Safety analyses of the INAVO120 randomised phase III trial of inavolisib or placebo with palbociclib–fulvestrant in patients with PIK3CA-mutated, hormone receptor-positive, HER2-negative, endocrine-resistant advanced breast cancer Authors: S.-A. Im, K. Kalinsky, K.L. Jhaveri, S. Loibl, N. Turner, C. Saura, P. Schmid, S. Loi, E. Hamilton, N. Karadurmus, S. Wang, A.A. Awan, E.M. Ciruelos, C.-F. Chung, E. Thanopoulou, N. Shankar, S. Lim, Y. Jin, C. Song, D. Juric Published: 2026 ABSTRACT BackgroundINAVO120 (NCT04191499) demonstrated significantly improved progression-free/overall survival with inavolisib plus palbociclib–fulvestrant versus placebo plus palbociclib–fulvestrant in patients with PIK3CA-mutated, hormone receptor-positive, HER2-negative, endocrine-resistant advanced breast cancer. We provide comprehensive analyses of key selected adverse events and their management.Materials and methodsInavolisib was given 9 mg orally once daily on days 1-28 of each 28-day cycle; palbociclib, at 125 mg orally once daily on days 1-21; fulvestrant, at 500 mg intramuscularly on days 1 and 15 of cycle 1, and every ∼28 days thereafter. Supportive therapies were used as clinically indicated. Key selected adverse events assessed included hyperglycaemia, stomatitis/mucosal inflammation, rash and diarrhoea, graded per National Cancer Institute Common Terminology Criteria for Adverse Events v5.0.ResultsData cut-off was 29 September 2023; median follow-up was 21.3 months (range 0-43.1; inavolisib arm) and 21.5 months (0.1-40.3; placebo arm). Inavolisib dose interruptions, reductions and discontinuations due to hyperglycaemia were observed in 27.2%, 2.5% and 0.6% of patients, respectively; the median time to first onset was 7.0 days (range 2.0-955.0). Metformin was the most commonly used antihyperglycaemic. No patients in the prediabetic population discontinued inavolisib due to hyperglycaemia. Inavolisib dose interruptions, reductions and discontinuations due to rash were observed in 1.2%, 0.6% and 0% of patients, respectively; the median time to first onset was 29.0 days (range 1.0-952.0). Topical hydrocortisone was most commonly used. Inavolisib dose interruptions, reductions and discontinuations due to stomatitis/mucosal inflammation were observed in 9.9%, 3.7% and 0.6% of patients, respectively; the median time to first onset was 13.0 days (range 1.0-610.0). Dexamethasone mouthwash was most commonly used. Inavolisib dose interruptions, reductions and discontinuations due to diarrhoea were observed in 6.8%, 1.2% and 0% of patients respectively; the median time to first onset was 13.0 days (range 1.0-610.0). Loperamide was most commonly used. Data were largely comparable across regions and ages. Inavolisib’s safety profile was consistent with that of long-term treated populations.ConclusionsThis analysis demonstrates the generally consistent, manageable and tolerable safety profile of inavolisib plus palbociclib–fulvestrant. Key selected adverse events were generally reversible and were controlled by concomitant medications and dose modifications. Introduction Inavolisib is an oral, highly potent and α-selective phosphoinositide 3-kinase (PI3K) inhibitor that also promotes mutated p110α degradation.1 The phase III, randomised, double-blind, placebo-controlled INAVO120 study (NCT04191499) assessed first-line inavolisib/placebo with palbociclib–fulvestrant in PIK3CA-mutated, hormone receptor-positive, HER2-negative, endocrine-resistant advanced breast cancer.2 INAVO120 met its primary endpoint (clinical cut-off, 29 September 2023): inavolisib plus palbociclib–fulvestrant resulted in a significant progression-free survival (PFS) benefit versus placebo plus palbociclib–fulvestrant [hazard ratio 0.43, 95% confidence interval (CI) 0.32-0.59, P<0.001; medians, 15.0 (95% CI 11.3-20.5) and 7.3 months (95% CI, 5.6-9.3), respectively].2 The updated PFS hazard ratio at the final analysis (clinical cut-off, 15 November 2024) was 0.42 [95% CI 0.32-0.55; medians, 17.2 (95% CI 11.6-22.2) and 7.3 months (95% CI 5.9-9.2), respectively] and overall survival (OS) was also significantly improved [hazard ratio 0.67 (95% CI 0.48-0.94; P=0.02; final OS boundary 0.0469); medians, 34.0 (95% CI 28.4-44.8) and 27.0 months (95% CI 22.8-38.7), respectively].3 INAVO120 confirmed that inavolisib can be combined with a cyclin-dependent kinase 4/6 (CDK4/6) inhibitor and endocrine therapy at full doses of each agent, with an acceptable, manageable safety and tolerability profile.2, 3, 4 In-depth knowledge of the safety profile of novel therapeutics is crucial for associated adverse event management. We provide comprehensive analyses of key selected adverse events and their management and an overview of inavolisib’s safety profile in selected subgroups. Materials and methods Study design and procedures The study design and protocol have been published.2 INAVO120 enrolled pre-, peri- and postmenopausal women, and men, with measurable disease by RECIST v1.1,5 endocrine-resistant disease by European Society for Medical Oncology criteria,6 fasting glucose <126 mg/dL (<7.0 mmol/L) and glycated haemoglobin (HbA1c) <6.0% (amended from <5.7% to more closely align with the eligible threshold for fasting glucose including prediabetes per the American Diabetes Association). Patients with type 1 or 2 diabetes requiring ongoing treatment were excluded. Patients were randomly assigned 1:1 to inavolisib [9 mg orally (p.o.) once daily (qd) on days 1-28 of each 28-day cycle]/placebo (p.o. qd) with palbociclib (125 mg p.o. qd on days 1-21 of each 28-day cycle) plus fulvestrant (500 mg intramuscularly on cycle 1, days 1 and 15, and every ∼28 days thereafter). Stratification factors were visceral disease (yes versus no), endocrine resistance (primary versus secondary6) and region [North America or Western Europe versus Asia versus Other (see Appendix for countries comprising each category)]. Inavolisib dose modifications, including interruptions, were permitted; reductions were to 6 mg qd and then 3 mg qd. Palbociclib dose modifications were permitted per local prescribing information. Fulvestrant dose reductions were not permitted (interruptions were). Treatment continued until disease progression, unacceptable toxicity, consent withdrawal or death. Supportive therapies were used as clinically indicated, per local standard practise. Antiemetic and antidiarrhoeal medications were not administered prophylactically before initial study treatment; they could be used per standard clinical practice before subsequent study drug doses. To reduce stomatitis risk, dexamethasone alcohol-free mouthwashes were allowed to be used per the SWISH study.7 Prophylactic metformin for hyperglycaemia was allowed (primary prophylaxis was recommended at the investigator’s discretion, where allowed by local regulations, and could be initiated on cycle 1, day 1 for patients at high risk); no prophylaxis for rash was recommended. Further details are available in the published protocol.2 INAVO120 was performed in accordance with the the International Council for Harmonisation (ICH) E6 guideline for Good Clinical Practice and the principles of the Declaration of Helsinki. Safety assessments Safety monitoring included assessment of the nature, frequency and severity of adverse events, with severity graded per National Cancer Institute Common Terminology Criteria for Adverse Events (NCI-CTCAE) v5.0.2 All events were reported after study drug initiation until 30 days after the final dose of study treatment or initiation of another anticancer therapy, whichever occurred first. Key selected adverse events included hyperglycaemia, stomatitis/mucosal inflammation, rash and diarrhoea; grouped terms for these and other selected events are defined in the Supplement, available at https://doi.org/10.1016/j.esmoop.2026.107735. Adverse event management Standard-of-care measures and adverse event management guidelines utilising dose modifications and interruptions were effective in managing adverse events, as reflected by the high relative dose intensity maintained and by low treatment discontinuation due to adverse events.2 For hyperglycaemia, time was allowed (up to 7 days) for resolution after interruptions and antihyperglycaemic treatment to avoid premature dose reduction or discontinuation. Insulin or sulphonylureas were recommended to be administered with caution, as subsequent inavolisib interruption could lead to rapid insulin level escalation and hypoglycaemia.8 Summaries of key selected adverse event management and follow-up are reported in the Supplement, available at https://doi.org/10.1016/j.esmoop.2026.107735, including additional guidelines for patients at high risk of hyperglycaemia that were added to the protocol. Statistical analysis The Safety Analysis Set (SAS) comprised all patients who received ≥1 dose of any study drug.2 The Full Analysis Set (FAS) included all randomly assigned patients.2 Patients in the SAS were analysed by treatment received. Median and range or interquartile range were used for continuous variables and patient counts and percentages for categorical variables. Crude methods were used for analyses of time to onset of key selected adverse events (only including patients who had such events). Results Patient characteristics INAVO120 enrolled 325 patients in 28 countries between 29 January 2020 and 14 September 2023; the clinical cut-off for this analysis was 29 September 2023.2 Median follow-up was 21.3 months (range 0-43.1) in the inavolisib arm and 21.5 months (0.1-40.3) in the placebo arm.2 The SAS comprised 162 patients per arm; the FAS included 161 and 164 patients in the inavolisib and placebo arms, respectively. Sixty-two of 161 patients (38.5%) in the FAS of the inavolisib arm and 63 of 164 (38.4%) in the placebo arm were in North America or Western Europe; 58 of 161 (36.0%) and 62 of 164 (37.8%), respectively, in Asia; and 41 of 161 (25.5%) and 39 of 164 (23.8%), respectively, in other regions.2 Baseline disease characteristics were well balanced between arms; 132 of 161 (82.0%) and 128 of 164 (78.0%) patients had visceral disease.2 Median ages were 53.0 (range 27-77) and 54.5 years (29-79), and proportions of patients aged <65 years were 84.5% (136 of 161) and 79.3% (130 of 164) in the inavolisib and placebo arms, respectively.2 In the FAS at clinical cut-off, 58.4% of patients (94 of 161) discontinued inavolisib and 70.1% (115 of 164) discontinued placebo; 67 of 161 (41.6%) and 49 of 164 (29.9%) remained on treatment, respectively.2 Overall safety profile Adverse events occurred in 160 of 162 (98.8%) patients in the inavolisib arm and in all patients in the placebo arm.2 The most common (by preferred term) in the inavolisib arm were neutropenia [88 of 162 (54.3%)], hyperglycaemia [87 of 162 (53.7%)], diarrhoea [78 of 162 (48.1%)], neutrophil count decreased [63 of 162 (38.9%)], anaemia [59 of 162 (36.4%)] and stomatitis [53 of 162 (32.7%)] (Table 1).Table 1Adverse events occurring in ≥15% of patients in either treatment arm and grade 3-4 adverse eventsPatients, n (%)Inavolisib + palbociclib + fulvestrant (n = 162)Placebo + palbociclib + fulvestrant (n = 162)Any gradeGrade 3-4Any gradeGrade 3-4Blood and lymphatic system disorders Neutropenia88 (54.3)77 (47.5)89 (54.9)78 (48.1) Anaemia59 (36.4)10 (6.2)59 (36.4)3 (1.9) Thrombocytopenia36 (22.2)8 (4.9)41 (25.3)3 (1.9) Leukopenia28 (17.3)11 (6.8)40 (24.7)17 (10.5)Investigations Neutrophil count decreased63 (38.9)56 (34.6)64 (39.5)53 (32.7) Platelet count decreased42 (25.9)15 (9.3)36 (22.2)4 (2.5) White blood cell count decreased35 (21.6)23 (14.2)35 (21.6)17 (10.5) Aspartate aminotransferase increased28 (17.3)2 (1.2)26 (16.0)2 (1.2) Alanine aminotransferase increased28 (17.3)6 (3.7)21 (13.0)2 (1.2) Weight decreased28 (17.3)6 (3.7)1 (0.6)0Gastrointestinal disorders Diarrhoea78 (48.1)6 (3.7)26 (16.0)0 Stomatitis53 (32.7)5 (3.1)27 (16.7)0 Nausea45 (27.8)1 (0.6)27 (16.7)0General disorders and administration site conditions Fatigue38 (23.5)021 (13.0)2 (1.2) Asthenia25 (15.4)3 (1.9)22 (13.6)0 Mucosal inflammation30 (18.5)3 (1.9)16 (9.9)0Metabolism and nutrition disorders Hyperglycaemia87 (53.7)9 (5.6)12 (7.4)0 Decreased appetite38 (23.5)014 (8.6)0 Hypokalaemia26 (16.0)4 (2.5)10 (6.2)0Infections and infestations COVID-1937 (22.8)3 (1.9)17 (10.5)1 (0.6)Skin and subcutaneous tissue disorders Rash26 (16.0)021 (13.0)0 Alopecia30 (18.5)09 (5.6)0Nervous system disorders Headache34 (21.0)022 (13.6)0Clinical cut-off for this analysis was 29 September 2023. Adverse events are preferred terms, not grouped terms.Adverse event severity was graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events version 5.0. Table 1Adverse events occurring in ≥15% of patients in either treatment arm and grade 3-4 adverse eventsPatients, n (%)Inavolisib + palbociclib + fulvestrant (n = 162)Placebo + palbociclib + fulvestrant (n = 162)Any gradeGrade 3-4Any gradeGrade 3-4Blood and lymphatic system disorders Neutropenia88 (54.3)77 (47.5)89 (54.9)78 (48.1) Anaemia59 (36.4)10 (6.2)59 (36.4)3 (1.9) Thrombocytopenia36 (22.2)8 (4.9)41 (25.3)3 (1.9) Leukopenia28 (17.3)11 (6.8)40 (24.7)17 (10.5)Investigations Neutrophil count decreased63 (38.9)56 (34.6)64 (39.5)53 (32.7) Platelet count decreased42 (25.9)15 (9.3)36 (22.2)4 (2.5) White blood cell count decreased35 (21.6)23 (14.2)35 (21.6)17 (10.5) Aspartate aminotransferase increased28 (17.3)2 (1.2)26 (16.0)2 (1.2) Alanine aminotransferase increased28 (17.3)6 (3.7)21 (13.0)2 (1.2) Weight decreased28 (17.3)6 (3.7)1 (0.6)0Gastrointestinal disorders Diarrhoea78 (48.1)6 (3.7)26 (16.0)0 Stomatitis53 (32.7)5 (3.1)27 (16.7)0 Nausea45 (27.8)1 (0.6)27 (16.7)0General disorders and administration site conditions Fatigue38 (23.5)021 (13.0)2 (1.2) Asthenia25 (15.4)3 (1.9)22 (13.6)0 Mucosal inflammation30 (18.5)3 (1.9)16 (9.9)0Metabolism and nutrition disorders Hyperglycaemia87 (53.7)9 (5.6)12 (7.4)0 Decreased appetite38 (23.5)014 (8.6)0 Hypokalaemia26 (16.0)4 (2.5)10 (6.2)0Infections and infestations COVID-1937 (22.8)3 (1.9)17 (10.5)1 (0.6)Skin and subcutaneous tissue disorders Rash26 (16.0)021 (13.0)0 Alopecia30 (18.5)09 (5.6)0Nervous system disorders Headache34 (21.0)022 (13.6)0Clinical cut-off for this analysis was 29 September 2023. Adverse events are preferred terms, not grouped terms.Adverse event severity was graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events version 5.0. Adverse events occurring in ≥15% of patients in either treatment arm and grade 3-4 adverse events Clinical cut-off for this analysis was 29 September 2023. Adverse events are preferred terms, not grouped terms. Adverse event severity was graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events version 5.0. Grade 3-4 adverse events occurred in 143 of 162 (88.3%) patients in the inavolisib arm and 133 of 162 (82.1%) in the placebo arm.2 The most common (by preferred term) were neutropenia [77 of 162 (47.5%) and 78 of 162 (48.1%), respectively], neutrophil count decreased [56 of 162 (34.6%) and 53 of 162 (32.7%)] and white blood cell count decreased [23 of 162 (14.2%) and 17 of 162 (10.5%)] (Table 1). Serious adverse events occurred in 39 of 162 (24.1%) patients in the inavolisib arm and 17 of 162 (10.5%) in the placebo arm.2 These were distributed across multiple preferred terms, and none had a ≥2% difference between arms.2 Grade 5 adverse events were reported in 6 of 162 (3.7%) patients in the inavolisib arm and 2 of 162 (1.2%) in the placebo arm.2 None were reported as related to any study drug by the investigator, and no preferred term was observed in >1 patient in either arm. In the inavolisib arm, 112 of 162 (69.1%) patients had an inavolisib dose interruption due to an adverse event, 115 of 162 (71.0%) had a palbociclib dose interruption due to an adverse event and 52 of 162 (32.1%) had a fulvestrant dose interruption due to an adverse event. The median duration of inavolisib dose interruption was 1 day (range 1.0-64.0). Preferred terms for inavolisib dose interruptions in ≥10% of patients were hyperglycaemia [40 of 162 (24.7%)], COVID-19 [26 of 162 (16.0%)], neutropenia [20 of 162 (12.3%)] and neutrophil count decreased [17 of 162 (10.5%)]. In the inavolisib arm, 23 of 162 (14.2%) patients had an inavolisib dose reduction due to an adverse event, and 61 of 162 (37.7%) had a palbociclib dose reduction due to an adverse event. Preferred terms leading to inavolisib dose reductions in >2 patients were hyperglycaemia [4 of 162 (2.5%)] and stomatitis [3 of 162 (1.9%)]. Discontinuation rates due to adverse events were generally low. In the inavolisib arm, 10 of 162 (6.2%) patients discontinued inavolisib, 8 of 162 (4.9%) discontinued palbociclib and 5 of 162 (3.1%) discontinued fulvestrant.2 The only preferred term leading to withdrawal of any study drug in ≥2 patients was hyperglycaemia [inavolisib arm; 2 of 162 patients (1.2%)]. Subgroups Tables 2 and 3 report an overview of safety by region (North America or Western Europe, Asia, and Other) and age [<65 years and ≥65 years; the ≥75-year group was excluded due to too few patients (five and six in the inavolisib and placebo arms, respectively)].Table 2Safety overview by geographical regionGeographical regionl patients with at least one:Inavolisib + palbociclib + fulvestrantPlacebo + palbociclib + fulvestrantAsia (n = 57)North America/Western Europe (n = 63)Other (n = 42)Asia (n = 57)North America/Western Europe (n = 63)Other (n = 42)Any-grade adverse event, n (%)57 (100)63 (100)40 (95.2)57 (100)63 (100)42 (100) Total number of adverse events, n12621369667693905460Treatment-related adverse event, n (%) Inavolisib/placebo53 (93.0)58 (92.1)31 (73.8)44 (77.2)49 (77.8)23 (54.8) Palbociclib55 (96.5)59 (93.7)38 (90.5)54 (94.7)61 (96.8)38 (90.5) Fulvestrant28 (49.1)40 (63.5)16 (38.1)21 (36.8)41 (65.1)10 (23.8) Any treatment57 (100)62 (98.4)39 (92.9)54 (94.7)61 (96.8)38 (90.5)Grade 3-4 adverse event, n (%)55 (96.5)54 (85.7)34 (81.0)48 (84.2)53 (84.1)32 (76.2)Grade 5 adverse event, n (%)2 (3.5)1 (1.6)3 (7.1)1 (1.8)01 (2.4)Serious adverse event, n (%)16 (28.1)8 (12.7)15 (35.7)5 (8.8)6 (9.5)6 (14.3)Adverse event leading to withdrawal from the study, n (%)1 (1.8)00000Adverse event leading to withdrawal from treatment, n (%) Inavolisib/placebo2 (3.5)6 (9.5)2 (4.8)01 (1.6)0 Palbociclib3 (5.3)4 (6.3)1 (2.4)000 Fulvestrant2 (3.5)2 (3.2)1 (2.4)000 Any treatment3 (5.3)6 (9.5)2 (4.8)01 (1.6)0Adverse event leading to dose modification/interruption of treatment, n (%) Inavolisib/placebo40 (70.2)45 (71.4)28 (66.7)15 (26.3)31 (49.2)11 (26.2) Palbociclib51 (89.5)46 (73.0)28 (66.7)42 (73.7)47 (74.6)27 (64.3) Fulvestrant14 (24.6)27 (42.9)11 (26.2)8 (14.0)18 (28.6)8 (19.0) Any treatment52 (91.2)50 (79.4)32 (76.2)44 (77.2)48 (76.2)29 (69.0)Clinical cut-off for this analysis was 29 September 2023. Adverse event severity was graded according to National Cancer Institute Common Terminology Criteria for Adverse Events version 5.0.Table 3Safety overview by ageAge group; patients with at least one:Inavolisib + palbociclib + fulvestrantPlacebo + palbociclib + fulvestrant<65 years (n = 138)65-74 years (n = 19)<65 years (n = 128)65-74 years (n = 28)Any-grade adverse event, n (%)136 (98.6)19 (100)128 (100)28 (100) Total number of adverse events, n29162951654359Treatment-related adverse event, n (%) Inavolisib/placebo121 (87.7)17 (89.5)90 (70.3)20 (71.4) Palbociclib131 (94.9)17 (89.5)119 (93.0)28 (100) Fulvestrant74 (53.6)9 (47.4)52 (40.6)17 (60.7) Any treatment136 (98.6)18 (94.7)119 (93.0)28 (100)Grade 3-4 adverse event, n (%)123 (89.1)16 (84.2)104 (81.2)23 (82.1)Grade 5 adverse event, n (%)2 (1.4)2 (10.5)2 (1.6)0Serious adverse event, n (%)29 (21.0)7 (36.8)12 (9.4)4 (14.3)Adverse event leading to withdrawal from the study, n (%)01 (5.3)00Adverse event leading to withdrawal from treatment, n (%) Inavolisib/placebo8 (5.8)2 (10.5)1 (0.8)0 Palbociclib6 (4.3)2 (10.5)00 Fulvestrant4 (2.9)1 (5.3)00 Any treatment8 (5.8)3 (15.8)1 (0.8)0Adverse event leading to dose modification/interruption of treatment, n (%) Inavolisib/placebo94 (68.1)15 (78.9)41 (32.0)12 (42.9) Palbociclib105 (76.1)16 (84.2)90 (70.3)20 (71.4) Fulvestrant43 (31.2)6 (31.6)23 (18.0)9 (32.1) Any treatment113 (81.9)17 (89.5)95 (74.2)20 (71.4)Clinical cut-off for this analysis was 29 September 2023.Adverse event severity was graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events version 5.0. Table 2Safety overview by geographical regionGeographical regionl patients with at least one:Inavolisib + palbociclib + fulvestrantPlacebo + palbociclib + fulvestrantAsia (n = 57)North America/Western Europe (n = 63)Other (n = 42)Asia (n = 57)North America/Western Europe (n = 63)Other (n = 42)Any-grade adverse event, n (%)57 (100)63 (100)40 (95.2)57 (100)63 (100)42 (100) Total number of adverse events, n12621369667693905460Treatment-related adverse event, n (%) Inavolisib/placebo53 (93.0)58 (92.1)31 (73.8)44 (77.2)49 (77.8)23 (54.8) Palbociclib55 (96.5)59 (93.7)38 (90.5)54 (94.7)61 (96.8)38 (90.5) Fulvestrant28 (49.1)40 (63.5)16 (38.1)21 (36.8)41 (65.1)10 (23.8) Any treatment57 (100)62 (98.4)39 (92.9)54 (94.7)61 (96.8)38 (90.5)Grade 3-4 adverse event, n (%)55 (96.5)54 (85.7)34 (81.0)48 (84.2)53 (84.1)32 (76.2)Grade 5 adverse event, n (%)2 (3.5)1 (1.6)3 (7.1)1 (1.8)01 (2.4)Serious adverse event, n (%)16 (28.1)8 (12.7)15 (35.7)5 (8.8)6 (9.5)6 (14.3)Adverse event leading to withdrawal from the study, n (%)1 (1.8)00000Adverse event leading to withdrawal from treatment, n (%) Inavolisib/placebo2 (3.5)6 (9.5)2 (4.8)01 (1.6)0 Palbociclib3 (5.3)4 (6.3)1 (2.4)000 Fulvestrant2 (3.5)2 (3.2)1 (2.4)000 Any treatment3 (5.3)6 (9.5)2 (4.8)01 (1.6)0Adverse event leading to dose modification/interruption of treatment, n (%) Inavolisib/placebo40 (70.2)45 (71.4)28 (66.7)15 (26.3)31 (49.2)11 (26.2) Palbociclib51 (89.5)46 (73.0)28 (66.7)42 (73.7)47 (74.6)27 (64.3) Fulvestrant14 (24.6)27 (42.9)11 (26.2)8 (14.0)18 (28.6)8 (19.0) Any treatment52 (91.2)50 (79.4)32 (76.2)44 (77.2)48 (76.2)29 (69.0)Clinical cut-off for this analysis was 29 September 2023. Adverse event severity was graded according to National Cancer Institute Common Terminology Criteria for Adverse Events version 5.0. Safety overview by geographical region Clinical cut-off for this analysis was 29 September 2023. Adverse event severity was graded according to National Cancer Institute Common Terminology Criteria for Adverse Events version 5.0. Table 3Safety overview by ageAge group; patients with at least one:Inavolisib + palbociclib + fulvestrantPlacebo + palbociclib + fulvestrant<65 years (n = 138)65-74 years (n = 19)<65 years (n = 128)65-74 years (n = 28)Any-grade adverse event, n (%)136 (98.6)19 (100)128 (100)28 (100) Total number of adverse events, n29162951654359Treatment-related adverse event, n (%) Inavolisib/placebo121 (87.7)17 (89.5)90 (70.3)20 (71.4) Palbociclib131 (94.9)17 (89.5)119 (93.0)28 (100) Fulvestrant74 (53.6)9 (47.4)52 (40.6)17 (60.7) Any treatment136 (98.6)18 (94.7)119 (93.0)28 (100)Grade 3-4 adverse event, n (%)123 (89.1)16 (84.2)104 (81.2)23 (82.1)Grade 5 adverse event, n (%)2 (1.4)2 (10.5)2 (1.6)0Serious adverse event, n (%)29 (21.0)7 (36.8)12 (9.4)4 (14.3)Adverse event leading to withdrawal from the study, n (%)01 (5.3)00Adverse event leading to withdrawal from treatment, n (%) Inavolisib/placebo8 (5.8)2 (10.5)1 (0.8)0 Palbociclib6 (4.3)2 (10.5)00 Fulvestrant4 (2.9)1 (5.3)00 Any treatment8 (5.8)3 (15.8)1 (0.8)0Adverse event leading to dose modification/interruption of treatment, n (%) Inavolisib/placebo94 (68.1)15 (78.9)41 (32.0)12 (42.9) Palbociclib105 (76.1)16 (84.2)90 (70.3)20 (71.4) Fulvestrant43 (31.2)6 (31.6)23 (18.0)9 (32.1) Any treatment113 (81.9)17 (89.5)95 (74.2)20 (71.4)Clinical cut-off for this analysis was 29 September 2023.Adverse event severity was graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events version 5.0. Safety overview by age Clinical cut-off for this analysis was 29 September 2023. Adverse event severity was graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events version 5.0. Key selected adverse events Key selected adverse events by month of exposure and median duration of key selected adverse events in the inavolisib arm are shown in Figure 1A-E, respectively. Key selected events occurred within the first 30 days of therapy and were generally of short duration. Therefore, careful monitoring during the first cycle and adherence to management guidelines are highly recommended.Figure 1(A-D) Key selected adverse events per month of exposurea and (E) median duration (range) of key selected adverse events in the inavolisib arm.baIf an adverse event was resolved and recurred in the same patient, it was not included a second time in this dataset. bDuration of adverse events was calculated using imputed start and end dates. The placebo arm, not shown as one of the study arms, is not comparable due to differences in patient populations. (A-D) Key selected adverse events per month of exposurea and (E) median duration (range) of key selected adverse events in the inavolisib arm.baIf an adverse event was resolved and recurred in the same patient, it was not included a second time in this dataset. bDuration of adverse events was calculated using imputed start and end dates. The placebo arm, not shown as one of the study arms, is not comparable due to differences in patient populations. A summary of key selected adverse event management strategies used is shown in the Supplement, available at https://doi.org/10.1016/j.esmoop.2026.107735. Hyperglycaemia In the SAS, approximately 40% of patients in the inavolisib (69 of 162) and placebo arms (66 of 162) had baseline risk factors for hyperglycaemia, including HbA1c ≥5.7% [16 of 162 (9.9%) and 21 of 162 (13.0%)], fasting glucose ≥100 mg/dL (5.6 mmol/L) [46 of 162 (28.4%) and 36 of 162 (22.2%)] and body mass index (BMI) ≥30 kg/m2 [29 of 162 (17.9%) and 28 of 162 (17.3%)].2 Sixteen of 162 (9.9%) and 20 of 162 (12.3%) met the American Diabetes Association prediabetes definition at baseline (HbA1c ≥5.7% to <6.5%) in the inavolisib and placebo arms, respectively. Hyperglycaemia (grouped term) was reported in 58.6% of patients (95 of 162) in the inavolisib arm (per NCI-CTCAE v5.0); most events were grade 1 [26 of 162 (16.0%)] or grade 2 [60 of 162 (37.0%)]; 5.6% (9 of 162) experienced a grade 3 event; none experienced a serious or grade 4-5 event.2 The preferred terms ‘hyperglycaemia’ and ‘blood glucose increased’ were reported in 53.7% (87 of 162) and 4.9% (8 of 162), respectively. No patients had diabetic ketoacidosis. Among the inavolisib-treated patients who had hyperglycaemia events, median time to first onset was 7.0 days (range 2.0-955.0); median time to resolution was 16.0 days (interquartile range 5.0-50.0). In the 157 patients with a laboratory baseline fasting glucose value and ≥1 posttreatment value in the inavolisib arm, high postbaseline fasting glucose was classed as grade 1 using NCI-CTCAE v4.0 in 80 of 157 (51.0%); grade 2, in 35 of 157 (22.3%); grade 3, in 18 of 157 (11.5%); and grade 4, in one of 157 (0.6%) patients. For patients who experienced an increased fasting glucose of >160 mg/dL (8.9 mmol/L), 96.3% (52 of 54) had an improvement in fasting glucose of ≥1 grade level with a median time to improvement from the first event of 8 days (range 2-43). Inavolisib dose interruptions, reductions and discontinuations due to hyperglycaemia were reported in 27.2% (44 of 162), 2.5% (4 of 162) and 0.6% (1 of 162) of patients, respectively. Of the 95 patients who had hyperglycaemia reported in the inavolisib arm, 75 (78.9%) reported an event as resolved and 20 (21.1%) did not have a resolved outcome by clinical cut-off (some patients were enrolled close to the cut-off date and adverse event follow-up is ongoing). Of the 66 patients who received ≥1 concomitant medication for management (prophylaxis and adverse event treatment) of hyperglycaemia in the inavolisib arm, metformin, either as a single agent and in combination with other antihyperglycaemic agents, was the most common medication (62 of 66; 93.9%). Twelve of 162 patients (7.4%) received prophylactic metformin (defined as flagged by the investigator or as patients who started metformin on/before cycle 1, day 1). Eleven of 162 patients (6.8%) received insulin during the study; median duration was 5.0 days (range 1.0-539.0; the high upper range was due to two patients with chronic insulin use). No other concomitant medications for hyperglycaemia were reported in ≥5% of patients (Supplement, available at https://doi.org/10.1016/j.esmoop.2026.107735). In the inavolisib arm, 38 (23.5%), 19 (11.7%), 14 (8.6%) and 3 of 162 patients (1.9%) received one, two, three and four unique concomitant medications for hyperglycaemia. Hyperglycaemia incidence by risk factor at baseline is shown in Table 4.Table 4Hyperglycaemia risk factor safety analysesRisk factor/number of risk factorsInavolisib + palbociclib + fulvestrant (n = 162)nPatients with a hyperglycaemia event (any grade), n (%)Patients with a hyperglycaemia event (grade 3–4), n (%)aOverall216295 (58.6)9 (5.6)BMI <30 kg/m213275 (56.8)4 (3.0) ≥30 kg/m22919 (65.5)5 (17.2) Missing11 (100)0HbA1c <5.7%14587 (60.0)8 (5.5) ≥5.7%167 (43.8)1 (6.2) Missing11 (100)0Fasting blood glucose <100 mg/dL (<5.6 mmol/L)11259 (52.7)4 (3.6) ≥100 mg/dL (≥5.6 mmol/L)4634 (73.9)5 (10.9) Missing42 (50.0)0Number of risk factors 09349 (52.7)2 (2.2) 15034 (68.0)4 (8.0) 21610 (62.5)2 (12.5) 332 (66.7)1 (33.3)Clinical cut-off for this analysis was 29 September 2023.Adverse event severity was graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events version 5.0.Percentages are based on the number of patients in each subgroup.BMI, body mass index; HbA1c, glycated haemoglobin.aThere were no grade 4 events. Table 4Hyperglycaemia risk factor safety analysesRisk factor/number of risk factorsInavolisib + palbociclib + fulvestrant (n = 162)nPatients with a hyperglycaemia event (any grade), n (%)Patients with a hyperglycaemia event (grade 3–4), n (%)aOverall216295 (58.6)9 (5.6)BMI <30 kg/m213275 (56.8)4 (3.0) ≥30 kg/m22919 (65.5)5 (17.2) Missing11 (100)0HbA1c <5.7%14587 (60.0)8 (5.5) ≥5.7%167 (43.8)1 (6.2) Missing11 (100)0Fasting blood glucose <100 mg/dL (<5.6 mmol/L)11259 (52.7)4 (3.6) ≥100 mg/dL (≥5.6 mmol/L)4634 (73.9)5 (10.9) Missing42 (50.0)0Number of risk factors 09349 (52.7)2 (2.2) 15034 (68.0)4 (8.0) 21610 (62.5)2 (12.5) 332 (66.7)1 (33.3)Clinical cut-off for this analysis was 29 September 2023.Adverse event severity was graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events version 5.0.Percentages are based on the number of patients in each subgroup.BMI, body mass index; HbA1c, glycated haemoglobin.aThere were no grade 4 events. Hyperglycaemia risk factor safety analyses Clinical cut-off for this analysis was 29 September 2023. Adverse event severity was graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events version 5.0. Percentages are based on the number of patients in each subgroup. BMI, body mass index; HbA1c, glycated haemoglobin. There were no grade 4 events. Subgroups In the prediabetic population enrolled in the inavolisib arm, 7 of 16 (43.8%) patients experienced any-grade hyperglycaemia [3 each at grade 1 and 2 (18.8%); 1 (6.3%) at grade 3], and none discontinued inavolisib due to hyperglycaemia. Metformin was the most common antihyperglycaemic used. In the inavolisib arm, 36 of 57 (63.2%), 34 of 63 (54.0%) and 25 of 42 (59.5%) patients in Asia, North America or Western Europe, and Other regions, respectively, experienced hyperglycaemia, as did 82 of 138 (59.4%) and 10 of 19 (52.6%) patients aged <65 and 65-74 years, respectively. Rash Rash (grouped term) was reported in 41 of 162 (25.3%) patients in the inavolisib arm2; all events were nonserious and grade 1 [34 of 162 (21.0%)] or grade 2 [7 of 162 (4.3%)]; there were no grade 3-5 events. ‘Rash’ [26 of 162 (16.0%)] and ‘rash maculo-papular’ [7 of 162 (4.3%)] were the most common preferred terms. No severe cutaneous adverse reactions {e.g. Stevens–Johnson syndrome, erythema multiforme, toxic epidermal necrolysis or drug hypersensitivity syndrome [also known as drug rash with eosinophilia and systemic symptoms (DRESS)]} were reported. Among the inavolisib-treated patients who had rash events, the median time to first onset was 29.0 days (range 1.0-952.0). Median time to resolution was 16.0 days (interquartile range 8.0-29.0). Of the 41 of 162 (25.3%) patients who experienced rash in the inavolisib arm, 35 of 41 (85.4%) reported an event as resolved. Inavolisib dose interruptions, reductions and discontinuations due to rash were reported in 1.2% (2 of 162), 0.6% (1 of 162) and 0% (0 of 162) of patients, respectively. Of the 26 patients who received ≥1 concomitant medication for rash in the inavolisib arm, the most common medications were topical hydrocortisone (5 of 26; 19.2%) and triamcinolone acetonide (4 of 26; 15.4%). Subgroups In the inavolisib arm, 15 of 57 (26.3%), 16 of 63 (25.4%) and 10 of 42 (23.8%) patients in Asia, North America or Western Europe, and Other regions, respectively, experienced rash, as did 39 of 138 (28.3%) and 2 of 19 (10.5%) patients aged <65 and 65-74 years, respectively. Stomatitis/mucosal inflammation Stomatitis/mucosal inflammation (grouped term) was reported in 83 of 162 (51.2%) patients in the inavolisib arm, mostly at grade 1 [52 of 162 (32.1%)] or grade 2 [22 of 162 (13.6%)]; 9 of 162 (5.6%) patients experienced a grade 3 event and none experienced a grade 4-5 event.2 ‘Stomatitis’ [53 of 162 (32.7%)] and ‘mucosal inflammation’ [30 of 162 (18.5%)] were the most common preferred terms. Among the inavolisib-treated patients who had stomatitis/mucosal inflammation events, the median time to first onset was 13.0 days (range 1.0-610.0). Median time to resolution was 20.0 days (interquartile range 10.0-49.0). Of the 83 of 162 (51.2%) patients who experienced stomatitis/mucosal inflammation in the inavolisib arm, 68 of 83 (81.9%) reported an event as resolved. Inavolisib dose interruptions, reductions, and discontinuations due to stomatitis/mucosal inflammation were reported in 9.9% (16 of 162), 3.7% (6 of 162) and 0.6% (1 of 162) of patients, respectively. Of the 69 patients who received ≥1 concomitant medication for stomatitis/mucosal inflammation in the inavolisib arm, the most common medications were dexamethasone [31 of 69 (44.9%)], sodium bicarbonate [7 of 69 (10.1%)] and nystatin [7 of 69 (10.1%)] (Supplement, available at https://doi.org/10.1016/j.esmoop.2026.107735). The majority were used as topical formulations (mouthwash). Thirty-two of 162 patients (19.8%) in the inavolisib arm received prophylactic dexamethasone/triamcinolone mouthwash/topical formulations. Subgroups In the inavolisib arm, 35 of 57 (61.4%), 31 of 63 (49.2%) and 17 of 42 (40.5%) patients in Asia, North America or Western Europe and Other regions, respectively, experienced stomatitis/mucosal inflammation, as did 76 of 138 (55.1%) and 6 of 19 (31.6%) patients aged <65 and 65-74 years, respectively. Diarrhoea Diarrhoea was reported in 78 of 162 (48.1%) patients in the inavolisib arm2; most patients had grade 1 [45 of 162 (27.8%)] or grade 2 diarrhoea events [27 of 162 (16.7%)]. Grade 3 events were only reported in the inavolisib arm [6 of 162 (3.7%) patients], and no grade 4-5 events or colitis were reported in either arm. Two of 162 (1.2%) patients in the inavolisib arm experienced serious events2; both were grade 3 and were reported as recovered or resolved at clinical cut-off. One of the events was considered unrelated to any study drug, required treatment and lasted 7 days. The other was considered related to inavolisib only, did not require antidiarrhoeal treatment (intravenous hydration and electrolytes provided) and lasted 5 days. Among the inavolisib-treated patients who had diarrhoea events, median time to first onset was 15.0 days (range 2.0-602.0); median time to resolution was 5.0 days (interquartile range 2.0-19.0). Of the 78 of 162 (48.1%) patients who experienced diarrhoea in the inavolisib arm, 68 of 78 (87.2%) reported an event as resolved. Inavolisib dose interruptions, reductions and discontinuations due to diarrhoea were reported in 6.8% (11 of 162), 1.2% (2 of 162) and 0% (0 of 162) of patients, respectively. Of the 46 patients who received ≥1 concomitant medication for diarrhoea in the inavolisib arm, the most common medications were loperamide [33 of 46 (82.6%)] and diosmectite [6 of 46 (13.0%)] (Supplement, available at https://doi.org/10.1016/j.esmoop.2026.107735). No escalation treatment (e.g. steroids and opioids) was required in any patient. Subgroups In the inavolisib arm, 24 of 57 (42.1%), 37 of 63 (58.7%) and 17 of 42 (40.5%) patients in Asia, North America or Western Europe and Other regions, respectively, experienced diarrhoea, as did 70 of 138 (50.7%) and 7 of 19 (36.8%) aged <65 and 65-74 years, respectively. Other selected adverse events Ocular toxicities were reported in 36 of 162 (22.2%) patients in the inavolisib arm; all were grade 1-22 and nonserious; the most common toxicities were dry eye [14 of 162 (8.6%)], blurred vision [6 of 162 (3.7%)] and increased lacrimation [4 of 162 (2.5%)]. Ocular toxicity events did not lead to interruption, reduction or withdrawal of any study drug. Pneumonitis (an overlapping toxicity with palbociclib) was reported in 3 of 162 (1.9%) patients in the inavolisib arm; all were grade 1-2 and nonserious. Pneumonitis led to inavolisib dose interruption in 1 of 162 (0.6%) patients but did not lead to dose reduction or withdrawal of any study drug. Blood creatinine increased was reported in 9 of 162 (5.6%) patients in the inavolisib arm; one event was grade 3. No treatment-emergent colitis was reported in the study. Long-term safety In the inavolisib arm, 69 of 162 (42.6%) patients received inavolisib treatment for at least 1 year at the clinical cut-off date of 2 February 2024. An overview of new-onset adverse events starting after 1 year of treatment in this population is reported in Table 5. No adverse events leading to withdrawal of study treatment occurred in the long-term safety population at the clinical cut-off date.Table 5Long-term safety overview: new-onset adverse events reported only after 1 year on treatment, with at least 1 year on treatment with inavolisib/placebo + palbociclib + fulvestrantPatients with at least one:Inavolisib + palbociclib + fulvestrant (n = 69)Placebo + palbociclib + fulvestrant (n = 36)Any-grade adverse event, n (%)57 (82.6)28 (77.8) Total number of adverse events, n303128Treatment-related adverse event, n (%) Inavolisib/placebo34 (49.3)8 (22.2) Palbociclib35 (50.7)10 (27.8) Fulvestrant14 (20.3)4 (11.1) Any treatment44 (63.8)13 (36.1)Grade 3-4 adverse event, n (%)18 (26.1)5 (13.9)Grade 5 adverse event, n (%)2 (2.9)0Serious adverse event, n (%)8 (11.6)2 (5.6)Adverse event leading to withdrawal from treatment, n (%) Inavolisib/placebo00 Palbociclib00 Fulvestrant00 Any treatment00Adverse event leading to dose modification/interruption of treatment, n (%) Inavolisib/placebo24 (34.8)7 (19.4) Palbociclib25 (36.2)9 (25.0) Fulvestrant13 (18.8)4 (11.1) Any treatment28 (40.6)9 (25.0)Clinical cut-off date for this analysis was 2 February 2024.Adverse event severity was graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events version 5.0. Table 5Long-term safety overview: new-onset adverse events reported only after 1 year on treatment, with at least 1 year on treatment with inavolisib/placebo + palbociclib + fulvestrantPatients with at least one:Inavolisib + palbociclib + fulvestrant (n = 69)Placebo + palbociclib + fulvestrant (n = 36)Any-grade adverse event, n (%)57 (82.6)28 (77.8) Total number of adverse events, n303128Treatment-related adverse event, n (%) Inavolisib/placebo34 (49.3)8 (22.2) Palbociclib35 (50.7)10 (27.8) Fulvestrant14 (20.3)4 (11.1) Any treatment44 (63.8)13 (36.1)Grade 3-4 adverse event, n (%)18 (26.1)5 (13.9)Grade 5 adverse event, n (%)2 (2.9)0Serious adverse event, n (%)8 (11.6)2 (5.6)Adverse event leading to withdrawal from treatment, n (%) Inavolisib/placebo00 Palbociclib00 Fulvestrant00 Any treatment00Adverse event leading to dose modification/interruption of treatment, n (%) Inavolisib/placebo24 (34.8)7 (19.4) Palbociclib25 (36.2)9 (25.0) Fulvestrant13 (18.8)4 (11.1) Any treatment28 (40.6)9 (25.0)Clinical cut-off date for this analysis was 2 February 2024.Adverse event severity was graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events version 5.0. Long-term safety overview: new-onset adverse events reported only after 1 year on treatment, with at least 1 year on treatment with inavolisib/placebo + palbociclib + fulvestrant Clinical cut-off date for this analysis was 2 February 2024. Adverse event severity was graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events version 5.0. Further details on the long-term safety profile have been reported.9 Discussion The inavolisib-based regimen had a safety profile consistent with the individual drugs and a low treatment discontinuation rate due to adverse events.2, 3, 4 Updated data with additional follow-up were consistent. Hyperglycaemia, rash, stomatitis/mucosal inflammation and diarrhoea were mostly low grade, manageable with protocol-recommended supportive measures and most resolved (reflected in the high dose intensity of each drug).2,4 The overall proportion of patients with ≥1 adverse event, events leading to inavolisib dose modification or interruption and events leading to study drug withdrawal were largely comparable across regions, supporting consistent manageability and tolerability of the inavolisib-based regimen. Fatal events were infrequent across regions, and serious events were mostly reported in the Other regions, followed by Asia and North America or Western Europe. There was a numerically higher incidence of grade 3-4 adverse events in Asia, compared with North America or Western Europe and Other regions. The proportion of patients experiencing hyperglycaemia and stomatitis/mucosal inflammation was numerically higher in Asia compared with Other regions. Proportions experiencing rash were consistent across regions, and diarrhoea was reported in a numerically higher proportion in North America or Western Europe than in Asia or Other regions. Patients were relatively younger in INAVO120, as it recruited a first-line, poor-prognosis endocrine-resistant population. Adverse events leading to inavolisib dose interruption or modification and events leading to study drug withdrawal were largely comparable across age groups, supporting consistent manageability and tolerability of the inavolisib-based regimen. Overall, the proportions of serious adverse events and events leading to dose interruption or modification of any study treatment were numerically higher in patients aged 65-74 years compared with those aged <65 years, which could be due to a higher prevalence of comorbidities in elderly patients and/or a reduction in the reserve capacity of the organs in general.10 Fatal adverse events were rare across all age groups. Hyperglycaemia, an on-target toxicity associated with PI3K pathway inhibitors, was more frequent in the inavolisib arm and occurred mostly in the first three treatment cycles, highlighting the importance of early, frequent fasting blood or plasma glucose monitoring and following hyperglycaemia management guidelines. Hyperglycaemia caused by inavolisib was an on-target effect that resolved upon treatment cessation, thereby eliminating the need for further antihyperglycaemic drugs (mostly metformin; a small number of patients received short-term insulin) in the majority of patients. Overall, hyperglycaemia was manageable and did not lead to a high proportion of patients needing inavolisib dose reductions or discontinuations. The incidence of hyperglycaemia in the inavolisib arm was higher in patients with one, two or three risk factors, namely fasting glucose ≥100 mg/dL (≥5.6 mmol/L), HbA1c ≥5.7% and BMI ≥30 kg/m2 at baseline, compared with patients with none. INAVO120 excluded patients who had fasting glucose ≥126 mg/dL (≥7.0 mmol/L) and HbA1c ≥6.0% (originally ≥5.7%), and patients with type 1 and 2 diabetes on systemic therapy. GO39374 (NCT03006172), however, enrolled patients with HbA1c <7% and fasting glucose <140 mg/dL.11 GO39374 showed that in patients with risk factors for hyperglycaemia (HbA1c ≥5.7% and <6.5%; fasting glucose ≥100 mg/dL and <126 mg/dL; BMI ≥30 kg/m2), hyperglycaemia was frequent (81% of patients); inavolisib cumulative dose intensity was high (92%), with a low rate of discontinuation due to hyperglycaemia (1%); hyperglycaemia was manageable with dose interruptions (42%), reductions (14%) or oral antihyperglycaemic medications (64%) and was reversible (86%).12 Hyperglycaemia, being a known toxicity associated with PI3Kα inhibitors, has been reported with alpelisib13; however, cross-trial comparisons should be made with caution due to differences in trial design, patient populations, analysis, NCI-CTCAE versions used (4.0 in SOLAR-1 and 5.0 in INAVO120) and reporting. Diarrhoea, stomatitis/mucosal inflammation and rash are commonly associated with PI3K pathway inhibitors; these were also more frequent in the inavolisib arm but were early-onset, low-grade, nonserious and manageable with supportive care and dose modifications. Prophylactic dexamethasone mouthwash was recommended for stomatitis based on the SWISH study7 and it was the most common concomitant medication. No prophylaxis for rash was recommended in INAVO120 based on the inavolisib mechanism of action (high α-selectivity) and data from the GO39374 study,11 whereas for alpelisib in SOLAR-1, prophylaxis with antihistamines was recommended.13 Ocular toxicities were mild, with dry eye and blurred vision being the most common ocular events reported. There were similar incidence rates of haematological adverse events across both arms for overall incidence and at grade 3 and 4 (with slightly higher rates of anaemia and thrombocytopenia), known CDK4/6 inhibitor-induced toxicities. The higher proportion of patients experiencing COVID-19 in the inavolisib arm may have been due to the longer time on study treatment, resulting in longer adverse event collection and observation period; median exposure to inavolisib was 9.2 months compared with 5.6 months for placebo.2 Specific starting dose adjustment guidance for patients with moderate–severe renal impairment is provided in the label. The safety profile of inavolisib was consistent with that of long-term treated populations.9 The main limitation of the current analysis is the small subgroups. It should be noted in regional analyses that ‘Other’ comprised multiple regions with small patient numbers; therefore, interpretation should be made with caution. Differences in hyperglycaemia, stomatitis/mucosal inflammation and diarrhoea may also be related to compliance with management guidelines. Similarly, no meaningful conclusion can be made on the impact of HbA1c on hyperglycaemia incidence or on comparisons by age. In summary, this analysis of INAVO120 demonstrates the generally consistent, manageable and tolerable safety profile of inavolisib in combination with palbociclib and fulvestrant. No new or unexpected safety signals were observed, and key selected adverse events that are commonly associated with PI3K inhibitors were generally nonserious, low-grade and mostly occurred in early treatment cycles (as such, focus should be on early monitoring, prevention and treatment). These events were generally reversible and were controlled with concomitant medications and dose modifications. Subgroup analyses showed numerical differences in certain adverse events, but they were manageable overall, and the long-term safety profile was consistent with previous reports. These analyses report a more detailed evaluation of the safety profile of inavolisib in combination with palbociclib and fulvestrant and provide guidance in adverse event management for patients receiving this regimen. This will help tailor the adverse event management and maximise the clinical benefit of inavolisib.