2667 Background: ICIs may cause severe skin toxicity associated with significant patient (pt) impact. Published data suggests Asian pts might be at higher risk of skin immune-related adverse events (SirAEs). Variation in human leukocyte antigen (HLA) is known to predispose to autoimmune (AI) conditions, but there is limited data to understand genetic drivers of severe SirAEs. Methods: The incidence of severe SirAEs was correlated with 30 HLA alleles of interest in 2 cohorts totaling 18 Asian pts (Dermatitis bullous n=4, Erythema multiforme n=10, Stevens-Johnson syndrome n=3, Toxic skin eruption n=1) treated with atezolizumab (A) across tumor types as monotherapy or as part of combinations- 8 Japanese pts received A in a non-interventional study (trial ID: UMIN000048702) and 10 Asian pts were enrolled in Roche trials with A. Data from the 18 cases was compared to 3 different controls of Asian pts identified from Roche trials with A (i.e. pts without any irAEs n=148, pts without SirAEs n=225, pts without severe SirAEs n=390). For each HLA allele, positive predictive value (PPV), negative predictive value (NPV), sensitivity, specificity, (unadjusted) odds ratio (OR) and corresponding 95%-confidence interval (CI) were evaluated. Case-control populations were obtained by risk-set sampling with exact matching on indication, treatment arm and sex, and used in age-adjusted conditional logistic regression models to assess associations between each HLA allele and severe SirAEs. Results: There was an association between some HLA alleles and severe SirAEs, but sensitivity was low (<30%) and 1-NPV was only slightly smaller than background prevalence. Results for five HLA alleles with OR>1.5 in either the unadjusted or adjusted analysis compared to controls without severe SirAEs are presented in table. Results were consistent across the three control populations. Conclusions: Five HLA alleles previously reported in AI disorders were associated with severe SirAEs in Asian pts receiving A. Although the effect size is insufficient to be considered clinically relevant, further research is warranted to better characterize the pt-level drivers of severe SirAEs in Asians pts. Clinical trial information: UMIN000048702 . [Table: see text]

<p>Differential gene expression (adj. p<0.01, FC>1.5) for GPAM-1 and GPAM-2 siRNA oligos with lists for up- and downregulated probe sets and genes for the individual oligos, as well as the overlapping probe sets and genes.</p>

Full description of all methods used, legends for the supplementary figures, and references for the supplementary.

Background IMvigor010 (NCT02450331) showed that atezolizumab reduced risk of death by approximately 40% in post-cystectomy ctDNA-positive patients.1 Whether ctDNA may predict toxicity of immune checkpoint inhibition is unknown. Theoretically the lack of tumor-derived immune suppression associated with the presence of cancer may result in more irAEs. Methods Patients with high-risk muscle invasive bladder cancer treated with atezolizumab (n=300) vs observation (n=281) after surgical resection and with available baseline ctDNA assessment were included. The association between baseline ctDNA (either as status [positive vs negative] or as continuous MTM/mL measurement [mean tumor molecules per mL of plasma on log-scale]) and the time to onset of the first irAE of a specific type per patient was assessed in the atezolizumab arm using a Cox regression model adjusted for PD-L1 status, prior chemotherapy, race, sex and age. Plots of the cumulative incidence function (CIF) derived from the Aalen-Johansen estimator were used for visualization of the irAE risk over time stratified by ctDNA status, treatment arm and severity grade. Death was considered as a competing event for the safety endpoints in the CIF plots and Cox model (i.e., cause-specific approach treating death as censoring). Results The risk of low grade (Grade 1/2) irAEs was higher in ctDNA-negative patients compared with ctDNA-positive patients at baseline (ctDNA-negative vs -positive: HR 1.97 [1.29, 2.99]) and in patients with lower baseline ctDNA levels (ctDNA log-MTM/mL: HR 0.92 [0.87, 0.97]) (figures 1 and 2). Low-grade irAEs were predominantly hepatitis, hypo-/hyperthyroidism and rash. For high-grade irAEs there was no clear association with baseline ctDNA; however, the number of events was very small. No conclusions could be drawn from on-treatment ctDNA-irAE analysis given the limited safety data available. Conclusions Our exploratory analysis showed evidence of a higher risk of low-grade irAEs in ctDNA-negative vs ctDNA-positive patients treated with atezolizumab in the adjuvant setting. This supports the hypothesis of increased treatment-related toxicity in patients without cancer. It supports other ctDNA trials such as IMvigor011 (NCT04660344). Acknowledgements Funded by F Hoffmann-La Roche Ltd/Genentech Trial Registration ClinicalTrials.gov ID: NCT02450331 References Jackson-Spence F, et al. IMvigor011: a study of adjuvant atezolizumab in patients with high-risk MIBC who are ctDNA+ post-surgery. Future Oncol. 2023;19(7):509–515. Ethics Approval The trial was performed per Good Clinical Practice and the Declaration of Helsinki. Protocol approval was obtained from ethics committees or independent review boards for each study site. All patients provided written informed consent. Further details, including the study protocol, are published in the primary manuscript: Bellmunt J et al, Lancet Oncol. 22, 525–537 (2021).

Intrinsic or acquired resistance to HER2-targeted therapy is often a problem when small molecule tyrosine kinase inhibitors or antibodies are used to treat patients with HER2 positive breast cancer. Therefore, the identification of new targets and therapies for this patient group is warranted. Activated choline metabolism, characterized by elevated levels of choline-containing compounds, has been previously reported in breast cancer. The glycerophosphodiesterase EDI3 (GPCPD1), which hydrolyses glycerophosphocholine to choline and glycerol-3-phosphate, directly influences choline and phospholipid metabolism, and has been linked to cancer-relevant phenotypes in vitro. While the importance of choline metabolism has been addressed in breast cancer, the role of EDI3 in this cancer type has not been explored. EDI3 mRNA and protein expression in human breast cancer tissue were investigated using publicly-available Affymetrix gene expression microarray datasets (n = 540) and with immunohistochemistry on a tissue microarray (n = 265), respectively. A panel of breast cancer cell lines of different molecular subtypes were used to investigate expression and activity of EDI3 in vitro. To determine whether EDI3 expression is regulated by HER2 signalling, the effect of pharmacological inhibition and siRNA silencing of HER2, as well as the influence of inhibiting key components of signalling cascades downstream of HER2 were studied. Finally, the influence of silencing and pharmacologically inhibiting EDI3 on viability was investigated in vitro and on tumour growth in vivo. In the present study, we show that EDI3 expression is highest in ER-HER2 + human breast tumours, and both expression and activity were also highest in ER-HER2 + breast cancer cell lines. Silencing HER2 using siRNA, as well as inhibiting HER2 signalling with lapatinib decreased EDI3 expression. Pathways downstream of PI3K/Akt/mTOR and GSK3β, and transcription factors, including HIF1α, CREB and STAT3 were identified as relevant in regulating EDI3 expression. Silencing EDI3 preferentially decreased cell viability in the ER-HER2 + cells. Furthermore, silencing or pharmacologically inhibiting EDI3 using dipyridamole in ER-HER2 + cells resistant to HER2-targeted therapy decreased cell viability in vitro and tumour growth in vivo. Our results indicate that EDI3 may be a potential novel therapeutic target in patients with HER2-targeted therapy-resistant ER-HER2 + breast cancer that should be further explored.

BackgroundImmune checkpoint inhibitors (ICIs) have revolutionized the treatment of cancer patients in the last decade, but immune-related adverse events (irAEs) pose significant clinical challenges. Despite advances in the management of these unique toxicities, there remains an unmet need to further characterize the patient-level drivers of irAEs in order to optimize the benefit/risk balance in patients receiving cancer immunotherapy.MethodsAn individual-patient data post-hoc meta-analysis was performed using data from 10,344 patients across 15 Roche sponsored clinical trials with atezolizumab in five different solid tumor types to assess the association between baseline risk factors and the time to onset of irAE. In this study, the overall analysis was conducted by treatment arm, indication, toxicity grade and irAE type, and the study design considered confounder adjustment to assess potential differences in risk factor profiles.ResultsThis analysis demonstrates that the safety profile of atezolizumab is generally consistent across indications in the 15 studies evaluated. In addition, our findings corroborate with prior reviews which suggest that reported rates of irAEs with PD-(L)1 inhibitors are nominally lower than CTLA-4 inhibitors. In our analysis, there were no remarkable differences in the distribution of toxicity grades between indications, but some indication-specific differences regarding the type of irAE were seen across treatment arms, where pneumonitis mainly occurred in lung cancer, and hypothyroidism and rash had a higher prevalence in advanced renal cell carcinoma compared to all other indications. Results showed consistency of risk factors across indications and by toxicity grade. The strongest and most consistent risk factors were mostly organ-specific such as elevated liver enzymes for hepatitis and thyroid stimulating hormone (TSH) for thyroid toxicities. Another strong but non-organ-specific risk factor was ethnicity, which was associated with rash, hepatitis and pneumonitis. Further understanding the impact of ethnicity on ICI associated irAEs is considered as an area for future research.ConclusionsOverall, this analysis demonstrated that atezolizumab safety profile is consistent across indications, is clinically distinguishable from comparator regimens without checkpoint inhibition, and in line with literature, seems to suggest a nominally lower reported rates of irAEs vs CTLA-4 inhibitors. This analysis demonstrates several risk factors for irAEs by indication, severity and location of irAE, and by patient ethnicity. Additionally, several potential irAE risk factors that have been published to date, such as demographic factors, liver enzymes, TSH and blood cell counts, are assessed in this large-scale meta-analysis, providing a more consistent picture of their relevance. However, given the small effects size, changes to clinical management of irAEs associated with the use of Anti-PDL1 therapy are not warranted.

Proteasome inhibition is associated with parkinsonian pathology in vivo and degeneration of dopaminergic neurons in vitro. We explored here the metabolome (386 metabolites) and transcriptome (3257 transcripts) regulations of human LUHMES neurons, following exposure to MG-132 [100 nM]. This proteasome inhibitor killed cells within 24 h but did not reduce viability for 12 h. Overall, 206 metabolites were changed in live neurons. The early (3 h) metabolome changes suggested a compromised energy metabolism. For instance, AMP, NADH and lactate were up-regulated, while glycolytic and citric acid cycle intermediates were down-regulated. At later time points, glutathione-related metabolites were up-regulated, most likely by an early oxidative stress response and activation of NRF2/ATF4 target genes. The transcriptome pattern confirmed proteostatic stress (fast up-regulation of proteasome subunits) and also suggested the progressive activation of additional stress response pathways. The early ones (e.g., HIF-1, NF-kB, HSF-1) can be considered a cytoprotective cellular counter-regulation, which maintained cell viability. For instance, a very strong up-regulation of AIFM2 (=FSP1) may have prevented fast ferroptotic death. For most of the initial period, a definite life-death decision was not taken, as neurons could be rescued for at least 10 h after the start of proteasome inhibition. Late responses involved p53 activation and catabolic processes such as a loss of pyrimidine synthesis intermediates. We interpret this as a phase of co-occurrence of protective and maladaptive cellular changes. Altogether, this combined metabolomics-transcriptomics analysis informs on responses triggered in neurons by proteasome dysfunction that may be targeted by novel therapeutic intervention in Parkinson's disease.

Lists of overlapping up- and downregulated probe sets and genes that were differentially expressed after GPAM knockdown with both oligos were compared to those differentially expressed after the knockdown of EDI3.

<p>Supplementary table 1 - Full list of all siRNA and shRNA used with catalog numbers. Supplementary table 2 - List on names and catalog numbers of Qiagen primer assays used in study Supplementary table 3 - Mass spectrometry settings and the fragment m/z data. Supplementary table 4 - Table of clinicopathological characteristics for all datasets of (A) ovarian (B) breast (C) colon and (D) lung cancer patients. Supplementary table 5 - Concentration of metabolites in MCF7 cells transfected with scrambled siRNA, EDI3, GPAM and CHKA siRNA and measured with NMR Supplementary table 8 - List of overlapping up- and down-regulated genes after silencing GPAM and EDI3 in MCF7 ranked according to adjusted p values (adj. p<0.01). Supplementary table 9 - Association of GPAM mRNA expression with metastasis-free survival (MFS), disease-free survival (DFS) or overall survival (OS) in different cancer types.</p>

Supplementary Fig. S1: Confirmation of the effect of GPAM and EDI3 on migration in HeLa cells, related to Fig.1. Supplementary Fig. S2: Confirmation that CHKA has no effect on viability and migration, related to Fig. 2. Supplementary Fig. S3: No decreased viability or off-target effect after GPAM silencing, related to Fig. 3. Supplementary Fig. S4: Confirmation of GPAM effect on migration with multiple siRNA oligos (HeLa cells), related to Fig. 3. Supplementary Fig. S5: Confirmation of GPAM effect on migration with two siRNA oligos (MDA-MB-231 cells), related to Fig. 3. Supplementary Fig. S6: Evidence that transfection accelerates the uptake of fluorescently-labeled LPA in cells, related to Fig. 5. Supplementary Fig. S7: Common gene expression changes after EDI3 and GPAM KD, related to Fig. 4 and 5. Supplementary Fig. S8: GPAM expression in ovarian cancer cohorts - 2nd probeset and dichotomizations, related to Fig. 6. Supplementary Fig. S9: GPAM silencing decreases migration and tumor growth - 2nd set of oligos and images of xenograft tumors, related to Fig. 7.

Published classifiers for prediction of pCR (a), study centers that contributed to the EXPRESSION trial (b), baseline characteristics of patients in the EXPRESSION trial (c), publicly available breast cancer datasets (d), comparison of patients who achieved versus not achieved a pCR (e), association of clinicopathologic parameters with pCR (f)

BACKGROUND: PIK3CA mutations have been shown to be associated with poor prognosis in HER2-positive breast cancer (BC). We combined data from three completed Phase III Roche-sponsored randomized trials of HER2-targeted therapy for the first-line treatment of HER2-positive metastatic BC (MBC); this allowed for exploration of the prognostic impact of PIK3CA mutations observed in the three individual trials across subgroups of interest. METHODS: Data from CLEOPATRA (pertuzumab + trastuzumab + docetaxel [PHD] vs. placebo [Pla] + HD; NCT00567190; N = 808), MARIANNE (HD vs. ado-trastuzumab emtansine [K] + Pla vs. K + P; NCT01120184; N = 1095), and PUFFIN (PHD vs. Pla + HD; NCT02896855; N = 243) were included. An individual patient data (IPD) meta-analysis was performed to test the association between PIK3CA mutation status in tumor tissue (mutated vs. wild type [WT]) and efficacy (progression-free and overall survival [PFS/OS]) in different biomarker and clinical subgroups. Confounder adjustment was conducted for age, Eastern Cooperative Oncology Group Performance Status, body mass index, treatment, disease type, and number of metastases (all at baseline). “Study” was included as a random effect in the IPD meta-analysis model to account for variability between studies. A landmark analysis was conducted on fast and non-fast progressors (cutoff of >137 days [i.e., after six chemotherapy cycles]) from CLEOPATRA and PUFFIN only, since they include the current standard-of-care regimens (PHD), by using Day 137 as the landmark time with separate Cox proportional hazards models. RESULTS: PIK3CA mutation data were available for 1905/2146 patients (89%; ~80% from primary tissue); mutation prevalence was 27% (n = 521). PIK3CA-mutated vs. WT in association with PFS in pooled treatment arms is shown in the table. OS data were consistent. CONCLUSIONS: PIK3CA mutations were associated with a worse prognosis across subgroups of interest, including in fast and non-fast progressors, in the two PHD-containing studies as compared with the overall ITT population. Table: PIK3CA-mutated vs. WT in association with PFS in pooled treatment arms Citation Format: Sandra Swain, Javier Cortés, Binghe Xu, Chiara Lambertini, Laurent Essioux, Adam Knott, Eleonora Restuccia, Katrin Madjar, Sanne Lysbet De Haas. Association of PIK3CA mutations with efficacy in HER2-positive first-line metastatic breast cancer: a meta-analysis [abstract]. In: Proceedings of the 2022 San Antonio Breast Cancer Symposium; 2022 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2023;83(5 Suppl):Abstract nr P2-11-07.

Classifier performance in the subgroup of ER+/HER2-patients

Despite the progress made in developmental toxicology, there is a great need for in vitro tests that identify developmental toxicants in relation to human oral doses and blood concentrations. In the present study, we established the hiPSC-based UKK2 in vitro test and analyzed genome-wide expression profiles of 23 known teratogens and 16 non-teratogens. Compounds were analyzed at the maximal plasma concentration (Cmax) and at 20-fold Cmax for a 24 h incubation period in three independent experiments. Based on the 1000 probe sets with the highest variance and including information on cytotoxicity, penalized logistic regression with leave-one-out cross-validation was used to classify the compounds as test-positive or test-negative, reaching an area under the curve (AUC), accuracy, sensitivity, and specificity of 0.96, 0.92, 0.96, and 0.88, respectively. Omitting the cytotoxicity information reduced the test performance to an AUC of 0.94, an accuracy of 0.79, and a sensitivity of 0.74. A second method, which used the number of significantly deregulated probe sets to classify the compounds, resulted in a specificity of 1; however, the AUC (0.90), accuracy (0.90), and sensitivity (0.83) were inferior compared to those of the logistic regression-based procedure. Finally, no increased performance was achieved when the high test concentrations (20-fold Cmax) were used, in comparison to testing within the realistic clinical range (1-fold Cmax). In conclusion, although further optimization is required, for example, by including additional readouts and cell systems that model different developmental processes, the UKK2-test in its present form can support the early discovery-phase detection of human developmental toxicants.

Human-relevant tests to predict developmental toxicity are urgently needed. A currently intensively studied approach makes use of differentiating human stem cells to measure chemically-induced deviations of the normal developmental program, as in a recent study based on cardiac differentiation (UKK2). Here, we (i) tested the performance of an assay modeling neuroepithelial differentiation (UKN1), and (ii) explored the benefit of combining assays (UKN1 and UKK2) that model different germ layers. Substance-induced cytotoxicity and genome-wide expression profiles of 23 teratogens and 16 non-teratogens at human-relevant concentrations were generated and used for statistical classification, resulting in accuracies of the UKN1 assay of 87–90%. A comparison to the UKK2 assay (accuracies of 90–92%) showed, in general, a high congruence in compound classification that may be explained by the fact that there was a high overlap of signaling pathways. Finally, the combination of both assays improved the prediction compared to each test alone, and reached accuracies of 92–95%. Although some compounds were misclassified by the individual tests, we conclude that UKN1 and UKK2 can be used for a reliable detection of teratogens in vitro, and that a combined analysis of tests that differentiate hiPSCs into different germ layers and cell types can even further improve the prediction of developmental toxicants.

An important task in clinical medicine is the construction of risk prediction models for specific subgroups of patients based on high-dimensional molecular measurements such as gene expression data. Major objectives in modeling high-dimensional data are good prediction performance and feature selection to find a subset of predictors that are truly associated with a clinical outcome such as a time-to-event endpoint. In clinical practice, this task is challenging since patient cohorts are typically small and can be heterogeneous with regard to their relationship between predictors and outcome. When data of several subgroups of patients with the same or similar disease are available, it is tempting to combine them to increase sample size, such as in multicenter studies. However, heterogeneity between subgroups can lead to biased results and subgroup-specific effects may remain undetected. For this situation, we propose a penalized Cox regression model with a weighted version of the Cox partial likelihood that includes patients of all subgroups but assigns them individual weights based on their subgroup affiliation. The weights are estimated from the data such that patients who are likely to belong to the subgroup of interest obtain higher weights in the subgroup-specific model. Our proposed approach is evaluated through simulations and application to real lung cancer cohorts, and compared to existing approaches. Simulation results demonstrate that our proposed model is superior to standard approaches in terms of prediction performance and variable selection accuracy when the sample size is small. The results suggest that sharing information between subgroups by incorporating appropriate weights into the likelihood can increase power to identify the prognostic covariates and improve risk prediction.

BACKGROUND:Important objectives in cancer research are the prediction of a patient's risk based on molecular measurements such as gene expression data and the identification of new prognostic biomarkers (e.g. genes). In clinical practice, this is often challenging because patient cohorts are typically small and can be heterogeneous. In classical subgroup analysis, a separate prediction model is fitted using only the data of one specific cohort. However, this can lead to a loss of power when the sample size is small. Simple pooling of all cohorts, on the other hand, can lead to biased results, especially when the cohorts are heterogeneous.RESULTS:We propose a new Bayesian approach suitable for continuous molecular measurements and survival outcome that identifies the important predictors and provides a separate risk prediction model for each cohort. It allows sharing information between cohorts to increase power by assuming a graph linking predictors within and across different cohorts. The graph helps to identify pathways of functionally related genes and genes that are simultaneously prognostic in different cohorts.CONCLUSIONS:Results demonstrate that our proposed approach is superior to the standard approaches in terms of prediction performance and increased power in variable selection when the sample size is small.

AbstractPurpose: Expression-based classifiers to predict pathologic complete response (pCR) after neoadjuvant chemotherapy (NACT) are not routinely used in the clinic. We aimed to build and validate a classifier for pCR after NACT. Patients and Methods: We performed a prospective multicenter study (EXPRESSION) including 114 patients treated with anthracycline/taxane-based NACT. Pretreatment core needle biopsies from 91 patients were used for gene expression analysis and classifier construction, followed by validation in five external cohorts (n = 619). Results: A 20-gene classifier established in the EXPRESSION cohort using a Youden index–based cut-off point predicted pCR in the validation cohorts with an accuracy, AUC, negative predictive value (NPV), positive predictive value, sensitivity, and specificity of 0.811, 0.768, 0.829, 0.587, 0.216, and 0.962, respectively. Alternatively, aiming for a high NPV by defining the cut-off point for classification based on the complete responder with the lowest predicted probability of pCR in the EXPRESSION cohort led to an NPV of 0.960 upon external validation. With this extreme-low cut-off point, a recommendation to not treat with anthracycline/taxane-based NACT would be possible for 121 of 619 unselected patients (19.5%) and 112 of 322 patients with luminal breast cancer (34.8%). The analysis of the molecular subtypes showed that the identification of patients who do not achieve a pCR by the 20-gene classifier was particularly relevant in luminal breast cancer. Conclusions: The novel 20-gene classifier reliably identifies patients who do not achieve a pCR in about one third of luminal breast cancers in both the EXPRESSION and combined validation cohorts.

Motivation To obtain a reliable prediction model for a specific cancer subgroup or cohort is often difficult due to limited sample size and, in survival analysis, due to potentially high censoring rates. Sometimes similar data from other patient subgroups are available, e.g. from other clinical centers. Simple pooling of all subgroups can decrease the variance of the predicted parameters of the prediction models, but also increase the bias due to heterogeneity between the cohorts. A promising compromise is to identify those subgroups with a similar relationship between covariates and target variable and then include only these for model building. Results We propose a subgroup-based weighted likelihood approach for survival prediction with high-dimensional genetic covariates. When predicting survival for a specific subgroup, for every other subgroup an individual weight determines the strength with which its observations enter into model building. MBO (model-based optimization) can be used to quickly find a good prediction model in the presence of a large number of hyperparameters. We use MBO to identify the best model for survival prediction of a specific subgroup by optimizing the weights for additional subgroups for a Cox model. The approach is evaluated on a set of lung cancer cohorts with gene expression measurements. The resulting models have competitive prediction quality, and they reflect the similarity of the corresponding cancer subgroups, with both weights close to 0 and close to 1 and medium weights. Availability and implementation mlrMBO is implemented as an R-package and is freely available at http://github.com/mlr-org/mlrMBO.