Identification of Early Symptoms Associated with Subsequent Immune-related Adverse Events in the I-SPY clinical trial

Identification of Early Symptoms Associated with Subsequent Immune-related Adverse Events in the I-SPY clinical trial | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Identification of Early Symptoms Associated with Subsequent Immune-related Adverse Events in the I-SPY clinical trial Amrita Basu, Saumya Umashankar, Michelle Melisko, Ritu Roy, Christina Yau, and 28 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8689063/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background Immune checkpoint inhibitors can result in serious, long-lasting immune-related adverse events (irAEs). Early identification of symptoms predictive of irAEs could enhance monitoring and timely intervention. This study assessed whether symptoms within the first 8 weeks of treatment could predict subsequent development of immune-related adrenal insufficiency(AI) or hypothyroidism. Methods This retrospective cohort study analyzed prospectively collected data from the I-SPY2 trial, a phase 2 platform trial for high-risk stage II/III breast cancer across 30 U.S. sites. The cohort included 482 women treated with experimental immunotherapy agents concurrent with weekly paclitaxel neoadjuvant chemotherapy. The primary outcomes were grade ≥ 1 hypothyroidism or AI, adjudicated by an independent safety group, up to 1-year post-treatment. Symptoms and irAEs were assessed using the Common Terminology Criteria for Adverse Events. Symptom burden was quantified as area under the curve (AUC) based on symptom grade and duration. Predictive modeling was performed using logistic regression and ROC analysis; symptom enrichment between cases and controls was evaluated using Fisher’s exact tests. Results Among 482 participants, 107 (22.2%) developed irAEs, with hypothyroidism (n = 61, 12.7%) occurring more frequently than AI (n = 38, 7.9%) at medians of 99 and 105 days from treatment initiation, respectively. Symptom enrichment analysis identified early predictive symptoms. Fatigue (17.2% vs 6.8%, p = 0.011) and rash (20.7% vs 7.8%, p = 0.0037) were predictive of hypothyroidism, while diarrhea (45.9% vs 31%, p = 0.048), constipation (5.4% vs 0.2%, p = 0.018), and taste changes (5.4% vs 0.5%, p = 0.034) were associated with AI. A predictive model demonstrated moderate performance (AUC 0.65 for AI, p < 0.0001; AUC 0.61 for hypothyroidism, p = 0.012). Model accuracy in an external validation cohort was 72.8% for AI and 74.7% for hypothyroidism. Conclusions This study presents a predictive framework to identify patients at risk for adrenal insufficiency and hypothyroidism as irAEs, enabling personalized care and proactive intervention to improve treatment outcomes and safety. Biological sciences/Cancer Health sciences/Diseases Health sciences/Gastroenterology Health sciences/Medical research Health sciences/Oncology breast cancer adverse events immunotherapy hypothyroidism adrenal insufficiency hypophysitis Figures Figure 1 Figure 2 Figure 3 Background Recent advances in cancer immunotherapy, including the advent of immune checkpoint inhibitors (ICIs), have revolutionized cancer treatment, providing survival benefits across an array of cancer types, including breast cancer[ 1 , 2 ]. ICIs harness patients’ adaptive immune system to treat cancer by antagonizing the receptors responsible for dampening the effective activity of immune cells and their ability to identify and remove tumor cells[ 1 ]. For breast cancer, pembrolizumab is approved in both early and metastatic triple-negative disease[ 3 – 5 ]. In addition, early trials have shown promising results of ICIs in early-stage hormone receptor-positive breast cancers[ 6 – 10 ]. ICIs are associated with immune related adverse events (irAE) that are often long-term[ 11 – 14 ]. These likely have multiple etiologies, leading to a broad range of autoimmune syndromes including potentially life-threatening events[ 15 ]. A systemic review by Balibegloo and colleagues[ 1 ] reported that approximately 12% of patients experienced any grade hypothyroidism, and about 1% of patients developed any grade adrenal insufficiency (AI), while being treated with PD-1 or PD-L1 inhibitors in the metastatic breast cancer setting. However, their incidence is higher in the adjuvant or neoadjuvant setting where these drugs are part of the standard of care for some tumor types[ 16 ]. Although the incidence of grade 3 or higher endocrine toxicity is low, even lower grade toxicity is nearly always permanent[ 15 , 17 , 18 ] and requires life-long medical management[ 19 ]. Identification of patients at higher risk of developing irAEs during treatment could offer an opportunity for early detection and intervention to prevent or minimize the impact of these conditions. For example, adjunctive immunomodulatory agents might help prevent or attenuate irAEs, and patients predicted to be at high risk for irAEs who experience rapid treatment response might consider discontinuing ICI therapy early. We investigated the feasibility of early identification of patients at high risk of developing AI or hypothyroidism by analyzing the early symptom landscape. We used prospectively collected toxicity data from the ISPY-2 trial[ 16 ] and computed an overall symptom burden (OSB) score incorporating duration and severity of symptom to construct a predictive framework for clinicians to identify early symptoms that could forecast a later irAE, and conducted statistical enrichment analyses to uncover symptoms enriched in patients with and without the irAEs of interest. Methods Study Population The I-SPY2 clinical trial is an ongoing, multicenter, open-label, response-adaptive phase 2 platform trial with multiple experimental arms evaluating new agents alone or in combination with weekly neoadjuvant paclitaxel followed by 4 additional cycles of doxorubicin/cyclophosphamide (AC) chemotherapy for women with high-risk stage II/III breast cancer (NCT01042379). Information regarding design, eligibility, and assessments has been previously reported[ 20 ]. This analysis used data from 2247 women enrolled between 2010–2022 at 30 clinical centers across the US. We identified 482 participants who received ICI therapy (± other experimental agents) concurrent with 12 cycles of weekly paclitaxel (Supplementary Figure S2 ) in 6 arms of I-SPY2. Information on treatment regimens is included in Supplementary Table 1. Predictor Variables 32 symptoms that occurred on or before weeks 4, 6 and 8 were included as predictor variables. These were symptoms selected by a group of clinicians, researchers and patient advocates in the I-SPY patient outcomes working group as key symptoms to assess independently or for overall patient toxicity, quality of life or symptom burden. Adverse events in I-SPY2 are documented and graded using the Common Terminology Criteria for Adverse Events (CTCAEv4.0) weekly from treatment start till the study timepoint for symptom assessment (week 8). Start and end dates (if resolved) of the symptoms during this time frame were documented for each event. Fifty-five CTCAE categories were collapsed into 32 symptoms predictors (Supplementary Table 2). For symptoms graded on a 1–5 scale, grade 1 was excluded (deemed occasional/asymptomatic and did not interfere with activities of daily living). Grade 1 symptom events were included for dizziness, loss of fingernails or toenails, painful urination, and palpitations where events are not on a 5-point scale. Outcomes The primary outcome was the development of immune-related adverse AI (due to hypophysitis or primary adrenalitis) and hypothyroidism defined as grade 1 or higher at any time during the study, or through follow-up up until 1 year. These AEs were adjudicated by the I-SPY Safety Working Group, an external independent data safety and monitoring working group (CCSA), and endocrinologists (ZQ and RNC). Diagnoses for hypothyroidism were triggered through routine TSH and free T4 testing and symptom triggers. Diagnoses for AI were primarily symptom triggered, followed by cortisol and ACTH testing. These irAEs were selected because they are common and often become life-altering long-term irAEs. Groups considered for statistical comparison were patients that developed an irAE(s) of interest after treatment with experimental immunotherapy(case) vs those that did not(control). Statistical Analysis A variable importance analysis was performed in two ways, first to characterize the association of OSB with the development of irAEs, and second, to understand sentinel symptoms associated with the development of irAEs of interest. Overall Symptom Burden For each patient, cumulative area under the curves (AUC) were calculated for each of the 32 symptoms based on their grade and duration at weeks 4, 6 and 8 as follows \(\:AUC\left(symptom\:|\:Week\:n\right)=\:\sum\:_{day\:1}^{n\times\:7}AE\:grade\:\times\:\:days\:(in\:specified\:duration\:that\:AE\:persisted\) ) Patients that developed an irAE of interest before the analysis timepoint (i.e. week 4, 6 or 8) were censored from the analysis of that timepoint for that outcome of interest. Logistic regression was performed first between each different symptom AUC and the outcome of interest; those symptoms that had positive association and Wald test p-values < 0.2 were selected to calculate the OSB. The OSB is the sum of the selected symptoms. A logistic regression was then performed between the continuous OSB and outcome. For each time point, p-values were adjusted for multiple testing using the Bonferroni correction and an association was considered significant if the adjusted p-value was < 0.05. ROC curves were generated for analyses with significant association of the OSB with outcome. The model developed above was validated in an external dataset comprising of 4 distinct arms in ISPY2 of 217 patients. Methods for additional statistical analyses and detailed validation methods are detailed in Supplementary Methods. All statistical analyses and visualizations were performed using R version 4.2.2. Results Patient Population Among the I-SPY patients, 482 were randomized to one of the six arms including an immunotherapy component (median age = 47.3 years (range, 20–79 years)). 382 patients were white(79.3%) and 95 were non-white(19.7%). 78 patients reported Hispanic ethnicity (16.2%). Detailed demographics are summarized in Table 1 . Table 1 Demographic characteristics of patient population by irAE type. Overall Hypothyroidism (N = 61) No Hypothyroidism (N = 421) Adrenal insufficiency (N = 38) No Adrenal insufficiency (N = 444) Age Mean (SD) 47.6 (11.6) 46.4 (10.3) 47.8 (11.8) 48.5 (11.7) 47.6 (11.6) Median [Min, Max] 47.3 [20.0, 79.0] 45.5 [28.8, 71.0] 47.9 [20.0, 79.0] 49.5 [31.0, 79.0] 47.0 [20.0, 76.0] Race American Indian/ Alaskan Native 3 (0.6%) 1 (1.6%) 2 (0.5%) 0 (0%) 3 (0.7%) Asian 32 (6.6%) 2 (3.3%) 30 (7.1%) 1 (2.6%) 31 (7.0%) Black 60 (12.4%) 2 (3.3%) 58 (13.8%) 5 (13.2%) 55 (12.4%) White 382 (79.3%) 53 (86.9%) 329 (78.1%) 32 (84.2%) 350 (78.8%) Ethnicity Hispanic 78 (16.2%) 7 (11.5%) 71 (16.9%) 8 (21.1%) 70 (15.8%) Not Hispanic 402 (83.4%) 54 (88.5%) 348 (82.7%) 30 (78.9%) 372 (83.8%) Among the cohort, hypothyroidism was more common(N = 61 or 12.7%) vs AI (38 or 7.9%). 21 of 61 hypothyroidism patients had hyperthyroidism an average of 4 weeks prior to hypothyroidism diagnosis. AI was most often attributed to confirmed or probable hypophysitis(65.8%), probable primary AI(2.6%), AI of unknown type(15.8%), and reported AI by site but unconfirmed at time of analysis by endocrinologist(15.8%)(Fig. 1 A). No significant differences were observed in irAE incidence based on age, race or ethnicity. Early Symptoms Predictive of Adrenal Insufficiency AI was diagnosed at a median of 15 weeks(105 days) from treatment start(range = 6.6–41.6 weeks(46–291 days)), which was after completion of the immunotherapy phase and close to the end of all neoadjuvant chemotherapy(Supplementary Figure S2 ). One patient developed AI between weeks 4 and 6 and was censored from the weeks 6 and 8 analyses, while an additional patient developed hypothyroidism between weeks 6–8 and was censored from the week 8 analyses. OSB at week 6 was the timepoint most predictive of AI (adjusted p < 0.001) (Fig. 2 A). Symptoms used to calculate OSB at week 6 were constipation, diarrhea, headache, joint pain, shortness of breath, taste changes, and vomiting (Supplementary Table 3). The trajectory of these symptoms as swimmers plots in patients with and without AI are shown on Fig. 1 A. Results indicated that shortness of breath, for example, presented early and persisted through follow up, as did GI symptoms such as diarrhea. Patients who would later develop AI had multiple occurrences of bouts of diarrhea (instances > 2) compared to patients who did not. Patients who would later develop AI also had multiple co-occurring symptoms, the foremost of which was diarrhea and vomiting, while patients without AI rarely had these symptoms. These symptoms preceded the AI diagnoses by an average of 63 days (9 weeks). Model Performance A logistic regression was performed between the continuous OSB and outcome. The Youden’s best OSB threshold was 14.5 with a sensitivity of 0.57, specificity of 0.74 and accuracy of 0.73 (Supplementary Table 6 and Supplementary Figure S1 A). Mean OSB at week 4 of patients that developed AI was almost double that of patients that did not develop AI (mean(SD) cases = 13.66 (19.73) vs controls = 5.2 (10.92)), and this gap widened at weeks 6 (31.11(39.58) vs 10.6(10.67)) and 8 (49.81(57.79) vs 18.5(30.68)(Fig. 2 A and Supplementary Table 3). Model Validation In an external validation set of 217 patients receiving ICI with an oral paclitaxel backbone, 11 patients later developed AI. Patients were demographically comparable to the training set above (Supplementary Table 8). The OSB was computed as the sum AUCs of the symptoms mentioned above. Patients that later developed AI had a higher mean OSB at week 6 than patients that did not develop AI(44.6 vs 17.4, p = 0.02). Using the Youden’s best threshold above of 14.5, the model’s sensitivity was 54.6%, specificity was 73.8%, positive predictive value was 10%, negative predictive value was 96.8% and model accuracy was 72.8%. Longitudinal Symptom Enrichment All symptoms were also analyzed as binary variables (present vs absent) at 4, 6, and 8 weeks (Supplementary Table 4). Unlike OSB, this analysis did not capture severity and duration of a given symptom and was performed so that we could identify additional symptoms that may be short lasting and more severe, or long lasting and less severe and may not be captured by the OSB analysis. Such an example was heart palpitations that was significant at the week 8 timepoint (adjusted p = 0.047) but receded p > 0.05 in the OSB analysis (Fig. 3 A and Table 2 ). Interestingly, palpitations had similar rates of incidence in patients that developed AI compared to those that did not up until week 6, after which, patients that developed AI had a 4-fold higher rate of palpitations (8.3% in case vs. 1.9% in control). Table 2 Binary (presence absence) enrichment analysis for symptoms. Symptoms with p-value < 0.2 at at least one timepoint are shown, ordered by p-value at week 4, with timepoints with adjusted p-value < 0.05 highlighted in green. WEEK 4 WEEK 6 WEEK 8 ADRENAL INSUFFICIENCY Symptom In Cases (N = 38) In Control (N = 423) Odds Ratio Adjusted p-value In Cases (N = 37) In Control (N = 423) Odds Ratio Adjusted p-value In Cases (N = 36) In Control (N = 423) Odds Ratio Adjusted p-value Diarrhea 18 (47.4%) 112 (26.5%) 2.49 0.0069 17 (45.9%) 131 (31%) 1.89 0.048 17 (47.2%) 143 (33.8%) 1.75 0.077 Constipation 2 (5.3%) 1 (0.2%) 23.03 0.019 2 (5.4%) 1 (0.2%) 23.68 0.018 3 (8.3%) 3 (0.7%) 12.56 0.0075 Shortness of breath 6 (15.8%) 23 (5.4%) 3.25 0.024 7 (18.9%) 37 (8.7%) 2.43 0.051 6 (16.7%) 41 (9.7%) 1.86 0.15 Headache 4 (10.5%) 14 (3.3%) 3.42 0.052 4 (10.8%) 17 (4%) 2.88 0.079 4 (11.1%) 20 (4.7%) 2.51 0.11 Palpitations 2 (5.3%) 4 (0.9%) 5.78 0.08 2 (5.4%) 7 (1.7%) 3.38 0.16 3 (8.3%) 8 (1.9%) 4.69 0.047 Fatigue 4 (10.5%) 18 (4.3%) 2.64 0.097 6 (16.2%) 31 (7.3%) 2.44 0.065 5 (13.9%) 49 (11.6%) 1.23 0.42 Taste changes 1 (2.6%) 2 (0.5%) 5.65 0.23 2 (5.4%) 2 (0.5%) 11.88 0.034 2 (5.6%) 2 (0.5%) 12.22 0.032 Dizziness 3 (7.9%) 29 (6.9%) 1.16 0.5 6 (16.2%) 39 (9.2%) 1.9 0.14 6 (16.7%) 46 (10.9%) 1.64 0.21 HYPOTHYROIDISM Symptom In Cases (N = 58) In Control (N = 400) Odds Ratio Adjusted p-value In Cases (N = 58) In Control (N = 400) Odds Ratio Adjusted p-value In Cases (N = 50) In Control (N = 400) Odds Ratio Adjusted p-value Fatigue 6 (10.3%) 16 (4%) 2.76 0.047 10 (17.2%) 27 (6.8%) 2.87 0.011 8 (16%) 44 (11%) 1.54 0.2 Rash 7 (12.1%) 22 (5.5%) 2.35 0.059 12 (20.7%) 31 (7.8%) 3.09 0.0037 12 (24%) 34 (8.5%) 3.39 0.0021 Itching 2 (3.4%) 4 (1%) 3.52 0.17 3 (5.2%) 6 (1.5%) 3.57 0.093 3 (6%) 8 (2%) 3.12 0.11 Shortness of breath 5 (8.6%) 24 (6%) 1.48 0.3 9 (15.5%) 36 (9%) 1.85 0.097 6 (12%) 39 (9.8%) 1.26 0.38 A) In terms of symptoms that were overlapping between the OSB and binary analyses, prior to week 4, diarrhea was the most enriched symptom in patients that later developed AI, followed by constipation and shortness of breath (OR = 2.49, 23.03 and 3.25, adjusted p = 0.0069, 0.019 and 0.024 respectively). At week 4, diarrhea was present in 47.4% of the patients that later developed AI compared to 26.5% in patients who did not develop AI (Table 2 ). The difference in the proportions declined by week 6 (45.9% vs 31%) and was no longer significant at week 8 (47.2% vs 33.8%). Constipation was the only symptom that was significantly enriched at all three study timepoints. Constipation was present in twice as many patients who later developed AI than those without at weeks 4 and 6 (5.3–5.4% in cases vs < 1% in control) and in almost 8 times as many patients with AI by week 8 (8.3% vs. 0.7%). Patients that developed AI had almost twice the proportion of fatigue at weeks 4 (10.5% vs 4.3%) and 6 (16.2% vs 7.3%) compared to patients that did not develop AI, though this proportion nearly equalized by week 8 (13.9% vs 11.6%). Similarly, patients who later developed AI had almost three times the reported shortness of breath (15.8% vs 5.4% ) and headache (10.5% vs. 3.3%), by week 4. All analyses were also performed excluding cases that were reported as AI but were not confirmed by endocrinologists, with no significant differences in results. To further analyze the co-occurrence of symptoms, heatmaps were generated with 32 symptoms that developed by week 6 to visualize symptom clusters (Fig. 3 C). Diarrhea, vomiting, headache, fatigue and shortness of breath co-occurred in the majority of patients that later developed AI. Multiplicity of symptoms and irAE development For each timepoint, a logistic regression was performed to determine if presence of a higher number of symptoms was associated with greater odds for later developing AI (Supplementary Table 5). At week 4, the odds of developing AI with each additional symptom increased by 1.46 (adjusted p-value = 0.008, mean number of symptoms in cases = 1.32 vs. controls = 0.75). At week 6, the odds of developing AI with each additional symptom decreased to 1.33 (adjusted p-value = 0.027, in cases = 1.65 vs. controls = 1.01). Number of symptoms was no longer a significant predictor by week 8 due to increasing number of symptoms in controls as well (adjusted p-value = 0.07, in cases = 1.83 vs. controls = 1.49). Early Symptoms Predictive of Hypothyroidism Hypothyroidism was diagnosed at a median of 14 weeks (99 days) from treatment start (range = 20–208) which was close to the end of the immunotherapy phase of the treatment. Two patients developed hypothyroidism prior to week 4, and were censored from weeks 4,6 and 8 analyses, while an additional eight patients developed hypothyroidism between weeks 6 and 8 and were censored from the week 8 analyses. For hypothyroidism, week 6 was also the most significant timepoint in the predictive model (adjusted p = 0.012) (Fig. 2 B). Symptoms included in the OSB calculation (i.e. that were independently associated with hypothyroidism with p > 0.2) were painful urination, rash, shortness of breath and fatigue (Supplementary Table 3). The Fig. 1 B swimmers plot illustrates that higher-grade fatigue and rash were significantly enriched in the patients that later developed hypothyroidism. These symptoms presented earlier and lasted longer in those with immune related hypothyroidism. Multiple occurrences of rash, each relatively short duration, were more frequent, and fatigue was present for longer durations in patients with hypothyroidism. Shortness of breath was also more frequent and longer lasting in patients who subsequently developed hypothyroidism. These symptoms preceded the hypothyroidism diagnoses by an average of 57 days (approx. 8 weeks). Model Performance A logistic regression was performed between the continuous OSB and outcome followed by . The Youden’s best threshold OSB was 9.5 with a sensitivity of 0.43, specificity of 0.82 and accuracy of 0.77 (Supplementary Table 7 and Supplementary Figure S1 B). As early as week 4, the mean OSB for patients that had hypothyroidism was more than two times higher than for patients that did not develop hypothyroidism (mean(SD) for case = 13.66(19.73) vs. control = 5.2(10.92)), and this gap widened greatly through week 8 (case = 14.6(22.66) and 20.64(37.21) vs control = 6.7(16.97) and 9.27(24.09) at weeks 6 and 8, respectively)(Fig. 2 B, Supplementary Table 3). Model Validation In the validation set of 217 patients receiving ICI with an oral paclitaxel backbone, 31 patients later developed hypothyroidism. Patients that later developed hypothyroidism had a higher mean OSB at week 6 than patients that did not develop hypothyroidism, though this difference was not statistically significant(13.5 vs 10). Model’s sensitivity was 16.1%, specificity was 84.4%, positive predictive value was 14.7%, negative predictive value was 85.8% and model accuracy was 74.7%. Longitudinal Symptom Enrichment Binary analysis of symptoms showed that prior to week 4, fatigue was the only significantly enriched symptom among patients that later developed hypothyroidism(OR = 2.76, adjusted p = 0.047)(Table 2 , Supplementary Table 4). At week 6, rash was the most significantly enriched symptom, followed by fatigue(OR = 3.09 and 2.87, p = 0.0037 and 0.011), and at week 8, only rash was significantly enriched in patients that later developed hypothyroidism(OR = 3.39, p = 0.0021). Fatigue was present in more than twice as many patients that developed hypothyroidism compared to those that did not at weeks 4(10.3% vs 4.0%) and 6(17.2% vs 6.8%)(Fig. 3 B and Table 2 ); however, it was not significantly enriched in cases compared to controls at week 8, indicating that while early presence of fatigue can be an indicator of hypothyroidism, the growing incidence of fatigue among all patients undergoing treatment renders this an insignificant predictor as treatment progresses. Rash was present in twice as many patients at week 6 (20.7% vs 7.8%) and three times as many patients at week 8(24% vs 8.5%)(Fig. 3 B and Table 2 ). Further analysis and visualization of co-occurring symptoms revealed that diarrhea, rash, dizziness, and fatigue commonly co-occurred prior to week 6 among patients that later developed hypothyroidism(Fig. 3 D). However, unlike for AI, the number of symptoms was not a significant predictor of later development of hypothyroidism(Supplementary Table 5). Discussion ICIs are now a standard of care as neoadjuvant therapy combined with chemotherapy for stage II-III triple negative breast cancer, and for patients with PD-L1 positive metastatic TNBC. The current study utilized an analysis framework to identify unique symptom clusters that can be observed early on during ICI therapy and are predictive of developing the two most common irAEs, hypothyroidism and AI. Using prospectively collected clinical trial data from I-SPY2, we developed and validated a framework for early identification of patients at elevated risk for endocrine immune-related adverse events(irAEs). Our findings suggest that cumulative early symptom burden, especially by week 6 of treatment, provides meaningful predictive insight and could inform clinical strategies for early intervention. Identification of patients who are high risk for these irAEs could allow more frequent diagnostic testing (i.e. hormone level measurements) and proactive early treatment. While TSH is part of the standard of care monitoring, and cortisol level monitoring is now standard in I-SPY, by the time these tests are abnormal, neither AI nor hypothyroidism is likely reversible. Thus, the ability to find earlier indicators could potentially lead to proactive preventive interventions. Overall, our results demonstrated that rash, shortness of breath, and fatigue commonly co-occurred among patients who later developed hypothyroidism (after a median of 57 days), while diarrhea, headache, vomiting, fatigue and shortness of breath were predictive of AI (after a median of 63 days). Dermatological symptoms, such as itching and rash, were also highly ranked symptoms in the first 4 to 8 weeks that were predictive for the development of hypothyroidism, while gastrointestinal symptoms, such as constipation and diarrhea, were predictive for the development of AI. Diarrhea often precedes constipation in both cases and controls, as symptom management for diarrhea can frequently lead to constipation in patients. Fatigue remains one the most frequently reported side effects of cancer treatment and can be highly burdensome for patients. In the present study, fatigue that occurred early in treatment was a predictor of later development of hypothyroidism, however its significance dissipated as treatment continued and a high number of patients in both groups began to develop fatigue. This reinforces the need to consider symptoms longitudinally and in the context of other symptoms as markers of irAEs. Previous studies have found that early immune dysregulation, specifically with respect to cytokines associated with T-cell activation and autoimmune disease eventually lead to a discrete case of severe irAE[ 21 ]. Many of these early symptoms included in the model, such as diarrhea and constipation, fatigue, rash, headache, joint pain, etc. are symptoms that are commonly associated with immune dysregulation, thus it is possible that early immune dysregulation that ultimately leads to irAE development manifests through a combination of these symptoms. We found that an overall symptom burden(OSB) score—capturing both severity and duration of early symptoms—was strongly associated with the later development of irAE, with a significant signal emerging as early as week 6 of therapy. Specific symptoms contributing to this burden included gastrointestinal (diarrhea, constipation, vomiting), respiratory (shortness of breath), and symptoms commonly associated with immune dysregulation (headache, joint pain, rash), These symptoms preceded the clinical diagnosis of AI by an average of nine weeks and hypothyroidism by 8 weeks, indicating a potential window for proactive monitoring and early evaluation. The consistency of these findings in an external validation cohort, despite different treatment backbones (oral paclitaxel rather than intravenous), reinforces the potential generalizability of this approach. The results of this work can directly inform clinical care in several ways: 1) help clinicians risk-stratify follow-up and monitoring of patients at high-risk for toxicity; and 2) help develop and evaluate strategies to mitigate toxicities. Ultimately, this aids the overall goal of helping with clinical decision weighing patient toxicities against the benefit from immunotherapy. Through identifying those patients who are at highest risk, one can test de-escalation or supportive care strategies to decrease toxicity risk. Our models provide a strategy for early guidance and permit more personalized care by balancing increased efficacy with reduced risks of chronic and debilitating toxicities, as well as promoting early supportive care interventions that could mitigate the long -term damage that could result from these toxicities. Previous studies examining ways by which to predict immunotherapy-related irAEs have focused on pre-treatment gene markers, host genetics, and circulating protein biomarkers including serum autoantibodies[ 22 – 25 ]. The current study utilizes a novel framework, leveraging early treatment-emergent symptoms in patients across 6 different immunotherapies, to determine risk of later development of serious irAEs. Early symptoms that we found to be enriched in patients with these irAE, provide clinicians a straightforward primary assessment tool in clinic for identifying patients that may have higher risk of these irAEs. The OSB, a sum of AUCs of selected symptoms, provides an assessment timepoint at week 6, where cumulative symptom burden in the early treatment phase provides a personalized risk estimates of the likelihood of later developing these key irAEs. Such a framework could pro-actively identify at-risk patients before the development of a serious irAE, with relatively minimal clinic and provider burden. The high specificity of our model would enable effective screening of patients that are at low risk of later irAEs and distinguishing them from those who might require additional monitoring or diagnostic interventions. Genetic markers predictive of predisposition to autoimmunity and patient reported symptoms have the potential to improve the predictive accuracy of our OSB score, which we are exploring. Patient reported outcomes might be particularly suited to strengthen this model as they can capture even earlier symptoms than the current assessments timed around clinic visits. Since June 2021, routine electronic patient reported outcomes data collection was launched at all I-SPY2 sites[ 26 ], and we are working on methodologies for incorporating real-time patient reported AE data with our CTCAE-symptom based irAE prediction model to further improve predictive power. The clinical implications of our study, along with a growing evidence base regarding the adverse effects of immunotherapy, could enable more proactive management of patients to prevent these irAEs. The identification of symptoms (both early and those that persist), in combination with the identification of timepoints at which their emergence becomes most predictive, can make an important contribution to comprehensive survivorship care for women with breast cancer. Going forward, we will prospectively validate our methods in the ongoing I-SPY2.2 platform trial, as well as in other cancer trials and work to develop strategies to redirect therapy or develop preventive interventions when there are early indicators of treatment-induced long-term toxicities. Limitations We analyzed data prospectively collected during a clinical trial, and the overall health and co-morbidity status of trial participants may not be representative of all patients with stage II/III breast cancer. All our patients received weekly paclitaxel, but the type of chemotherapy partner administered along with ICIs can influence the type and frequency of early symptoms. However, three of the six treatment arms used in this analysis included additional experimental therapies along with ICI and paclitaxel which suggest generalizability. Since many irAEs are rare, we could only analyze associations between early symptoms and subsequent irAE for the two most common AEs, hypothyroidism and AI. If early symptoms predict development of other irAEs remains unknown. Although our models performed well in the validation cohort, the positive predictive value remained modest, reflecting the low absolute incidence of these irAEs, and underscoring that symptom burden is likely only one piece of the risk puzzle. Conclusions We show that a constellation of non-specific treatment emergent symptoms that develop as early as four to six weeks of ICI therapy could predict higher risk of future immune-related hypothyroidism and AI in patients with breast cancer. Since routine monitoring for AI with serial hormone level measurements is not currently recommended and the frequency of thyroid hormone assessments during therapy also vary, early onset fatigue, diarrhea, rash, shortness of breath could alert physicians to be more diligent in on-treatment and post-treatment monitoring for these endocrinopathies. Work is ongoing to develop strategies for monitoring and potential treatment mitigation to reduce the risk of permanent endocrinopathies. Declarations Funding This resear ch was supported by the National Cancer Institute of the National Institutes of Health under award number P01CA210961 and R21CA2588218. The authors wish to acknowledge the generous support of the study sponsors, Quantum Leap Healthcare Collaborative (QLHC, 2013 to present) and the Foundation for the National Institutes of Health (2010 to 2012). The authors sincerely appreciate the ongoing support for the I-SPY2 Trial from the Safeway Foundation, the William K. Bowes, Jr. Foundation, Give Breast Cancer the Boot, QLHC and the Breast Cancer Research Foundation. DISCLOSURES CY reports institutional research grant from NCI/NIH; salary support and travel reimbursement from Quantum Leap Healthcare Collaborative; US patent titled, “Breast cancer response prediction subtypes,” (No. 18/174,491); and University of California Inventor Share. RN reports research funding from Arvinas, AstraZeneca, BMS, Corcept Therapeutics, Genentech/Roche, Gilead, GSK, Merck, Novartis, OBI Pharma, OncoSec, Pfizer, Relay, Seattle Genetics, Sun Pharma, Taiho; advisory roles with AstraZeneca, BeyondSpring, Daiichi Sankyo, Exact Sciences, Fujifilm, GE, Gilead, Guardant Health, Infinity, iTeos, Merck, Moderna, Novartis, OBI, Oncosec, Pfizer, Sanofi, Seagen, Stemline. LP reports institutional research funding from Susan Komen Foundation, Breast Cancer Research Foundation, National Cancer Institute, Pfizer, AstraZeneca, Menarini/Stemline, Bristol Myers Squibb, Merck and Co,; consulting fees from AstraZeneca, Merck, Novartis, Genentech, Natera, Personalis, ExactSciences, Stemline/Menarini; patent titled, “Method of measuring residual cancer and predicting patient survival,” (No. 7711494); and DSMB member of the DYNASTY Breast02, OPTIMA and PARTNER Trials. AJC reports institutional research funding from Merck, Amgen, Puma, Seagen, Pfizer, Olema; advisory roles with Astra Zeneca and Genentech. MCL is an employee of Natera, Inc., with stock or options to own stock, and further reports grants/contracts (funding to institution: Mayo Clinic) from Eisai, Exact Sciences Corporation, Genentech, Genomic Health, GRAIL, Menarini Silicon Biosystems, Merck, Novartis, Seattle Genetics and Tesaro; travel support reimbursement from AstraZeneca, Genomic Health and Ionis; and ad hoc advisory board meetings (all funds to Mayo Clinic; no personal compensation) from AstraZeneca, Celgene, Roche/Genentech, Genomic Health, GRAIL, Ionis, Merck, Pfizer, Seattle Genetics and Syndax. HH reports participating on advisory board for Pfizer. HS reports consulting fees from Eli Lilly, Novartis, AstraZeneca, PUMA and Sermonix. CI reports institutional research funding from Tesaro/GSK, Seattle Genetics, Pfizer, AstraZeneca, BMS, Genentech, Novartis and Regeneron; consultancy roles with AstraZeneca, Genentech, Gilead, ION, Merck, Medscape, MJH Holdings, Novartis, Pfizer, PUMA, Seagen; royalties from Wolters Kluwer (UptoDate); McGraw Hill (Goodman and Gillman). RAS reports institutional research funding from OBI Pharmaceuticals, Quantum Leap Healthcare Collaborative, AstraZeneca and Gilead; serves on AstraZeneca and Stemline Advisory Boards; Gilead Speaker’s Bureau; consultancy role with Quantum Leap Healthcare Collaborative. ESR reports grants from V Foundation, NIH, Susan G. Komen; institutional research funding from GSK, Seagen, Pfizer, Lilly; consulting and honoraria from Novartis, Merck, Seagen, Astrazeneca, Lilly; Cancer Awareness Network Board member and support from ASCO, NCCN. AA is an employee of Quantum Leap Healthcare Collaborative. ADM reports institutional research funding from Novartis, Pfizer, Genentech and Neogenomics; Program Chair, Scientific Advisory Committee, ASCO. LJvV is an advisor for Exai BIo; part-time employee and owns stock in Agendia. JP reports honoraria from Methods in Clinical Research - Faculty SCION Workshop; support from ASCO and Advocate Scholarship; AACR - SSP Program; VIVLI, U Wisc SPORE - EAB, QuantumLEAD - COVID DSMB, PCORI - Reviewer and ISPY Advocate Lead. DY reports research funding from NIH/NCI P30 CA 077598, P01 CA234228-01 and R01CA251600; consulting fees from Martell Diagnostics; and honoraria and travel for speaking at the "International Breast Cancer Conference." NMH reports institutional research finding from NIH. LJE reports past funding from Merck & Co. and Moderna for an unrelated trial; participation on an advisory board for Blue Cross Blue Shield; personal fees from UpToDate; unpaid board member of QLHC. HSR reports institutional research support from AstraZeneca, Daiichi Sankyo, Inc., F. Hoffmann-La Roche AG/Genentech, Inc., Gilead Sciences, Inc., Lilly; Merck & Co., Novartis Pharmaceuticals Corporation, Pfizer, Stemline Therapeutics, OBI Pharma, Ambrx, Greenwich Pharma; advisory and consulting roles with Chugai, Puma, Sanofi, Napo, and Mylan. ZQ has research support from NIH NIDDK DiabDocs K12DK133995 and a Larry L Hillblom Foundation Start Up Grant consulting role with Sanofi. All other authors declare no competing interests. Data Availability Statement Subject-level data for this study is available to approved investigators completing a request form available at: https://www.quantumleaphealth.org/for-investigators/clinicians-proposal-submissions/. Author Contribution A.B: Conceptualization; Formal analysis; Funding acquisition; Investigation; Methodology; Project administration; Resources; Software; Supervision; Validation; Writing – original draft; Writing – review & editingS.U: Data curation; Formal analysis; Investigation; Software; Validation; Visualization; Writing – original draft; Writing – review & editingM.M: Data curation; Investigation; Writing – review & editingR.R: Formal analysis; Software; VisualizationC.Y, A.J.C, M.L, H.H, H.S, E.S.R, P.R, K.Y, K.K, A.T: Data curation; InvestigationL.P: Data curation; Investigation; Writing – review & editingC.I, R.S: Data curation; Investigation; Project administrationE.P, D.H: Writing – review & editingM.Mu: Formal AnalysisG.H: Data CurationA.A: Data Curation, SoftwareA.D, L.vV, D.Y, N.H, J.P: Project AdministrationA.O: SupervisionZ.Q, R.N.C: Data curation; Writing – review & editingL.E: Funding acquisition; Project administration; Resources; Supervision; Writing – original draft; Writing – review & editingH.R, R.N: Investigation; Project administration; Supervision; Writing – review& editing References Esfahani, K., et al., A review of cancer immunotherapy: from the past, to the present, to the future. Curr Oncol, 2020. 27(Suppl 2): p. S87-s97. Nathan, M.R. and P. Schmid, The emerging world of breast cancer immunotherapy . Breast, 2018. 37: p. 200–206. Cortes, J., et al., Pembrolizumab plus chemotherapy versus placebo plus chemotherapy for previously untreated locally recurrent inoperable or metastatic triple-negative breast cancer (KEYNOTE-355): a randomised, placebo-controlled, double-blind, phase 3 clinical trial . Lancet, 2020. 396(10265): p. 1817–1828. Rizzo, A., et al., KEYNOTE-522, IMpassion031 and GeparNUEVO: changing the paradigm of neoadjuvant immune checkpoint inhibitors in early triple-negative breast cancer . Future Oncol, 2022. 18(18): p. 2301–2309. Schmid, P., et al., Pembrolizumab for Early Triple-Negative Breast Cancer . N Engl J Med, 2020. 382(9): p. 810–821. Henriques, B., F. Mendes, and D. Martins, Immunotherapy in Breast Cancer: When, How, and What Challenges? Biomedicines, 2021. 9(11). Setordzi, P., et al., The recent advances of PD-1 and PD-L1 checkpoint signaling inhibition for breast cancer immunotherapy . Eur J Pharmacol, 2021. 895: p. 173867. Pusztai, L., et al., Durvalumab with olaparib and paclitaxel for high-risk HER2-negative stage II/III breast cancer: Results from the adaptively randomized I-SPY2 trial . Cancer Cell, 2021. 39(7): p. 989–998 e5. Cardoso F, M.H., Schmid P, et al, LBA21 KEYNOTE-756: Phase III study of neoadjuvant pembrolizumab (pembro) or placebo (pbo) + chemotherapy (chemo), followed by adjuvant pembro or pbo + endocrine therapy (ET) for early-stage high-risk ER+/HER2– breast cancer . Annals of Oncology. Loi S, C.G., Salgado RF, et al, LBA20 A randomized, double-blind trial of nivolumab (NIVO) vs placebo (PBO) with neoadjuvant chemotherapy (NACT) followed by adjuvant endocrine therapy (ET) ± NIVO in patients (pts) with high-risk, ER+ HER2 – primary breast cancer (BC) . Annals of Oncology. Criscitiello, C., et al., Managing side effects of immune checkpoint inhibitors in breast cancer . Crit Rev Oncol Hematol, 2021. 162: p. 103354. Li, Y., et al., Efficacy and Safety of Adding Immune Checkpoint Inhibitors to Neoadjuvant Chemotherapy Against Triple-Negative Breast Cancer: A Meta-Analysis of Randomized Controlled Trials . Front Oncol, 2021. 11: p. 657634. Sternschuss, M., et al., Efficacy and safety of neoadjuvant immune checkpoint inhibitors in early-stage triple-negative breast cancer: a systematic review and meta-analysis . J Cancer Res Clin Oncol, 2021. 147(11): p. 3369–3379. Xin, Y., et al., Immune checkpoint inhibitors plus neoadjuvant chemotherapy in early triple-negative breast cancer: a systematic review and meta-analysis . BMC Cancer, 2021. 21(1): p. 1261. Balibegloo, M., et al., Adverse events associated with immune checkpoint inhibitors in patients with breast cancer: A systematic review and meta-analysis . Int Immunopharmacol, 2021. 96: p. 107796. Nanda, R., et al., Effect of Pembrolizumab Plus Neoadjuvant Chemotherapy on Pathologic Complete Response in Women With Early-Stage Breast Cancer: An Analysis of the Ongoing Phase 2 Adaptively Randomized I-SPY2 Trial . JAMA Oncol, 2020. 6(5): p. 676–684. Cui, K., et al., Immune checkpoint inhibitors and adrenal insufficiency: a large-sample case series study . Ann Transl Med, 2022. 10(5): p. 251. Grouthier, V., et al., Immune Checkpoint Inhibitor-Associated Primary Adrenal Insufficiency: WHO VigiBase Report Analysis . Oncologist, 2020. 25(8): p. 696–701. D'Abreo, N. and S. Adams, Immune-checkpoint inhibition for metastatic triple-negative breast cancer: safety first? Nat Rev Clin Oncol, 2019. 16(7): p. 399–400. Wang, H. and D. Yee, I-SPY 2: a Neoadjuvant Adaptive Clinical Trial Designed to Improve Outcomes in High-Risk Breast Cancer . Curr Breast Cancer Rep, 2019. 11(4): p. 303–310. Khan, S., et al., Immune dysregulation in cancer patients developing immune-related adverse events . Br J Cancer, 2019. 120(1): p. 63–68. Chennamadhavuni, A., et al., Risk Factors and Biomarkers for Immune-Related Adverse Events: A Practical Guide to Identifying High-Risk Patients and Rechallenging Immune Checkpoint Inhibitors . Front Immunol, 2022. 13: p. 779691. Chin, I.S., et al., Germline genetic variation and predicting immune checkpoint inhibitor induced toxicity . NPJ Genom Med, 2022. 7(1): p. 73. Jing, Y., et al., Multi-omics prediction of immune-related adverse events during checkpoint immunotherapy . Nat Commun, 2020. 11(1): p. 4946. Les, I., et al., Predictive Biomarkers for Checkpoint Inhibitor Immune-Related Adverse Events . Cancers (Basel), 2023. 15(5). Northrop, A., et al., Implementation and impact of an electronic patient reported outcomes system in a phase II multi-site adaptive platform clinical trial for early-stage breast cancer . J Am Med Inform Assoc, 2025. 32(1): p. 172–180. Additional Declarations Competing interest reported. CY reports institutional research grant from NCI/NIH; salary support and travel reimbursement from Quantum Leap Healthcare Collaborative; US patent titled, “Breast cancer response prediction subtypes,” (No. 18/174,491); and University of California Inventor Share. RN reports research funding from Arvinas, AstraZeneca, BMS, Corcept Therapeutics, Genentech/Roche, Gilead, GSK, Merck, Novartis, OBI Pharma, OncoSec, Pfizer, Relay, Seattle Genetics, Sun Pharma, Taiho; advisory roles with AstraZeneca, BeyondSpring, Daiichi Sankyo, Exact Sciences, Fujifilm, GE, Gilead, Guardant Health, Infinity, iTeos, Merck, Moderna, Novartis, OBI, Oncosec, Pfizer, Sanofi, Seagen, Stemline. LP reports institutional research funding from Susan Komen Foundation, Breast Cancer Research Foundation, National Cancer Institute, Pfizer, AstraZeneca, Menarini/Stemline, Bristol Myers Squibb, Merck and Co,; consulting fees from AstraZeneca, Merck, Novartis, Genentech, Natera, Personalis, ExactSciences, Stemline/Menarini; patent titled, “Method of measuring residual cancer and predicting patient survival,” (No. 7711494); and DSMB member of the DYNASTY Breast02, OPTIMA and PARTNER Trials. AJC reports institutional research funding from Merck, Amgen, Puma, Seagen, Pfizer, Olema; advisory roles with Astra Zeneca and Genentech. MCL is an employee of Natera, Inc., with stock or options to own stock, and further reports grants/contracts (funding to institution: Mayo Clinic) from Eisai, Exact Sciences Corporation, Genentech, Genomic Health, GRAIL, Menarini Silicon Biosystems, Merck, Novartis, Seattle Genetics and Tesaro; travel support reimbursement from AstraZeneca, Genomic Health and Ionis; and ad hoc advisory board meetings (all funds to Mayo Clinic; no personal compensation) from AstraZeneca, Celgene, Roche/Genentech, Genomic Health, GRAIL, Ionis, Merck, Pfizer, Seattle Genetics and Syndax. HH reports participating on advisory board for Pfizer. HS reports consulting fees from Eli Lilly, Novartis, AstraZeneca, PUMA and Sermonix. CI reports institutional research funding from Tesaro/GSK, Seattle Genetics, Pfizer, AstraZeneca, BMS, Genentech, Novartis and Regeneron; consultancy roles with AstraZeneca, Genentech, Gilead, ION, Merck, Medscape, MJH Holdings, Novartis, Pfizer, PUMA, Seagen; royalties from Wolters Kluwer (UptoDate); McGraw Hill (Goodman and Gillman). RAS reports institutional research funding from OBI Pharmaceuticals, Quantum Leap Healthcare Collaborative, AstraZeneca and Gilead; serves on AstraZeneca and Stemline Advisory Boards; Gilead Speaker’s Bureau; consultancy role with Quantum Leap Healthcare Collaborative. ESR reports grants from V Foundation, NIH, Susan G. Komen; institutional research funding from GSK, Seagen, Pfizer, Lilly; consulting and honoraria from Novartis, Merck, Seagen, Astrazeneca, Lilly; Cancer Awareness Network Board member and support from ASCO, NCCN. AA is an employee of Quantum Leap Healthcare Collaborative. ADM reports institutional research funding from Novartis, Pfizer, Genentech and Neogenomics; Program Chair, Scientific Advisory Committee, ASCO. LJvV is an advisor for Exai BIo; part-time employee and owns stock in Agendia. JP reports honoraria from Methods in Clinical Research - Faculty SCION Workshop; support from ASCO and Advocate Scholarship; AACR - SSP Program; VIVLI, U Wisc SPORE - EAB, QuantumLEAD - COVID DSMB, PCORI - Reviewer and ISPY Advocate Lead. DY reports research funding from NIH/NCI P30 CA 077598, P01 CA234228-01 and R01CA251600; consulting fees from Martell Diagnostics; and honoraria and travel for speaking at the "International Breast Cancer Conference." NMH reports institutional research finding from NIH. LJE reports past funding from Merck & Co. and Moderna for an unrelated trial; participation on an advisory board for Blue Cross Blue Shield; personal fees from UpToDate; unpaid board member of QLHC. HSR reports institutional research support from AstraZeneca, Daiichi Sankyo, Inc., F. Hoffmann-La Roche AG/Genentech, Inc., Gilead Sciences, Inc., Lilly; Merck & Co., Novartis Pharmaceuticals Corporation, Pfizer, Stemline Therapeutics, OBI Pharma, Ambrx, Greenwich Pharma; advisory and consulting roles with Chugai, Puma, Sanofi, Napo, and Mylan. ZQ has research support from NIH NIDDK DiabDocs K12DK133995 and a Larry L Hillblom Foundation Start Up Grant consulting role with Sanofi. All other authors declare no competing interests. Supplementary Files SupplementaryTable1Armdetailsofpatients.xlsx Supplementary Table 1: Arm details of patients included in study. SupplementaryTable2CTCAEitemsirae.xlsx Supplementary Table 2: CTCAE Items selection criteria for analysis SupplementaryTable3SymptomfeaturesforOSB.xlsx Supplementary Table 3: Symptom features for computation of OSB. SupplementaryTable4enrichment.xlsx Supplementary Table 4: Binary (presence absence) enrichment analysis for all symptoms. SupplementaryTable5LogisticRegressionofNumberofSymptomswithirAE.xlsx Supplementary Table 5: Logistic Regression of Number of Symptoms with irAE upSupplementaryTable6AdrenalThresholds.xlsx Supplementary Table 6: ROC curve thresholds with sensitivity and specificity for adrenal insufficiency at week 6. upSupplementaryTable7HypoThresholds.xlsx Supplementary Table 7: ROC curve thresholds with sensitivity and specificity for hypothyroidism at week 6. SupplementaryTable8Demographiccharacteristicsofvalidationset.xlsx Supplementary Table 8: Demographic characteristics of validation set patient population. SUPPLEMENTARYFIGURES.docx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8689063","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":587936370,"identity":"560a8c9d-92ac-42c7-a228-fd5511c48a2c","order_by":0,"name":"Amrita Basu","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA5ElEQVRIie3QIQvCQBTA8SeDrdxYfRb3FbZyGPwwrswyLAciKONAMGozCX4FLeaTwdLEOriifWHJKN7UYDoXDfeHd+n94O4ATKY/TagZqulc1eGpCX4T8SZWs9rlbQh8iI2tiLc537J6DmNvXdAZmacIzuKIOoJyFAiRA8MyoZLkGQLJJ1oCMlZvsSHiJaHS5SIFTKiW+C/ygGh3KShzubqYX+lJ0JDTEqK9SKjlcgsBiZ6EDTmvkIVlzLpb9RabxKyvIz0ZW/X0Phj3LtmhrtSPeU52KHXk09dN7BbrJpPJZPrRE/s5SO7t4JgtAAAAAElFTkSuQmCC","orcid":"","institution":"University of California, San Francisco","correspondingAuthor":true,"prefix":"","firstName":"Amrita","middleName":"","lastName":"Basu","suffix":""},{"id":587936372,"identity":"2dbd1939-8768-48b6-80a4-9f31a6cfcdfd","order_by":1,"name":"Saumya Umashankar","email":"","orcid":"","institution":"University of California, San Francisco","correspondingAuthor":false,"prefix":"","firstName":"Saumya","middleName":"","lastName":"Umashankar","suffix":""},{"id":587936373,"identity":"c54137ed-7c64-4d95-95d8-95af31d01385","order_by":2,"name":"Michelle Melisko","email":"","orcid":"","institution":"University of California, San Francisco","correspondingAuthor":false,"prefix":"","firstName":"Michelle","middleName":"","lastName":"Melisko","suffix":""},{"id":587936374,"identity":"1e1d0e51-0426-4d68-aa61-13952b27b3fa","order_by":3,"name":"Ritu Roy","email":"","orcid":"","institution":"University of California, San Francisco","correspondingAuthor":false,"prefix":"","firstName":"Ritu","middleName":"","lastName":"Roy","suffix":""},{"id":587936375,"identity":"d26ffc01-0b74-42ce-a896-e5fc3d996690","order_by":4,"name":"Christina Yau","email":"","orcid":"","institution":"University of California, San Francisco","correspondingAuthor":false,"prefix":"","firstName":"Christina","middleName":"","lastName":"Yau","suffix":""},{"id":587936376,"identity":"0c1ffb66-f190-4094-b3e8-ed863a505d02","order_by":5,"name":"Lajos Pusztai","email":"","orcid":"","institution":"Yale University","correspondingAuthor":false,"prefix":"","firstName":"Lajos","middleName":"","lastName":"Pusztai","suffix":""},{"id":587936377,"identity":"c7eac9ff-009a-4a8e-8be4-7215ae2c465c","order_by":6,"name":"A. Jo Chien","email":"","orcid":"","institution":"University of California, San Francisco","correspondingAuthor":false,"prefix":"","firstName":"A.","middleName":"Jo","lastName":"Chien","suffix":""},{"id":587936378,"identity":"90af3b3f-553a-4c74-be4d-fe1f12d15d2e","order_by":7,"name":"Minetta Liu","email":"","orcid":"","institution":"Mayo Clinic","correspondingAuthor":false,"prefix":"","firstName":"Minetta","middleName":"","lastName":"Liu","suffix":""},{"id":587936379,"identity":"c9206dab-e970-49a1-b0f1-32ee53e540e7","order_by":8,"name":"Hyo Han","email":"","orcid":"","institution":"Moffitt Cancer Center","correspondingAuthor":false,"prefix":"","firstName":"Hyo","middleName":"","lastName":"Han","suffix":""},{"id":587936380,"identity":"850fa2c4-8f21-4096-8788-b6e037dcbd86","order_by":9,"name":"Hatem Soliman","email":"","orcid":"","institution":"Moffitt Cancer Center","correspondingAuthor":false,"prefix":"","firstName":"Hatem","middleName":"","lastName":"Soliman","suffix":""},{"id":587936381,"identity":"708cf7af-be83-460e-b5ac-9fd87641c8f2","order_by":10,"name":"Claudine Isaacs","email":"","orcid":"","institution":"Georgetown University","correspondingAuthor":false,"prefix":"","firstName":"Claudine","middleName":"","lastName":"Isaacs","suffix":""},{"id":587936383,"identity":"e3feef54-77de-4d5f-8a8a-eb7d011dfd87","order_by":11,"name":"Rebecca Shatsky","email":"","orcid":"","institution":"University of California, San Diego","correspondingAuthor":false,"prefix":"","firstName":"Rebecca","middleName":"","lastName":"Shatsky","suffix":""},{"id":587936385,"identity":"1d1d679d-bb62-4719-af56-97b9e61db50c","order_by":12,"name":"Erica Stringer-Reasor","email":"","orcid":"","institution":"University of Alabama at Birmingham","correspondingAuthor":false,"prefix":"","firstName":"Erica","middleName":"","lastName":"Stringer-Reasor","suffix":""},{"id":587936387,"identity":"57ceeb30-a4ec-4faa-8d7c-c50244eeb90a","order_by":13,"name":"Patricia Robinson","email":"","orcid":"","institution":"Loyola University Chicago","correspondingAuthor":false,"prefix":"","firstName":"Patricia","middleName":"","lastName":"Robinson","suffix":""},{"id":587936388,"identity":"99469360-11fe-45b8-a436-47fd63ed9a3e","order_by":14,"name":"Kay Yeung","email":"","orcid":"","institution":"University of California, San Diego","correspondingAuthor":false,"prefix":"","firstName":"Kay","middleName":"","lastName":"Yeung","suffix":""},{"id":587936389,"identity":"4f2d3844-4237-4985-8c3e-dfe71bf4d873","order_by":15,"name":"Kevin Kalinsky","email":"","orcid":"","institution":"Emory University","correspondingAuthor":false,"prefix":"","firstName":"Kevin","middleName":"","lastName":"Kalinsky","suffix":""},{"id":587936390,"identity":"5ca927a9-a2c1-4bf3-9914-47dbfe8504e0","order_by":16,"name":"Alexandra Thomas","email":"","orcid":"","institution":"Wake Forest University","correspondingAuthor":false,"prefix":"","firstName":"Alexandra","middleName":"","lastName":"Thomas","suffix":""},{"id":587936391,"identity":"c0cbd569-b4a9-4c21-b253-382b09215d80","order_by":17,"name":"Errol Philip","email":"","orcid":"","institution":"University of California, San Francisco","correspondingAuthor":false,"prefix":"","firstName":"Errol","middleName":"","lastName":"Philip","suffix":""},{"id":587936392,"identity":"6598d5c5-2b7c-4f6a-a803-6ab06e7e0c7b","order_by":18,"name":"Mina Musthafa","email":"","orcid":"","institution":"University of California, San Francisco","correspondingAuthor":false,"prefix":"","firstName":"Mina","middleName":"","lastName":"Musthafa","suffix":""},{"id":587936394,"identity":"f1b42d66-3fcd-460c-8d7d-917d56500be4","order_by":19,"name":"Gillian Hirst","email":"","orcid":"","institution":"University of California, San Francisco","correspondingAuthor":false,"prefix":"","firstName":"Gillian","middleName":"","lastName":"Hirst","suffix":""},{"id":587936396,"identity":"4fbec8c7-0332-4ecb-87d1-b9364c1572e4","order_by":20,"name":"Adam Asare","email":"","orcid":"","institution":"University of California, San Francisco","correspondingAuthor":false,"prefix":"","firstName":"Adam","middleName":"","lastName":"Asare","suffix":""},{"id":587936398,"identity":"4838cdd4-8764-461a-a644-49c453285659","order_by":21,"name":"Angela DeMichele","email":"","orcid":"","institution":"University of Pennsylvania","correspondingAuthor":false,"prefix":"","firstName":"Angela","middleName":"","lastName":"DeMichele","suffix":""},{"id":587936401,"identity":"3c2a5cf2-0d56-4faa-8594-ab8dddb3dec7","order_by":22,"name":"Laura van't Veer","email":"","orcid":"","institution":"University of California, San Francisco","correspondingAuthor":false,"prefix":"","firstName":"Laura","middleName":"van't","lastName":"Veer","suffix":""},{"id":587936404,"identity":"8a80ccbf-4f35-4ddf-a485-6140a5280548","order_by":23,"name":"Douglas Yee","email":"","orcid":"","institution":"University of Minnesota","correspondingAuthor":false,"prefix":"","firstName":"Douglas","middleName":"","lastName":"Yee","suffix":""},{"id":587936406,"identity":"afbbc614-8252-47ac-9bfb-507c8f7009b3","order_by":24,"name":"Nola Hylton","email":"","orcid":"","institution":"University of California, San Francisco","correspondingAuthor":false,"prefix":"","firstName":"Nola","middleName":"","lastName":"Hylton","suffix":""},{"id":587936407,"identity":"dfa3f0cc-f9b7-492a-8a99-7f3074829c0f","order_by":25,"name":"Adam Olshen","email":"","orcid":"","institution":"University of California, San Francisco","correspondingAuthor":false,"prefix":"","firstName":"Adam","middleName":"","lastName":"Olshen","suffix":""},{"id":587936409,"identity":"c6d9d146-bb58-42f3-bd22-876ce722ab35","order_by":26,"name":"Jane Perlmutter","email":"","orcid":"","institution":"Gemini Group","correspondingAuthor":false,"prefix":"","firstName":"Jane","middleName":"","lastName":"Perlmutter","suffix":""},{"id":587936410,"identity":"e1d830e8-dbe2-4c3c-9f76-cf69ecc8f5b8","order_by":27,"name":"Ronald N. Cohen","email":"","orcid":"","institution":"University of Chicago","correspondingAuthor":false,"prefix":"","firstName":"Ronald","middleName":"N.","lastName":"Cohen","suffix":""},{"id":587936411,"identity":"41c4d382-f073-4b0e-af01-069dc0f90424","order_by":28,"name":"Zoe Quandt","email":"","orcid":"","institution":"University of California, San Francisco","correspondingAuthor":false,"prefix":"","firstName":"Zoe","middleName":"","lastName":"Quandt","suffix":""},{"id":587936412,"identity":"97b0197b-c23c-4e96-9f21-e77edb7c4564","order_by":29,"name":"Laura Esserman","email":"","orcid":"","institution":"University of California, San Francisco","correspondingAuthor":false,"prefix":"","firstName":"Laura","middleName":"","lastName":"Esserman","suffix":""},{"id":587936413,"identity":"0e903486-a1e6-47b6-b642-b1cc3d50dbc9","order_by":30,"name":"Dawn Hershman","email":"","orcid":"","institution":"Columbia University","correspondingAuthor":false,"prefix":"","firstName":"Dawn","middleName":"","lastName":"Hershman","suffix":""},{"id":587936414,"identity":"dfe0565a-2cfa-4141-8af9-77fcda7c1237","order_by":31,"name":"Hope Rugo","email":"","orcid":"","institution":"University of California, San Francisco","correspondingAuthor":false,"prefix":"","firstName":"Hope","middleName":"","lastName":"Rugo","suffix":""},{"id":587936415,"identity":"ebd86805-fb27-430c-9075-687e773ef159","order_by":32,"name":"Rita Nanda","email":"","orcid":"","institution":"University of Chicago","correspondingAuthor":false,"prefix":"","firstName":"Rita","middleName":"","lastName":"Nanda","suffix":""}],"badges":[],"createdAt":"2026-01-24 20:53:12","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8689063/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8689063/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":102376995,"identity":"05134658-3ede-44f5-a650-d5352fc35ff8","added_by":"auto","created_at":"2026-02-11 05:40:02","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":256757,"visible":true,"origin":"","legend":"\u003cp\u003eSwimmers plots showing the incidence, duration and co-occurrence of the symptom drivers symptoms in patients that had adrenal insufficiency vs. without (A), and in patients that had hypothyroidism vs. without (B). Each item on y axis represents a patient, with gray bars representing period of active treatment. Black points represent irAE start date\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-8689063/v1/ef290ba018b5c0a3cf197056.png"},{"id":102377014,"identity":"75910716-7e60-4ace-b536-6bffe1b62200","added_by":"auto","created_at":"2026-02-11 05:40:16","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":50427,"visible":true,"origin":"","legend":"\u003cp\u003e(A-B) Violin plots showing the difference in the overall symptom burden (OSB) between cases (irAE) and controls (no irAE) for adrenal insufficiency (A) and hypothyroidism (B). The p-value (PV) and Bonferroni adjusted p-value (bonf) for each model is provided above the plot.\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-8689063/v1/b6b240324700767fd0a2f562.png"},{"id":102376992,"identity":"a573d9ca-b2ab-42a1-a5ad-3a51bb587193","added_by":"auto","created_at":"2026-02-11 05:40:00","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":257361,"visible":true,"origin":"","legend":"\u003cp\u003e(A-B) Incidence of symptom presence over time. Symptoms with p value \u0026lt;0.2 are visualized in patients with adrenal insufficiency vs those without (A) and in patients with hypothyroidism vs those without (B). (C-D) Symptom co-occurrence heatmap for symptoms present prior to week 6 for patients with adrenal insufficiency (C) and hypothyroidism (D).\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-8689063/v1/4c27b54d2370d25728603609.png"},{"id":108977291,"identity":"9d8295ce-501c-4ab8-bd22-8e98fae1c614","added_by":"auto","created_at":"2026-05-11 11:31:14","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":979857,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8689063/v1/f58ef67e-e548-4bc1-9df3-ee75d9e35277.pdf"},{"id":102376996,"identity":"baa1a8c8-7db4-4a7b-82fe-a1d36a3e2757","added_by":"auto","created_at":"2026-02-11 05:40:02","extension":"xlsx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":9459,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary Table 1: Arm details of patients included in study.\u003c/p\u003e","description":"","filename":"SupplementaryTable1Armdetailsofpatients.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8689063/v1/89d16d9519e1fb34f83edb28.xlsx"},{"id":102377008,"identity":"d2828846-1f57-454b-a3da-ea5d8a84e976","added_by":"auto","created_at":"2026-02-11 05:40:05","extension":"xlsx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":10622,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary Table 2: CTCAE Items selection criteria for analysis\u003c/p\u003e","description":"","filename":"SupplementaryTable2CTCAEitemsirae.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8689063/v1/7c9d5ee1a9734031fc029654.xlsx"},{"id":102377010,"identity":"d51019c9-a730-488d-a089-8571c25b34c4","added_by":"auto","created_at":"2026-02-11 05:40:12","extension":"xlsx","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":10048,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary Table 3: Symptom features for computation of OSB.\u003c/p\u003e","description":"","filename":"SupplementaryTable3SymptomfeaturesforOSB.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8689063/v1/5330d8ccc7d4a881b4bdc392.xlsx"},{"id":102377011,"identity":"25283d1a-f3af-494d-a332-27634b1e95af","added_by":"auto","created_at":"2026-02-11 05:40:12","extension":"xlsx","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":15236,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary Table 4: Binary (presence absence) enrichment analysis for all symptoms.\u003c/p\u003e","description":"","filename":"SupplementaryTable4enrichment.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8689063/v1/5273abb9c007f0a79cdec5c3.xlsx"},{"id":102377009,"identity":"1f6302e1-22cd-45db-b14f-f4ae4adef7be","added_by":"auto","created_at":"2026-02-11 05:40:06","extension":"xlsx","order_by":5,"title":"","display":"","copyAsset":false,"role":"supplement","size":9919,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary Table 5: Logistic Regression of Number of Symptoms with irAE\u003c/p\u003e","description":"","filename":"SupplementaryTable5LogisticRegressionofNumberofSymptomswithirAE.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8689063/v1/eb0348fe4afeee761a1ed23c.xlsx"},{"id":102376988,"identity":"2abe37d4-5479-4ba6-b221-d3ef57077ce6","added_by":"auto","created_at":"2026-02-11 05:39:59","extension":"xlsx","order_by":6,"title":"","display":"","copyAsset":false,"role":"supplement","size":12104,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary Table 6: ROC curve thresholds with sensitivity and specificity for adrenal insufficiency at week 6.\u003c/p\u003e","description":"","filename":"upSupplementaryTable6AdrenalThresholds.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8689063/v1/384c7212b803543cb40ada82.xlsx"},{"id":102398480,"identity":"7a78e78b-c9b7-4430-b4e6-b86f3fba7cd6","added_by":"auto","created_at":"2026-02-11 10:22:59","extension":"xlsx","order_by":7,"title":"","display":"","copyAsset":false,"role":"supplement","size":11618,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary Table 7: ROC curve thresholds with sensitivity and specificity for hypothyroidism at week 6.\u003c/p\u003e","description":"","filename":"upSupplementaryTable7HypoThresholds.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8689063/v1/67bb106750d665ff851e3892.xlsx"},{"id":102377015,"identity":"f60a3de2-4ed3-4d8f-b633-97f501dcf031","added_by":"auto","created_at":"2026-02-11 05:40:17","extension":"xlsx","order_by":8,"title":"","display":"","copyAsset":false,"role":"supplement","size":9599,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary Table 8: Demographic characteristics of validation set patient population.\u003c/p\u003e","description":"","filename":"SupplementaryTable8Demographiccharacteristicsofvalidationset.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8689063/v1/30171c8497ff9d2b2a9cfe17.xlsx"},{"id":102376994,"identity":"7c8ad2a5-2508-4de1-ab4d-211790bd83cd","added_by":"auto","created_at":"2026-02-11 05:40:01","extension":"docx","order_by":9,"title":"","display":"","copyAsset":false,"role":"supplement","size":637585,"visible":true,"origin":"","legend":"","description":"","filename":"SUPPLEMENTARYFIGURES.docx","url":"https://assets-eu.researchsquare.com/files/rs-8689063/v1/c3ad9cae5fa54ba8a18572cb.docx"}],"financialInterests":"Competing interest reported. CY reports institutional research grant from NCI/NIH; salary support and travel reimbursement from Quantum Leap Healthcare Collaborative; US patent titled, “Breast cancer response prediction subtypes,” (No. 18/174,491); and University of California Inventor Share. RN reports research funding from Arvinas, AstraZeneca, BMS, Corcept Therapeutics, Genentech/Roche, Gilead, GSK, Merck, Novartis, OBI Pharma, OncoSec, Pfizer, Relay, Seattle Genetics, Sun Pharma, Taiho; advisory roles with AstraZeneca, BeyondSpring, Daiichi Sankyo, Exact Sciences, Fujifilm, GE, Gilead, Guardant Health, Infinity, iTeos, Merck, Moderna, Novartis, OBI, Oncosec, Pfizer, Sanofi, Seagen, Stemline. LP reports institutional research funding from Susan Komen Foundation, Breast Cancer Research Foundation, National Cancer Institute, Pfizer, AstraZeneca, Menarini/Stemline, Bristol Myers Squibb, Merck and Co,; consulting fees from AstraZeneca, Merck, Novartis, Genentech, Natera, Personalis, ExactSciences, Stemline/Menarini; patent titled, “Method of measuring residual cancer and predicting patient survival,” (No. 7711494); and DSMB member of the DYNASTY Breast02, OPTIMA and PARTNER Trials. AJC reports institutional research funding from Merck, Amgen, Puma, Seagen, Pfizer, Olema; advisory roles with Astra Zeneca and Genentech. MCL is an employee of Natera, Inc., with stock or options to own stock, and further reports grants/contracts (funding to institution: Mayo Clinic) from Eisai, Exact Sciences Corporation, Genentech, Genomic Health, GRAIL, Menarini Silicon Biosystems, Merck, Novartis, Seattle Genetics and Tesaro; travel support reimbursement from AstraZeneca, Genomic Health and Ionis; and ad hoc advisory board meetings (all funds to Mayo Clinic; no personal compensation) from AstraZeneca, Celgene, Roche/Genentech, Genomic Health, GRAIL, Ionis, Merck, Pfizer, Seattle Genetics and Syndax. HH reports participating on advisory board for Pfizer. HS reports consulting fees from Eli Lilly, Novartis, AstraZeneca, PUMA and Sermonix. CI reports institutional research funding from Tesaro/GSK, Seattle Genetics, Pfizer, AstraZeneca, BMS, Genentech, Novartis and Regeneron; consultancy roles with AstraZeneca, Genentech, Gilead, ION, Merck, Medscape, MJH Holdings, Novartis, Pfizer, PUMA, Seagen; royalties from Wolters Kluwer (UptoDate); McGraw Hill (Goodman and Gillman). RAS reports institutional research funding from OBI Pharmaceuticals, Quantum Leap Healthcare Collaborative, AstraZeneca and Gilead; serves on AstraZeneca and Stemline Advisory Boards; Gilead Speaker’s Bureau; consultancy role with Quantum Leap Healthcare Collaborative. ESR reports grants from V Foundation, NIH, Susan G. Komen; institutional research funding from GSK, Seagen, Pfizer, Lilly; consulting and honoraria from Novartis, Merck, Seagen, Astrazeneca, Lilly; Cancer Awareness Network Board member and support from ASCO, NCCN. AA is an employee of Quantum Leap Healthcare Collaborative. ADM reports institutional research funding from Novartis, Pfizer, Genentech and Neogenomics; Program Chair, Scientific Advisory Committee, ASCO. LJvV is an advisor for Exai BIo; part-time employee and owns stock in Agendia. JP reports honoraria from Methods in Clinical Research - Faculty SCION Workshop; support from ASCO and Advocate Scholarship; AACR - SSP Program; VIVLI, U Wisc SPORE - EAB, QuantumLEAD - COVID DSMB, PCORI - Reviewer and ISPY Advocate Lead. DY reports research funding from NIH/NCI P30 CA 077598, P01 CA234228-01 and R01CA251600; consulting fees from Martell Diagnostics; and honoraria and travel for speaking at the \"International Breast Cancer Conference.\" NMH reports institutional research finding from NIH. LJE reports past funding from Merck \u0026 Co. and Moderna for an unrelated trial; participation on an advisory board for Blue Cross Blue Shield; personal fees from UpToDate; unpaid board member of QLHC. HSR reports institutional research support from AstraZeneca, Daiichi Sankyo, Inc., F. Hoffmann-La Roche AG/Genentech, Inc., Gilead Sciences, Inc., Lilly; Merck \u0026 Co., Novartis Pharmaceuticals Corporation, Pfizer, Stemline Therapeutics, OBI Pharma, Ambrx, Greenwich Pharma; advisory and consulting roles with Chugai, Puma, Sanofi, Napo, and Mylan. ZQ has research support from NIH NIDDK DiabDocs K12DK133995 and a Larry L Hillblom Foundation Start Up Grant consulting role with Sanofi. All other authors declare no competing interests.","formattedTitle":"Identification of Early Symptoms Associated with Subsequent Immune-related Adverse Events in the I-SPY clinical trial","fulltext":[{"header":"Background","content":"\u003cp\u003eRecent advances in cancer immunotherapy, including the advent of immune checkpoint inhibitors (ICIs), have revolutionized cancer treatment, providing survival benefits across an array of cancer types, including breast cancer[\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. ICIs harness patients\u0026rsquo; adaptive immune system to treat cancer by antagonizing the receptors responsible for dampening the effective activity of immune cells and their ability to identify and remove tumor cells[\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eFor breast cancer, pembrolizumab is approved in both early and metastatic triple-negative disease[\u003cspan additionalcitationids=\"CR4\" citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. In addition, early trials have shown promising results of ICIs in early-stage hormone receptor-positive breast cancers[\u003cspan additionalcitationids=\"CR7 CR8 CR9\" citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eICIs are associated with immune related adverse events (irAE) that are often long-term[\u003cspan additionalcitationids=\"CR12 CR13\" citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]. These likely have multiple etiologies, leading to a broad range of autoimmune syndromes including potentially life-threatening events[\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eA systemic review by Balibegloo and colleagues[\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e] reported that approximately 12% of patients experienced any grade hypothyroidism, and about 1% of patients developed any grade adrenal insufficiency (AI), while being treated with PD-1 or PD-L1 inhibitors in the metastatic breast cancer setting. However, their incidence is higher in the adjuvant or neoadjuvant setting where these drugs are part of the standard of care for some tumor types[\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e]. Although the incidence of grade 3 or higher endocrine toxicity is low, even lower grade toxicity is nearly always permanent[\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e] and requires life-long medical management[\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eIdentification of patients at higher risk of developing irAEs during treatment could offer an opportunity for early detection and intervention to prevent or minimize the impact of these conditions. For example, adjunctive immunomodulatory agents might help prevent or attenuate irAEs, and patients predicted to be at high risk for irAEs who experience rapid treatment response might consider discontinuing ICI therapy early.\u003c/p\u003e \u003cp\u003eWe investigated the feasibility of early identification of patients at high risk of developing AI or hypothyroidism by analyzing the early symptom landscape. We used prospectively collected toxicity data from the ISPY-2 trial[\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e] and computed an overall symptom burden (OSB) score incorporating duration and severity of symptom to construct a predictive framework for clinicians to identify early symptoms that could forecast a later irAE, and conducted statistical enrichment analyses to uncover symptoms enriched in patients with and without the irAEs of interest.\u003c/p\u003e"},{"header":"Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eStudy Population\u003c/h2\u003e \u003cp\u003e The I-SPY2 clinical trial is an ongoing, multicenter, open-label, response-adaptive phase 2 platform trial with multiple experimental arms evaluating new agents alone or in combination with weekly neoadjuvant paclitaxel followed by 4 additional cycles of doxorubicin/cyclophosphamide (AC) chemotherapy for women with high-risk stage II/III breast cancer (NCT01042379). Information regarding design, eligibility, and assessments has been previously reported[\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eThis analysis used data from 2247 women enrolled between 2010\u0026ndash;2022 at 30 clinical centers across the US. We identified 482 participants who received ICI therapy (\u0026plusmn;\u0026thinsp;other experimental agents) concurrent with 12 cycles of weekly paclitaxel (Supplementary Figure \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e) in 6 arms of I-SPY2. Information on treatment regimens is included in Supplementary Table\u0026nbsp;1.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003ePredictor Variables\u003c/h3\u003e\n\u003cp\u003e32 symptoms that occurred on or before weeks 4, 6 and 8 were included as predictor variables. These were symptoms selected by a group of clinicians, researchers and patient advocates in the I-SPY patient outcomes working group as key symptoms to assess independently or for overall patient toxicity, quality of life or symptom burden. Adverse events in I-SPY2 are documented and graded using the Common Terminology Criteria for Adverse Events (CTCAEv4.0) weekly from treatment start till the study timepoint for symptom assessment (week 8). Start and end dates (if resolved) of the symptoms during this time frame were documented for each event. Fifty-five CTCAE categories were collapsed into 32 symptoms predictors (Supplementary Table\u0026nbsp;2). For symptoms graded on a 1\u0026ndash;5 scale, grade 1 was excluded (deemed occasional/asymptomatic and did not interfere with activities of daily living). Grade 1 symptom events were included for dizziness, loss of fingernails or toenails, painful urination, and palpitations where events are not on a 5-point scale.\u003c/p\u003e\n\u003ch3\u003eOutcomes\u003c/h3\u003e\n\u003cp\u003eThe primary outcome was the development of immune-related adverse AI (due to hypophysitis or primary adrenalitis) and hypothyroidism defined as grade 1 or higher at any time during the study, or through follow-up up until 1 year. These AEs were adjudicated by the I-SPY Safety Working Group, an external independent data safety and monitoring working group (CCSA), and endocrinologists (ZQ and RNC). Diagnoses for hypothyroidism were triggered through routine TSH and free T4 testing and symptom triggers. Diagnoses for AI were primarily symptom triggered, followed by cortisol and ACTH testing. These irAEs were selected because they are common and often become life-altering long-term irAEs. Groups considered for statistical comparison were patients that developed an irAE(s) of interest after treatment with experimental immunotherapy(case) vs those that did not(control).\u003c/p\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003eStatistical Analysis\u003c/h2\u003e \u003cp\u003eA variable importance analysis was performed in two ways, first to characterize the association of OSB with the development of irAEs, and second, to understand sentinel symptoms associated with the development of irAEs of interest.\u003c/p\u003e \u003cp\u003e \u003cstrong\u003eOverall Symptom Burden\u003c/strong\u003e \u003cp\u003eFor each patient, cumulative area under the curves (AUC) were calculated for each of the 32 symptoms based on their grade and duration at weeks 4, 6 and 8 as follows\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cspan class=\"InlineEquation\"\u003e \u003cspan class=\"mathinline\"\u003e\\(\\:AUC\\left(symptom\\:|\\:Week\\:n\\right)=\\:\\sum\\:_{day\\:1}^{n\\times\\:7}AE\\:grade\\:\\times\\:\\:days\\:(in\\:specified\\:duration\\:that\\:AE\\:persisted\\)\u003c/span\u003e \u003c/span\u003e)\u003c/p\u003e \u003cp\u003ePatients that developed an irAE of interest before the analysis timepoint (i.e. week 4, 6 or 8) were censored from the analysis of that timepoint for that outcome of interest.\u003c/p\u003e \u003cp\u003eLogistic regression was performed first between each different symptom AUC and the outcome of interest; those symptoms that had positive association and Wald test p-values\u0026thinsp;\u0026lt;\u0026thinsp;0.2 were selected to calculate the OSB. The OSB is the sum of the selected symptoms. A logistic regression was then performed between the continuous OSB and outcome. For each time point, p-values were adjusted for multiple testing using the Bonferroni correction and an association was considered significant if the adjusted p-value was \u0026lt;\u0026thinsp;0.05. ROC curves were generated for analyses with significant association of the OSB with outcome.\u003c/p\u003e \u003cp\u003eThe model developed above was validated in an external dataset comprising of 4 distinct arms in ISPY2 of 217 patients. Methods for additional statistical analyses and detailed validation methods are detailed in Supplementary Methods. All statistical analyses and visualizations were performed using R version 4.2.2.\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003ePatient Population\u003c/h2\u003e \u003cp\u003eAmong the I-SPY patients, 482 were randomized to one of the six arms including an immunotherapy component (median age\u0026thinsp;=\u0026thinsp;47.3 years (range, 20\u0026ndash;79 years)). 382 patients were white(79.3%) and 95 were non-white(19.7%). 78 patients reported Hispanic ethnicity (16.2%). Detailed demographics are summarized in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eDemographic characteristics of patient population by irAE type.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eOverall\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eHypothyroidism\u003c/p\u003e \u003cp\u003e(N\u0026thinsp;=\u0026thinsp;61)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eNo Hypothyroidism\u003c/p\u003e \u003cp\u003e(N\u0026thinsp;=\u0026thinsp;421)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eAdrenal insufficiency\u003c/p\u003e \u003cp\u003e(N\u0026thinsp;=\u0026thinsp;38)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eNo Adrenal insufficiency\u003c/p\u003e \u003cp\u003e(N\u0026thinsp;=\u0026thinsp;444)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMean (SD)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e47.6 (11.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e46.4 (10.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e47.8 (11.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e48.5 (11.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e47.6 (11.6)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMedian [Min, Max]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e47.3 [20.0, 79.0]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e45.5 [28.8, 71.0]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e47.9 [20.0, 79.0]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e49.5 [31.0, 79.0]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e47.0 [20.0, 76.0]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eRace\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAmerican Indian/ Alaskan Native\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e3 (0.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1 (1.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e2 (0.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e3 (0.7%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAsian\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e32 (6.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2 (3.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e30 (7.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1 (2.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e31 (7.0%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBlack\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e60 (12.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2 (3.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e58 (13.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e5 (13.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e55 (12.4%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eWhite\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e382 (79.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e53 (86.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e329 (78.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e32 (84.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e350 (78.8%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eEthnicity\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHispanic\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e78 (16.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e7 (11.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e71 (16.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e8 (21.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e70 (15.8%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNot Hispanic\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e402 (83.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e54 (88.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e348 (82.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e30 (78.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e372 (83.8%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eAmong the cohort, hypothyroidism was more common(N\u0026thinsp;=\u0026thinsp;61 or 12.7%) vs AI (38 or 7.9%). 21 of 61 hypothyroidism patients had hyperthyroidism an average of 4 weeks prior to hypothyroidism diagnosis. AI was most often attributed to confirmed or probable hypophysitis(65.8%), probable primary AI(2.6%), AI of unknown type(15.8%), and reported AI by site but unconfirmed at time of analysis by endocrinologist(15.8%)(Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e1\u003c/span\u003eA). No significant differences were observed in irAE incidence based on age, race or ethnicity.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eEarly Symptoms Predictive of Adrenal Insufficiency\u003c/h3\u003e\n\u003cp\u003eAI was diagnosed at a median of 15 weeks(105 days) from treatment start(range\u0026thinsp;=\u0026thinsp;6.6\u0026ndash;41.6 weeks(46\u0026ndash;291 days)), which was after completion of the immunotherapy phase and close to the end of all neoadjuvant chemotherapy(Supplementary Figure \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e). One patient developed AI between weeks 4 and 6 and was censored from the weeks 6 and 8 analyses, while an additional patient developed hypothyroidism between weeks 6\u0026ndash;8 and was censored from the week 8 analyses.\u003c/p\u003e \u003cp\u003eOSB at week 6 was the timepoint most predictive of AI (adjusted p\u0026thinsp;\u0026lt;\u0026thinsp;0.001) (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e2\u003c/span\u003eA). Symptoms used to calculate OSB at week 6 were constipation, diarrhea, headache, joint pain, shortness of breath, taste changes, and vomiting (Supplementary Table\u0026nbsp;3). The trajectory of these symptoms as swimmers plots in patients with and without AI are shown on Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e1\u003c/span\u003eA. Results indicated that shortness of breath, for example, presented early and persisted through follow up, as did GI symptoms such as diarrhea. Patients who would later develop AI had multiple occurrences of bouts of diarrhea (instances\u0026thinsp;\u0026gt;\u0026thinsp;2) compared to patients who did not. Patients who would later develop AI also had multiple co-occurring symptoms, the foremost of which was diarrhea and vomiting, while patients without AI rarely had these symptoms. These symptoms preceded the AI diagnoses by an average of 63 days (9 weeks).\u003c/p\u003e\n\u003ch3\u003eModel Performance\u003c/h3\u003e\n\u003cp\u003eA logistic regression was performed between the continuous OSB and outcome.\u003c/p\u003e \u003cp\u003eThe Youden\u0026rsquo;s best OSB threshold was 14.5 with a sensitivity of 0.57, specificity of 0.74 and accuracy of 0.73 (Supplementary Table\u0026nbsp;6 and Supplementary Figure \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003eA). Mean OSB at week 4 of patients that developed AI was almost double that of patients that did not develop AI (mean(SD) cases\u0026thinsp;=\u0026thinsp;13.66 (19.73) vs controls\u0026thinsp;=\u0026thinsp;5.2 (10.92)), and this gap widened at weeks 6 (31.11(39.58) vs 10.6(10.67)) and 8 (49.81(57.79) vs 18.5(30.68)(Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e2\u003c/span\u003eA and Supplementary Table\u0026nbsp;3).\u003c/p\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eModel Validation\u003c/h2\u003e \u003cp\u003eIn an external validation set of 217 patients receiving ICI with an oral paclitaxel backbone, 11 patients later developed AI. Patients were demographically comparable to the training set above (Supplementary Table\u0026nbsp;8). The OSB was computed as the sum AUCs of the symptoms mentioned above. Patients that later developed AI had a higher mean OSB at week 6 than patients that did not develop AI(44.6 vs 17.4, p\u0026thinsp;=\u0026thinsp;0.02). Using the Youden\u0026rsquo;s best threshold above of 14.5, the model\u0026rsquo;s sensitivity was 54.6%, specificity was 73.8%, positive predictive value was 10%, negative predictive value was 96.8% and model accuracy was 72.8%.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eLongitudinal Symptom Enrichment\u003c/h2\u003e \u003cp\u003eAll symptoms were also analyzed as binary variables (present vs absent) at 4, 6, and 8 weeks (Supplementary Table\u0026nbsp;4). Unlike OSB, this analysis did not capture severity and duration of a given symptom and was performed so that we could identify additional symptoms that may be short lasting and more severe, or long lasting and less severe and may not be captured by the OSB analysis. Such an example was heart palpitations that was significant at the week 8 timepoint (adjusted p\u0026thinsp;=\u0026thinsp;0.047) but receded p\u0026thinsp;\u0026gt;\u0026thinsp;0.05 in the OSB analysis (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA and Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). Interestingly, palpitations had similar rates of incidence in patients that developed AI compared to those that did not up until week 6, after which, patients that developed AI had a 4-fold higher rate of palpitations (8.3% in case vs. 1.9% in control).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eBinary (presence absence) enrichment analysis for symptoms. Symptoms with p-value\u0026thinsp;\u0026lt;\u0026thinsp;0.2 at at least one timepoint are shown, ordered by p-value at week 4, with timepoints with adjusted p-value\u0026thinsp;\u0026lt;\u0026thinsp;0.05 highlighted in green.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"13\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c10\" colnum=\"10\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c11\" colnum=\"11\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c12\" colnum=\"12\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c13\" colnum=\"13\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colspan=\"4\" nameend=\"c5\" namest=\"c2\"\u003e \u003cp\u003eWEEK 4\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"4\" nameend=\"c9\" namest=\"c6\"\u003e \u003cp\u003eWEEK 6\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"4\" nameend=\"c13\" namest=\"c10\"\u003e \u003cp\u003eWEEK 8\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colspan=\"13\" nameend=\"c13\" namest=\"c1\"\u003e \u003cp\u003eADRENAL INSUFFICIENCY\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSymptom\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eIn Cases (N\u0026thinsp;=\u0026thinsp;38)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eIn Control (N\u0026thinsp;=\u0026thinsp;423)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eOdds Ratio\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eAdjusted p-value\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eIn Cases (N\u0026thinsp;=\u0026thinsp;37)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eIn Control (N\u0026thinsp;=\u0026thinsp;423)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003eOdds Ratio\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c9\"\u003e \u003cp\u003eAdjusted p-value\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c10\"\u003e \u003cp\u003eIn Cases (N\u0026thinsp;=\u0026thinsp;36)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c11\"\u003e \u003cp\u003eIn Control (N\u0026thinsp;=\u0026thinsp;423)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c12\"\u003e \u003cp\u003eOdds Ratio\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c13\"\u003e \u003cp\u003eAdjusted p-value\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDiarrhea\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e18 (47.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e112 (26.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2.49\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.0069\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e17 (45.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e131 (31%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e1.89\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0.048\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e17 (47.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e143 (33.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e1.75\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e0.077\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eConstipation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2 (5.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1 (0.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e23.03\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.019\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e2 (5.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e1 (0.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e23.68\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0.018\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e3 (8.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e3 (0.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e12.56\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e0.0075\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eShortness of breath\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e6 (15.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e23 (5.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3.25\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.024\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e7 (18.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e37 (8.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e2.43\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0.051\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e6 (16.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e41 (9.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e1.86\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e0.15\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHeadache\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e4 (10.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e14 (3.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3.42\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.052\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e4 (10.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e17 (4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e2.88\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0.079\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e4 (11.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e20 (4.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e2.51\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e0.11\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePalpitations\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2 (5.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4 (0.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e5.78\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.08\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e2 (5.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e7 (1.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e3.38\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0.16\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e3 (8.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e8 (1.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e4.69\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e0.047\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFatigue\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e4 (10.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e18 (4.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2.64\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.097\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e6 (16.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e31 (7.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e2.44\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0.065\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e5 (13.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e49 (11.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e1.23\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e0.42\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTaste changes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1 (2.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2 (0.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e5.65\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.23\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e2 (5.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e2 (0.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e11.88\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0.034\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e2 (5.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e2 (0.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e12.22\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e0.032\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDizziness\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3 (7.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e29 (6.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.16\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e6 (16.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e39 (9.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e1.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0.14\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e6 (16.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e46 (10.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e1.64\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e0.21\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"13\" nameend=\"c13\" namest=\"c1\"\u003e \u003cp\u003e\u003cb\u003eHYPOTHYROIDISM\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eSymptom\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003eIn Cases (N\u0026thinsp;=\u0026thinsp;58)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cb\u003eIn Control (N\u0026thinsp;=\u0026thinsp;400)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cb\u003eOdds Ratio\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u003cb\u003eAdjusted p-value\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cb\u003eIn Cases (N\u0026thinsp;=\u0026thinsp;58)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u003cb\u003eIn Control (N\u0026thinsp;=\u0026thinsp;400)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u003cb\u003eOdds Ratio\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e\u003cb\u003eAdjusted p-value\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e\u003cb\u003eIn Cases (N\u0026thinsp;=\u0026thinsp;50)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e\u003cb\u003eIn Control (N\u0026thinsp;=\u0026thinsp;400)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e\u003cb\u003eOdds Ratio\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e\u003cb\u003eAdjusted p-value\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFatigue\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e6 (10.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e16 (4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2.76\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.047\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e10 (17.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e27 (6.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e2.87\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0.011\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e8 (16%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e44 (11%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e1.54\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e0.2\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRash\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e7 (12.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e22 (5.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2.35\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.059\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e12 (20.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e31 (7.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e3.09\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0.0037\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e12 (24%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e34 (8.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e3.39\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e0.0021\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eItching\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2 (3.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4 (1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3.52\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.17\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e3 (5.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e6 (1.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e3.57\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0.093\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e3 (6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e8 (2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e3.12\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e0.11\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eShortness of breath\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5 (8.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e24 (6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.48\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e9 (15.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e36 (9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e1.85\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0.097\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e6 (12%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e39 (9.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e1.26\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e0.38\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"13\"\u003eA)\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eIn terms of symptoms that were overlapping between the OSB and binary analyses, prior to week 4, diarrhea was the most enriched symptom in patients that later developed AI, followed by constipation and shortness of breath (OR\u0026thinsp;=\u0026thinsp;2.49, 23.03 and 3.25, adjusted p\u0026thinsp;=\u0026thinsp;0.0069, 0.019 and 0.024 respectively). At week 4, diarrhea was present in 47.4% of the patients that later developed AI compared to 26.5% in patients who did not develop AI (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). The difference in the proportions declined by week 6 (45.9% vs 31%) and was no longer significant at week 8 (47.2% vs 33.8%). Constipation was the only symptom that was significantly enriched at all three study timepoints. Constipation was present in twice as many patients who later developed AI than those without at weeks 4 and 6 (5.3\u0026ndash;5.4% in cases vs\u0026thinsp;\u0026lt;\u0026thinsp;1% in control) and in almost 8 times as many patients with AI by week 8 (8.3% vs. 0.7%). Patients that developed AI had almost twice the proportion of fatigue at weeks 4 (10.5% vs 4.3%) and 6 (16.2% vs 7.3%) compared to patients that did not develop AI, though this proportion nearly equalized by week 8 (13.9% vs 11.6%). Similarly, patients who later developed AI had almost three times the reported shortness of breath (15.8% vs 5.4% ) and headache (10.5% vs. 3.3%), by week 4. All analyses were also performed excluding cases that were reported as AI but were not confirmed by endocrinologists, with no significant differences in results.\u003c/p\u003e \u003cp\u003eTo further analyze the co-occurrence of symptoms, heatmaps were generated with 32 symptoms that developed by week 6 to visualize symptom clusters (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eC). Diarrhea, vomiting, headache, fatigue and shortness of breath co-occurred in the majority of patients that later developed AI.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003eMultiplicity of symptoms and irAE development\u003c/h2\u003e \u003cp\u003eFor each timepoint, a logistic regression was performed to determine if presence of a higher number of symptoms was associated with greater odds for later developing AI (Supplementary Table\u0026nbsp;5). At week 4, the odds of developing AI with each additional symptom increased by 1.46 (adjusted p-value\u0026thinsp;=\u0026thinsp;0.008, mean number of symptoms in cases\u0026thinsp;=\u0026thinsp;1.32 vs. controls\u0026thinsp;=\u0026thinsp;0.75). At week 6, the odds of developing AI with each additional symptom decreased to 1.33 (adjusted p-value\u0026thinsp;=\u0026thinsp;0.027, in cases\u0026thinsp;=\u0026thinsp;1.65 vs. controls\u0026thinsp;=\u0026thinsp;1.01). Number of symptoms was no longer a significant predictor by week 8 due to increasing number of symptoms in controls as well (adjusted p-value\u0026thinsp;=\u0026thinsp;0.07, in cases\u0026thinsp;=\u0026thinsp;1.83 vs. controls\u0026thinsp;=\u0026thinsp;1.49).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003eEarly Symptoms Predictive of Hypothyroidism\u003c/h2\u003e \u003cp\u003eHypothyroidism was diagnosed at a median of 14 weeks (99 days) from treatment start (range\u0026thinsp;=\u0026thinsp;20\u0026ndash;208) which was close to the end of the immunotherapy phase of the treatment. Two patients developed hypothyroidism prior to week 4, and were censored from weeks 4,6 and 8 analyses, while an additional eight patients developed hypothyroidism between weeks 6 and 8 and were censored from the week 8 analyses.\u003c/p\u003e \u003cp\u003eFor hypothyroidism, week 6 was also the most significant timepoint in the predictive model (adjusted p\u0026thinsp;=\u0026thinsp;0.012) (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e2\u003c/span\u003eB). Symptoms included in the OSB calculation (i.e. that were independently associated with hypothyroidism with p\u0026thinsp;\u0026gt;\u0026thinsp;0.2) were painful urination, rash, shortness of breath and fatigue (Supplementary Table\u0026nbsp;3).\u003c/p\u003e \u003cp\u003eThe Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e1\u003c/span\u003eB swimmers plot illustrates that higher-grade fatigue and rash were significantly enriched in the patients that later developed hypothyroidism. These symptoms presented earlier and lasted longer in those with immune related hypothyroidism. Multiple occurrences of rash, each relatively short duration, were more frequent, and fatigue was present for longer durations in patients with hypothyroidism. Shortness of breath was also more frequent and longer lasting in patients who subsequently developed hypothyroidism. These symptoms preceded the hypothyroidism diagnoses by an average of 57 days (approx. 8 weeks).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003eModel Performance\u003c/h2\u003e \u003cp\u003eA logistic regression was performed between the continuous OSB and outcome followed by .\u003c/p\u003e \u003cp\u003eThe Youden\u0026rsquo;s best threshold OSB was 9.5 with a sensitivity of 0.43, specificity of 0.82 and accuracy of 0.77 (Supplementary Table\u0026nbsp;7 and Supplementary Figure \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003eB). As early as week 4, the mean OSB for patients that had hypothyroidism was more than two times higher than for patients that did not develop hypothyroidism (mean(SD) for case\u0026thinsp;=\u0026thinsp;13.66(19.73) vs. control\u0026thinsp;=\u0026thinsp;5.2(10.92)), and this gap widened greatly through week 8 (case\u0026thinsp;=\u0026thinsp;14.6(22.66) and 20.64(37.21) vs control\u0026thinsp;=\u0026thinsp;6.7(16.97) and 9.27(24.09) at weeks 6 and 8, respectively)(Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e2\u003c/span\u003eB, Supplementary Table\u0026nbsp;3).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003eModel Validation\u003c/h2\u003e \u003cp\u003eIn the validation set of 217 patients receiving ICI with an oral paclitaxel backbone, 31 patients later developed hypothyroidism. Patients that later developed hypothyroidism had a higher mean OSB at week 6 than patients that did not develop hypothyroidism, though this difference was not statistically significant(13.5 vs 10). Model\u0026rsquo;s sensitivity was 16.1%, specificity was 84.4%, positive predictive value was 14.7%, negative predictive value was 85.8% and model accuracy was 74.7%.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003eLongitudinal Symptom Enrichment\u003c/h2\u003e \u003cp\u003eBinary analysis of symptoms showed that prior to week 4, fatigue was the only significantly enriched symptom among patients that later developed hypothyroidism(OR\u0026thinsp;=\u0026thinsp;2.76, adjusted p\u0026thinsp;=\u0026thinsp;0.047)(Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e, Supplementary Table\u0026nbsp;4). At week 6, rash was the most significantly enriched symptom, followed by fatigue(OR\u0026thinsp;=\u0026thinsp;3.09 and 2.87, p\u0026thinsp;=\u0026thinsp;0.0037 and 0.011), and at week 8, only rash was significantly enriched in patients that later developed hypothyroidism(OR\u0026thinsp;=\u0026thinsp;3.39, p\u0026thinsp;=\u0026thinsp;0.0021). Fatigue was present in more than twice as many patients that developed hypothyroidism compared to those that did not at weeks 4(10.3% vs 4.0%) and 6(17.2% vs 6.8%)(Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eB and Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e); however, it was not significantly enriched in cases compared to controls at week 8, indicating that while early presence of fatigue can be an indicator of hypothyroidism, the growing incidence of fatigue among all patients undergoing treatment renders this an insignificant predictor as treatment progresses. Rash was present in twice as many patients at week 6 (20.7% vs 7.8%) and three times as many patients at week 8(24% vs 8.5%)(Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eB and Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eFurther analysis and visualization of co-occurring symptoms revealed that diarrhea, rash, dizziness, and fatigue commonly co-occurred prior to week 6 among patients that later developed hypothyroidism(Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eD). However, unlike for AI, the number of symptoms was not a significant predictor of later development of hypothyroidism(Supplementary Table\u0026nbsp;5).\u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eICIs are now a standard of care as neoadjuvant therapy combined with chemotherapy for stage II-III triple negative breast cancer, and for patients with PD-L1 positive metastatic TNBC. The current study utilized an analysis framework to identify unique symptom clusters that can be observed early on during ICI therapy and are predictive of developing the two most common irAEs, hypothyroidism and AI. Using prospectively collected clinical trial data from I-SPY2, we developed and validated a framework for early identification of patients at elevated risk for endocrine immune-related adverse events(irAEs). Our findings suggest that cumulative early symptom burden, especially by week 6 of treatment, provides meaningful predictive insight and could inform clinical strategies for early intervention. Identification of patients who are high risk for these irAEs could allow more frequent diagnostic testing (i.e. hormone level measurements) and proactive early treatment. While TSH is part of the standard of care monitoring, and cortisol level monitoring is now standard in I-SPY, by the time these tests are abnormal, neither AI nor hypothyroidism is likely reversible. Thus, the ability to find earlier indicators could potentially lead to proactive preventive interventions.\u003c/p\u003e \u003cp\u003eOverall, our results demonstrated that rash, shortness of breath, and fatigue commonly co-occurred among patients who later developed hypothyroidism (after a median of 57 days), while diarrhea, headache, vomiting, fatigue and shortness of breath were predictive of AI (after a median of 63 days). Dermatological symptoms, such as itching and rash, were also highly ranked symptoms in the first 4 to 8 weeks that were predictive for the development of hypothyroidism, while gastrointestinal symptoms, such as constipation and diarrhea, were predictive for the development of AI. Diarrhea often precedes constipation in both cases and controls, as symptom management for diarrhea can frequently lead to constipation in patients.\u003c/p\u003e \u003cp\u003eFatigue remains one the most frequently reported side effects of cancer treatment and can be highly burdensome for patients. In the present study, fatigue that occurred early in treatment was a predictor of later development of hypothyroidism, however its significance dissipated as treatment continued and a high number of patients in both groups began to develop fatigue. This reinforces the need to consider symptoms longitudinally and in the context of other symptoms as markers of irAEs. Previous studies have found that early immune dysregulation, specifically with respect to cytokines associated with T-cell activation and autoimmune disease eventually lead to a discrete case of severe irAE[\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]. Many of these early symptoms included in the model, such as diarrhea and constipation, fatigue, rash, headache, joint pain, etc. are symptoms that are commonly associated with immune dysregulation, thus it is possible that early immune dysregulation that ultimately leads to irAE development manifests through a combination of these symptoms.\u003c/p\u003e \u003cp\u003eWe found that an overall symptom burden(OSB) score\u0026mdash;capturing both severity and duration of early symptoms\u0026mdash;was strongly associated with the later development of irAE, with a significant signal emerging as early as week 6 of therapy. Specific symptoms contributing to this burden included gastrointestinal (diarrhea, constipation, vomiting), respiratory (shortness of breath), and symptoms commonly associated with immune dysregulation (headache, joint pain, rash), These symptoms preceded the clinical diagnosis of AI by an average of nine weeks and hypothyroidism by 8 weeks, indicating a potential window for proactive monitoring and early evaluation. The consistency of these findings in an external validation cohort, despite different treatment backbones (oral paclitaxel rather than intravenous), reinforces the potential generalizability of this approach.\u003c/p\u003e \u003cp\u003eThe results of this work can directly inform clinical care in several ways: 1) help clinicians risk-stratify follow-up and monitoring of patients at high-risk for toxicity; and 2) help develop and evaluate strategies to mitigate toxicities. Ultimately, this aids the overall goal of helping with clinical decision weighing patient toxicities against the benefit from immunotherapy. Through identifying those patients who are at highest risk, one can test de-escalation or supportive care strategies to decrease toxicity risk. Our models provide a strategy for early guidance and permit more personalized care by balancing increased efficacy with reduced risks of chronic and debilitating toxicities, as well as promoting early supportive care interventions that could mitigate the long -term damage that could result from these toxicities.\u003c/p\u003e \u003cp\u003ePrevious studies examining ways by which to predict immunotherapy-related irAEs have focused on pre-treatment gene markers, host genetics, and circulating protein biomarkers including serum autoantibodies[\u003cspan additionalcitationids=\"CR23 CR24\" citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e]. The current study utilizes a novel framework, leveraging early treatment-emergent symptoms in patients across 6 different immunotherapies, to determine risk of later development of serious irAEs. Early symptoms that we found to be enriched in patients with these irAE, provide clinicians a straightforward primary assessment tool in clinic for identifying patients that may have higher risk of these irAEs. The OSB, a sum of AUCs of selected symptoms, provides an assessment timepoint at week 6, where cumulative symptom burden in the early treatment phase provides a personalized risk estimates of the likelihood of later developing these key irAEs. Such a framework could pro-actively identify at-risk patients before the development of a serious irAE, with relatively minimal clinic and provider burden. The high specificity of our model would enable effective screening of patients that are at low risk of later irAEs and distinguishing them from those who might require additional monitoring or diagnostic interventions. Genetic markers predictive of predisposition to autoimmunity and patient reported symptoms have the potential to improve the predictive accuracy of our OSB score, which we are exploring. Patient reported outcomes might be particularly suited to strengthen this model as they can capture even earlier symptoms than the current assessments timed around clinic visits. Since June 2021, routine electronic patient reported outcomes data collection was launched at all I-SPY2 sites[\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e], and we are working on methodologies for incorporating real-time patient reported AE data with our CTCAE-symptom based irAE prediction model to further improve predictive power.\u003c/p\u003e \u003cp\u003eThe clinical implications of our study, along with a growing evidence base regarding the adverse effects of immunotherapy, could enable more proactive management of patients to prevent these irAEs. The identification of symptoms (both early and those that persist), in combination with the identification of timepoints at which their emergence becomes most predictive, can make an important contribution to comprehensive survivorship care for women with breast cancer. Going forward, we will prospectively validate our methods in the ongoing I-SPY2.2 platform trial, as well as in other cancer trials and work to develop strategies to redirect therapy or develop preventive interventions when there are early indicators of treatment-induced long-term toxicities.\u003c/p\u003e \u003cdiv id=\"Sec19\" class=\"Section2\"\u003e \u003ch2\u003eLimitations\u003c/h2\u003e \u003cp\u003eWe analyzed data prospectively collected during a clinical trial, and the overall health and co-morbidity status of trial participants may not be representative of all patients with stage II/III breast cancer. All our patients received weekly paclitaxel, but the type of chemotherapy partner administered along with ICIs can influence the type and frequency of early symptoms. However, three of the six treatment arms used in this analysis included additional experimental therapies along with ICI and paclitaxel which suggest generalizability. Since many irAEs are rare, we could only analyze associations between early symptoms and subsequent irAE for the two most common AEs, hypothyroidism and AI. If early symptoms predict development of other irAEs remains unknown. Although our models performed well in the validation cohort, the positive predictive value remained modest, reflecting the low absolute incidence of these irAEs, and underscoring that symptom burden is likely only one piece of the risk puzzle.\u003c/p\u003e \u003c/div\u003e"},{"header":"Conclusions","content":"\u003cp\u003eWe show that a constellation of non-specific treatment emergent symptoms that develop as early as four to six weeks of ICI therapy could predict higher risk of future immune-related hypothyroidism and AI in patients with breast cancer. Since routine monitoring for AI with serial hormone level measurements is not currently recommended and the frequency of thyroid hormone assessments during therapy also vary, early onset fatigue, diarrhea, rash, shortness of breath could alert physicians to be more diligent in on-treatment and post-treatment monitoring for these endocrinopathies. Work is ongoing to develop strategies for monitoring and potential treatment mitigation to reduce the risk of permanent endocrinopathies.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eThis resear\u003c/strong\u003ech was supported by the National Cancer Institute of the National Institutes of Health under award number P01CA210961 and R21CA2588218. The authors wish to acknowledge the generous support of the study sponsors, Quantum Leap Healthcare Collaborative (QLHC, 2013 to present) and the Foundation for the National Institutes of Health (2010 to 2012). The authors sincerely appreciate the ongoing support for the I-SPY2 Trial from the Safeway Foundation, the William K. Bowes, Jr. Foundation, Give Breast Cancer the Boot, QLHC and the Breast Cancer Research Foundation.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDISCLOSURES\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eCY reports institutional research grant from NCI/NIH; salary support and travel reimbursement from Quantum Leap Healthcare Collaborative; US patent titled, \u0026ldquo;Breast cancer response prediction subtypes,\u0026rdquo; (No. 18/174,491); and University of California Inventor Share. RN reports research funding from Arvinas, AstraZeneca, BMS, Corcept Therapeutics, Genentech/Roche, Gilead, GSK, Merck, Novartis, OBI Pharma, OncoSec, Pfizer, Relay, Seattle Genetics, Sun Pharma, Taiho; advisory roles with AstraZeneca, BeyondSpring, Daiichi Sankyo, Exact Sciences, Fujifilm, GE, Gilead, Guardant Health, Infinity, iTeos, Merck, Moderna, Novartis, OBI, Oncosec, Pfizer, Sanofi, Seagen, Stemline. LP reports institutional research funding from Susan Komen Foundation, Breast Cancer Research Foundation, National Cancer Institute, Pfizer, AstraZeneca, Menarini/Stemline, Bristol Myers Squibb, Merck and Co,; consulting fees from AstraZeneca, Merck, Novartis, Genentech, Natera, Personalis, ExactSciences, Stemline/Menarini; patent titled, \u0026ldquo;Method of measuring residual cancer and predicting patient survival,\u0026rdquo; (No. 7711494); and DSMB member of the DYNASTY Breast02, OPTIMA and PARTNER Trials. AJC reports institutional research funding from Merck, Amgen, Puma, Seagen, Pfizer, Olema; advisory roles with Astra Zeneca and Genentech. MCL is an employee of Natera, Inc., with stock or options to own stock, and further reports grants/contracts (funding to institution: Mayo Clinic) from Eisai, Exact Sciences Corporation, Genentech, Genomic Health, GRAIL, Menarini Silicon Biosystems, Merck, Novartis, Seattle Genetics and Tesaro; travel support reimbursement from AstraZeneca, Genomic Health and Ionis; and ad hoc advisory board meetings (all funds to Mayo Clinic; no personal compensation) from AstraZeneca, Celgene, Roche/Genentech, Genomic Health, GRAIL, Ionis, Merck, Pfizer, Seattle Genetics and Syndax. HH reports participating on advisory board for Pfizer. HS reports consulting fees from Eli Lilly, Novartis, AstraZeneca, PUMA and Sermonix. CI reports institutional research funding from Tesaro/GSK, Seattle Genetics, Pfizer, AstraZeneca, BMS, Genentech, Novartis and Regeneron; consultancy roles with AstraZeneca, Genentech, Gilead, ION, Merck, Medscape, MJH Holdings, Novartis, Pfizer, PUMA, Seagen; royalties from Wolters Kluwer (UptoDate); McGraw Hill (Goodman and Gillman). RAS reports institutional research funding from OBI Pharmaceuticals, Quantum Leap Healthcare Collaborative, AstraZeneca and Gilead; serves on AstraZeneca and Stemline Advisory Boards; Gilead Speaker\u0026rsquo;s Bureau; consultancy role with Quantum Leap Healthcare Collaborative. ESR reports grants from V Foundation, NIH, Susan G. Komen; institutional research funding from GSK, Seagen, Pfizer, Lilly; consulting and honoraria from Novartis, Merck, Seagen, Astrazeneca, Lilly; Cancer Awareness Network Board member and support from ASCO, NCCN. AA is an employee of Quantum Leap Healthcare Collaborative. ADM reports institutional research funding from Novartis, Pfizer, Genentech and Neogenomics; Program Chair, Scientific Advisory Committee, ASCO. LJvV is an advisor for Exai BIo; part-time employee and owns stock in Agendia. JP reports honoraria from Methods in Clinical Research - Faculty SCION Workshop; support from ASCO and Advocate Scholarship; AACR - SSP Program; VIVLI, U Wisc SPORE - EAB, QuantumLEAD - COVID DSMB, PCORI - Reviewer and ISPY Advocate Lead. DY reports research funding from NIH/NCI P30 CA 077598, P01 CA234228-01 and R01CA251600; consulting fees from Martell Diagnostics; and honoraria and travel for speaking at the \u0026quot;International Breast Cancer Conference.\u0026quot; NMH reports institutional research finding from NIH. LJE reports past funding from Merck \u0026amp; Co. and Moderna for an unrelated trial; participation on an advisory board for Blue Cross Blue Shield; personal fees from UpToDate; unpaid board member of QLHC. HSR reports institutional research support from AstraZeneca, Daiichi Sankyo, Inc., F. Hoffmann-La Roche AG/Genentech, Inc., Gilead Sciences, Inc., Lilly; Merck \u0026amp; Co., Novartis Pharmaceuticals Corporation, Pfizer, Stemline Therapeutics, OBI Pharma, Ambrx, Greenwich Pharma; advisory and consulting roles with Chugai, Puma, Sanofi, Napo, and Mylan. ZQ has research support from NIH NIDDK DiabDocs K12DK133995 and a Larry L Hillblom Foundation Start Up Grant consulting role with Sanofi. All other authors declare no competing interests.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData Availability Statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSubject-level data for this study is available to approved investigators completing a request form available at: https://www.quantumleaphealth.org/for-investigators/clinicians-proposal-submissions/. \u003c/p\u003e\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eA.B: Conceptualization; Formal analysis; Funding acquisition; Investigation; Methodology; Project administration; Resources; Software; Supervision; Validation; Writing \u0026ndash; original draft; Writing \u0026ndash; review \u0026amp; editingS.U: Data curation; Formal analysis; Investigation; Software; Validation; Visualization; Writing \u0026ndash; original draft; Writing \u0026ndash; review \u0026amp; editingM.M: Data curation; Investigation; Writing \u0026ndash; review \u0026amp; editingR.R: Formal analysis; Software; VisualizationC.Y, A.J.C, M.L, H.H, H.S, E.S.R, P.R, K.Y, K.K, A.T: Data curation; InvestigationL.P: Data curation; Investigation; Writing \u0026ndash; review \u0026amp; editingC.I, R.S: Data curation; Investigation; Project administrationE.P, D.H: Writing \u0026ndash; review \u0026amp; editingM.Mu: Formal AnalysisG.H: Data CurationA.A: Data Curation, SoftwareA.D, L.vV, D.Y, N.H, J.P: Project AdministrationA.O: SupervisionZ.Q, R.N.C: Data curation; Writing \u0026ndash; review \u0026amp; editingL.E: Funding acquisition; Project administration; Resources; Supervision; Writing \u0026ndash; original draft; Writing \u0026ndash; review \u0026amp; editingH.R, R.N: Investigation; Project administration; Supervision; Writing \u0026ndash; review\u0026amp; editing\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eEsfahani, K., et al., \u003cem\u003eA review of cancer immunotherapy: from the past, to the present, to the future.\u003c/em\u003e Curr Oncol, 2020. 27(Suppl 2): p. S87-s97.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNathan, M.R. and P. Schmid, \u003cem\u003eThe emerging world of breast cancer immunotherapy\u003c/em\u003e. Breast, 2018. 37: p. 200\u0026ndash;206.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCortes, J., et al., \u003cem\u003ePembrolizumab plus chemotherapy versus placebo plus chemotherapy for previously untreated locally recurrent inoperable or metastatic triple-negative breast cancer (KEYNOTE-355): a randomised, placebo-controlled, double-blind, phase 3 clinical trial\u003c/em\u003e. Lancet, 2020. 396(10265): p. 1817\u0026ndash;1828.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRizzo, A., et al., \u003cem\u003eKEYNOTE-522, IMpassion031 and GeparNUEVO: changing the paradigm of neoadjuvant immune checkpoint inhibitors in early triple-negative breast cancer\u003c/em\u003e. Future Oncol, 2022. 18(18): p. 2301\u0026ndash;2309.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchmid, P., et al., \u003cem\u003ePembrolizumab for Early Triple-Negative Breast Cancer\u003c/em\u003e. N Engl J Med, 2020. 382(9): p. 810\u0026ndash;821.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHenriques, B., F. Mendes, and D. Martins, \u003cem\u003eImmunotherapy in Breast Cancer: When, How, and What Challenges?\u003c/em\u003e Biomedicines, 2021. 9(11).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSetordzi, P., et al., \u003cem\u003eThe recent advances of PD-1 and PD-L1 checkpoint signaling inhibition for breast cancer immunotherapy\u003c/em\u003e. Eur J Pharmacol, 2021. 895: p. 173867.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePusztai, L., et al., \u003cem\u003eDurvalumab with olaparib and paclitaxel for high-risk HER2-negative stage II/III breast cancer: Results from the adaptively randomized I-SPY2 trial\u003c/em\u003e. Cancer Cell, 2021. 39(7): p. 989\u0026ndash;998 e5.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCardoso F, M.H., Schmid P, et al, \u003cem\u003eLBA21 KEYNOTE-756: Phase III study of neoadjuvant pembrolizumab (pembro) or placebo (pbo) + chemotherapy (chemo), followed by adjuvant pembro or pbo\u0026thinsp;+\u0026thinsp;endocrine therapy (ET) for early-stage high-risk ER+/HER2\u0026ndash; breast cancer\u003c/em\u003e. Annals of Oncology.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLoi S, C.G., Salgado RF, et al, \u003cem\u003eLBA20 A randomized, double-blind trial of nivolumab (NIVO) vs placebo (PBO) with neoadjuvant chemotherapy (NACT) followed by adjuvant endocrine therapy (ET) \u0026plusmn; NIVO in patients (pts) with high-risk, ER+ HER2\u0026thinsp;\u0026ndash;\u0026thinsp;primary breast cancer (BC)\u003c/em\u003e. Annals of Oncology.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCriscitiello, C., et al., \u003cem\u003eManaging side effects of immune checkpoint inhibitors in breast cancer\u003c/em\u003e. Crit Rev Oncol Hematol, 2021. 162: p. 103354.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi, Y., et al., \u003cem\u003eEfficacy and Safety of Adding Immune Checkpoint Inhibitors to Neoadjuvant Chemotherapy Against Triple-Negative Breast Cancer: A Meta-Analysis of Randomized Controlled Trials\u003c/em\u003e. Front Oncol, 2021. 11: p. 657634.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSternschuss, M., et al., \u003cem\u003eEfficacy and safety of neoadjuvant immune checkpoint inhibitors in early-stage triple-negative breast cancer: a systematic review and meta-analysis\u003c/em\u003e. J Cancer Res Clin Oncol, 2021. 147(11): p. 3369\u0026ndash;3379.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eXin, Y., et al., \u003cem\u003eImmune checkpoint inhibitors plus neoadjuvant chemotherapy in early triple-negative breast cancer: a systematic review and meta-analysis\u003c/em\u003e. BMC Cancer, 2021. 21(1): p. 1261.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBalibegloo, M., et al., \u003cem\u003eAdverse events associated with immune checkpoint inhibitors in patients with breast cancer: A systematic review and meta-analysis\u003c/em\u003e. Int Immunopharmacol, 2021. 96: p. 107796.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNanda, R., et al., \u003cem\u003eEffect of Pembrolizumab Plus Neoadjuvant Chemotherapy on Pathologic Complete Response in Women With Early-Stage Breast Cancer: An Analysis of the Ongoing Phase 2 Adaptively Randomized I-SPY2 Trial\u003c/em\u003e. JAMA Oncol, 2020. 6(5): p. 676\u0026ndash;684.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCui, K., et al., \u003cem\u003eImmune checkpoint inhibitors and adrenal insufficiency: a large-sample case series study\u003c/em\u003e. Ann Transl Med, 2022. 10(5): p. 251.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGrouthier, V., et al., \u003cem\u003eImmune Checkpoint Inhibitor-Associated Primary Adrenal Insufficiency: WHO VigiBase Report Analysis\u003c/em\u003e. Oncologist, 2020. 25(8): p. 696\u0026ndash;701.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eD'Abreo, N. and S. Adams, \u003cem\u003eImmune-checkpoint inhibition for metastatic triple-negative breast cancer: safety first?\u003c/em\u003e Nat Rev Clin Oncol, 2019. 16(7): p. 399\u0026ndash;400.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang, H. and D. Yee, \u003cem\u003eI-SPY 2: a Neoadjuvant Adaptive Clinical Trial Designed to Improve Outcomes in High-Risk Breast Cancer\u003c/em\u003e. Curr Breast Cancer Rep, 2019. 11(4): p. 303\u0026ndash;310.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKhan, S., et al., \u003cem\u003eImmune dysregulation in cancer patients developing immune-related adverse events\u003c/em\u003e. Br J Cancer, 2019. 120(1): p. 63\u0026ndash;68.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChennamadhavuni, A., et al., \u003cem\u003eRisk Factors and Biomarkers for Immune-Related Adverse Events: A Practical Guide to Identifying High-Risk Patients and Rechallenging Immune Checkpoint Inhibitors\u003c/em\u003e. Front Immunol, 2022. 13: p. 779691.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChin, I.S., et al., \u003cem\u003eGermline genetic variation and predicting immune checkpoint inhibitor induced toxicity\u003c/em\u003e. NPJ Genom Med, 2022. 7(1): p. 73.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJing, Y., et al., \u003cem\u003eMulti-omics prediction of immune-related adverse events during checkpoint immunotherapy\u003c/em\u003e. Nat Commun, 2020. 11(1): p. 4946.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLes, I., et al., \u003cem\u003ePredictive Biomarkers for Checkpoint Inhibitor Immune-Related Adverse Events\u003c/em\u003e. Cancers (Basel), 2023. 15(5).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNorthrop, A., et al., \u003cem\u003eImplementation and impact of an electronic patient reported outcomes system in a phase II multi-site adaptive platform clinical trial for early-stage breast cancer\u003c/em\u003e. J Am Med Inform Assoc, 2025. 32(1): p. 172\u0026ndash;180.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"breast cancer, adverse events, immunotherapy, hypothyroidism, adrenal insufficiency, hypophysitis","lastPublishedDoi":"10.21203/rs.3.rs-8689063/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8689063/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e \u003cp\u003eImmune checkpoint inhibitors can result in serious, long-lasting immune-related adverse events (irAEs). Early identification of symptoms predictive of irAEs could enhance monitoring and timely intervention. This study assessed whether symptoms within the first 8 weeks of treatment could predict subsequent development of immune-related adrenal insufficiency(AI) or hypothyroidism.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003eThis retrospective cohort study analyzed prospectively collected data from the I-SPY2 trial, a phase 2 platform trial for high-risk stage II/III breast cancer across 30 U.S. sites. The cohort included 482 women treated with experimental immunotherapy agents concurrent with weekly paclitaxel neoadjuvant chemotherapy. The primary outcomes were grade\u0026thinsp;\u0026ge;\u0026thinsp;1 hypothyroidism or AI, adjudicated by an independent safety group, up to 1-year post-treatment. Symptoms and irAEs were assessed using the Common Terminology Criteria for Adverse Events. Symptom burden was quantified as area under the curve (AUC) based on symptom grade and duration. Predictive modeling was performed using logistic regression and ROC analysis; symptom enrichment between cases and controls was evaluated using Fisher\u0026rsquo;s exact tests.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eAmong 482 participants, 107 (22.2%) developed irAEs, with hypothyroidism (n\u0026thinsp;=\u0026thinsp;61, 12.7%) occurring more frequently than AI (n\u0026thinsp;=\u0026thinsp;38, 7.9%) at medians of 99 and 105 days from treatment initiation, respectively. Symptom enrichment analysis identified early predictive symptoms. Fatigue (17.2% vs 6.8%, p\u0026thinsp;=\u0026thinsp;0.011) and rash (20.7% vs 7.8%, p\u0026thinsp;=\u0026thinsp;0.0037) were predictive of hypothyroidism, while diarrhea (45.9% vs 31%, p\u0026thinsp;=\u0026thinsp;0.048), constipation (5.4% vs 0.2%, p\u0026thinsp;=\u0026thinsp;0.018), and taste changes (5.4% vs 0.5%, p\u0026thinsp;=\u0026thinsp;0.034) were associated with AI. A predictive model demonstrated moderate performance (AUC 0.65 for AI, p\u0026thinsp;\u0026lt;\u0026thinsp;0.0001; AUC 0.61 for hypothyroidism, p\u0026thinsp;=\u0026thinsp;0.012). Model accuracy in an external validation cohort was 72.8% for AI and 74.7% for hypothyroidism.\u003c/p\u003e\u003ch2\u003eConclusions\u003c/h2\u003e \u003cp\u003eThis study presents a predictive framework to identify patients at risk for adrenal insufficiency and hypothyroidism as irAEs, enabling personalized care and proactive intervention to improve treatment outcomes and safety.\u003c/p\u003e","manuscriptTitle":"Identification of Early Symptoms Associated with Subsequent Immune-related Adverse Events in the I-SPY clinical trial","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-02-11 05:38:45","doi":"10.21203/rs.3.rs-8689063/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"d75708ed-7028-460f-b62f-cdae57f373a6","owner":[],"postedDate":"February 11th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":62545505,"name":"Biological sciences/Cancer"},{"id":62545506,"name":"Health sciences/Diseases"},{"id":62545507,"name":"Health sciences/Gastroenterology"},{"id":62545508,"name":"Health sciences/Medical research"},{"id":62545509,"name":"Health sciences/Oncology"}],"tags":[],"updatedAt":"2026-05-11T02:11:33+00:00","versionOfRecord":[],"versionCreatedAt":"2026-02-11 05:38:45","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8689063","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8689063","identity":"rs-8689063","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}