients diagnosed with melanoma (OR=14.7; 95%CI=10.7-20.2) appear to be at higher risk of developing uveitis than those with lung cancer (OR=2.67; 95%CI=1.68-4.23).85