Prediction Of Response Research Articles

Genomic fingerprints of the world's soil ecosystems.

Despite the explosion of soil metagenomic data, we lack a synthesized understanding of patterns in the distribution and functions of soil microorganisms. These patterns are critical to predictions of soil microbiome responses to climate change and resulting feedbacks that regulate greenhouse gas release from soils. To address this gap, we assay 1,512 manually curated soil metagenomes using complementary annotation databases, read-based taxonomy, and machine learning to extract multidimensional genomic fingerprints of global soil microbiomes. Our objective is to uncover novel biogeographical patterns of soil microbiomes across environmental factors and ecological biomes with high molecular resolution. We reveal shifts in the potential for (i) microbial nutrient acquisition across pH gradients; (ii) stress-, transport-, and redox-based processes across changes in soil bulk density; and (iii) greenhouse gas emissions across biomes. We also use an unsupervised approach to reveal a collection of soils with distinct genomic signatures, characterized by coordinated changes in soil organic carbon, nitrogen, and cation exchange capacity and in bulk density and clay content that may ultimately reflect soil environments with high microbial activity. Genomic fingerprints for these soils highlight the importance of resource scavenging, plant-microbe interactions, fungi, and heterotrophic metabolisms. Across all analyses, we observed phylogenetic coherence in soil microbiomes-more closely related microorganisms tended to move congruently in response to soil factors. Collectively, the genomic fingerprints uncovered here present a basis for global patterns in the microbial mechanisms underlying soil biogeochemistry and help beget tractable microbial reaction networks for incorporation into process-based models of soil carbon and nutrient cycling.IMPORTANCEWe address a critical gap in our understanding of soil microorganisms and their functions, which have a profound impact on our environment. We analyzed 1,512 global soils with advanced analytics to create detailed genetic profiles (fingerprints) of soil microbiomes. Our work reveals novel patterns in how microorganisms are distributed across different soil environments. For instance, we discovered shifts in microbial potential to acquire nutrients in relation to soil acidity, as well as changes in stress responses and potential greenhouse gas emissions linked to soil structure. We also identified soils with putative high activity that had unique genomic characteristics surrounding resource acquisition, plant-microbe interactions, and fungal activity. Finally, we observed that closely related microorganisms tend to respond in similar ways to changes in their surroundings. Our work is a significant step toward comprehending the intricate world of soil microorganisms and its role in the global climate.

Global convergence in terrestrial gross primary production response to atmospheric vapor pressure deficit.

Atmospheric vapor pressure deficit (VPD) increases with climate warming and may limit plant growth. However, gross primary production (GPP) responses to VPD remain a mystery, offering a significant source of uncertainty in the estimation of global terrestrial ecosystems carbon dynamics. In this study, in-situ measurements, satellite-derived data, and Earth System Models (ESMs) simulations were analysed to show that the GPP of most ecosystems has a similar threshold in response to VPD: first increasing and then declining. When VPD exceeds these thresholds, atmospheric drought stress reduces soil moisture and stomatal conductance, thereby decreasing the productivity of terrestrial ecosystems. Current ESMs underscore CO2 fertilization effects but predict significant GPP decline in low-latitude ecosystems when VPD exceeds the thresholds. These results emphasize the impacts of climate warming on VPD and propose limitations to future ecosystems productivity caused by increased atmospheric water demand. Incorporating VPD, soil moisture, and canopy conductance interactions into ESMs enhances the prediction of terrestrial ecosystem responses to climate change.

Prediction of anticancer drug sensitivity using an interpretable model guided by deep learning

BackgroundThe prediction of drug sensitivity plays a crucial role in improving the therapeutic effect of drugs. However, testing the effectiveness of drugs is challenging due to the complex mechanism of drug reactions and the lack of interpretability in most machine learning and deep learning methods. Therefore, it is imperative to establish an interpretable model that receives various cell line and drug feature data to learn drug response mechanisms and achieve stable predictions between available datasets.ResultsThis study proposes a new and interpretable deep learning model, DrugGene, which integrates gene expression, gene mutation, gene copy number variation of cancer cells, and chemical characteristics of anticancer drugs to predict their sensitivity. This model comprises two different branches of neural networks, where the first involves a hierarchical structure of biological subsystems that uses the biological processes of human cells to form a visual neural network (VNN) and an interpretable deep neural network for human cancer cells. DrugGene receives genotype input from the cell line and detects changes in the subsystem states. We also employ a traditional artificial neural network (ANN) to capture the chemical structural features of drugs. DrugGene generates final drug response predictions by combining VNN and ANN and integrating their outputs into a fully connected layer. The experimental results using drug sensitivity data extracted from the Cancer Drug Sensitivity Genome Database and the Cancer Treatment Response Portal v2 reveal that the proposed model is better than existing prediction methods. Therefore, our model achieves higher accuracy, learns the reaction mechanisms between anticancer drugs and cell lines from various features, and interprets the model’s predicted results.ConclusionsOur method utilizes biological pathways to construct neural networks, which can use genotypes to monitor changes in the state of network subsystems, thereby interpreting the prediction results in the model and achieving satisfactory prediction accuracy. This will help explore new directions in cancer treatment. More available code resources can be downloaded for free from GitHub (https://github.com/pangweixiong/DrugGene).

Interpretable machine learning for selection of site-specific soil constitutive models and consolidation settlement analysis

Machine learning (ML) has achieved great breakthroughs in many fields, such as computer vision and speech recognition, and it also gains an increasing attention in geotechnical engineering. However, it is often criticized for a lack of interpretability. This study proposes an interpretable sparse dictionary learning (ISDL) approach for selection of suitable soil constitutive models (SCMs) and geotechnical consolidation analysis in a specific site. ISDL is inherently interpretable since it expresses prediction of geotechnical responses (e.g., consolidation settlement) as a weighted summation of many elementary datasets (i.e., dictionary atoms in ISDL). The dictionary atoms are constructed using numerical analyses (e.g., finite element model) with different candidate SCMs and parameters. Interpretability of ISDL is improved by a small number of non-trivial atoms (or SCMs) selected for model prediction using site-specific monitoring data and further enhanced by contribution ranking of the selected SCMs using a game theory-based approach. The proposed method is illustrated using a real reclamation project in Hong Kong, and it is shown to effectively select the most suitable constitutive models for the given site, quantify contributions of the selected models, and significantly improve settlement predictions with rigorously quantified prediction uncertainty, particularly at locations without monitoring or future time steps.

Prediction and mediation analysis for treatment responses to combined cognitive and physical training for older adults

Diminished cognitive and physical functions negatively affect the daily functions of individuals. Although combined cognitive and physical training prevents instrumental activities of daily living (IADL) disability in older adults, no predictive model or mediation analysis of IADL after combined training exists. This study aims to employ prediction and mediation analysis to identify the predictors of IADL performance and to elucidate the mediators of the association between baseline global cognition and subsequent IADL performance following combined cognitive and physical training. This study involved 177 participants aged 60 years and older who underwent combined training. Cognitive function was measured with the Montreal Cognitive Assessment (MoCA), Digit Symbol Substitution Test (DSST), Color Trails Test, Word List, and a dual task; physical function with the Timed Up and Go (TUG) test; daily function with the Lawton IADL Scale. We conducted regression analyses to identify the predictors of IADL performance, and mediation analysis to examine whether DSST and TUG mediate the relationship between MoCA and IADL. The pre-intervention DSST and TUG were significant independent predictors of post-intervention IADL. The association between the pre-intervention MoCA and post-intervention IADL was mediated by pre-intervention DSST and TUG. This study highlighted the importance of measuring and improving processing speed and functional mobility before and during interventions to enhance IADL outcomes.Trial registration: NCT03619577, 23/07/2018; NCT04689776, 29/12/2020.

Broadening the horizon of immune thrombocytopenia through Omics approaches.

Immune thrombocytopenia (ITP) is a highly heterogeneous autoimmune bleeding disorder characterized by low platelet counts due to an immune-mediated platelet destruction and impaired platelet production. The pathophysiology is multifactorial and remains to be fully unravelled. Consequently, disease trajectories and responses to therapeutics, despite the availability of multiple agents, can be unpredictable and differing between patients. There is an urgent need for the identification of diagnostic and therapeutic biomarkers, but this has proven to be challenging to achieve. To shed light on this, two studies in this issue of the British Journal of Haematology have recognized the opportunity of using high-throughput Omics technologies in ITP. Sun etal. performed proteomics, and Li etal. metabolomics, on bone marrow biopsy samples of patients with ITP. This was conducted using mass spectrometry and, due to the generation of large datasets, in combination with machine learning. These studies set the stage for further investigations exploring the high potential of multi-omics technologies in order to shed light on the heterogeneity in ITP, accelerating the path towards a much needed personalized medicine approach. Commentary on: Li etal. Metabolomics profile and machine learning prediction of treatment responses in immune thrombocytopenia: A prospective cohort study. Br J Haematol 2024 (Online ahead of print). doi: 10.1111/bjh.19391 Commentary on: Sun etal. Proteomics landscape and machine learning prediction of long-term response to splenectomy in primary immune thrombocytopenia. Br J Haematol 2024 (Online ahead of print). doi: 10.1111/bjh.19420.

Examining the role of AI technology in online mental healthcare: opportunities, challenges, and implications, a mixed-methods review

IntroductionOnline mental healthcare has gained significant attention due to its effectiveness, accessibility, and scalability in the management of mental health symptoms. Despite these advantages over traditional in-person formats, including higher availability and accessibility, issues with low treatment adherence and high dropout rates persist. Artificial intelligence (AI) technologies could help address these issues, through powerful predictive models, language analysis, and intelligent dialogue with users, however the study of these applications remains underexplored. The following mixed methods review aimed to supplement this gap by synthesizing the available evidence on the applications of AI in online mental healthcare.MethodWe searched the following databases: MEDLINE, CINAHL, PsycINFO, EMBASE, and Cochrane. This review included peer-reviewed randomized controlled trials, observational studies, non-randomized experimental studies, and case studies that were selected using the PRISMA guidelines. Data regarding pre and post-intervention outcomes and AI applications were extracted and analyzed. A mixed-methods approach encompassing meta-analysis and network meta-analysis was used to analyze pre and post-intervention outcomes, including main effects, depression, anxiety, and study dropouts. We applied the Cochrane risk of bias tool and the Grading of Recommendations Assessment, Development and Evaluation (GRADE) to assess the quality of the evidence.ResultsTwenty-nine studies were included revealing a variety of AI applications including triage, psychotherapy delivery, treatment monitoring, therapy engagement support, identification of effective therapy features, and prediction of treatment response, dropout, and adherence. AI-delivered self-guided interventions demonstrated medium to large effects on managing mental health symptoms, with dropout rates comparable to non-AI interventions. The quality of the data was low to very low.DiscussionThe review supported the use of AI in enhancing treatment response, adherence, and improvements in online mental healthcare. Nevertheless, given the low quality of the available evidence, this study highlighted the need for additional robust and high-powered studies in this emerging field.Systematic review registrationhttps://www.crd.york.ac.uk/prospero/display_record.php?RecordID=443575, identifier CRD42023443575.

Predicting response to neoadjuvant chemotherapy for colorectal liver metastasis using deep learning on prechemotherapy cross-sectional imaging.

Deep learning models (DLMs) are applied across domains of health sciences to generate meaningful predictions. DLMs make use of neural networks to generate predictions from discrete data inputs. This study employs DLM on prechemotherapy cross-sectional imaging to predict patients' response to neoadjuvant chemotherapy. Adult patients with colorectal liver metastasis who underwent surgery after neoadjuvant chemotherapy were included. A DLM was trained on computed tomography images using attention-based multiple-instance learning. A logistic regression model incorporating clinical parameters of the Fong clinical risk score was used for comparison. Both model performances were benchmarked against the Response Evaluation Criteria in Solid Tumors criteria. A receiver operating curve was created and resulting area under the curve (AUC) was determined. Ninety-five patients were included, with 33,619 images available for study inclusion. Ninety-five percent of patients underwent 5-fluorouracil-based chemotherapy with oxaliplatin and/or irinotecan. Sixty percent of the patients were categorized as chemotherapy responders (30% reduction in tumor diameter). The DLM had an AUC of 0.77. The AUC for the clinical model was 0.41. Image-based DLM for prediction of response to neoadjuvant chemotherapy in patients with colorectal cancer liver metastases was superior to a clinical-based model. These results demonstrate potential to identify nonresponders to chemotherapy and guide select patients toward earlier curative resection.

An Adaptive Contextual Relation Model for Improving Response Generation

An Adaptive Contextual Relation Model for Improving Response Generation

A road surface reconstruction dataset for autonomous driving

Recent developments in intelligent robot systems, especially autonomous vehicles, put forward higher requirements for safety and comfort. Road conditions are crucial factors affecting the comprehensive performance of ground vehicles. Nonetheless, existing environment perception datasets for autonomous driving lack attention to road surface areas. In this paper, we introduce the road surface reconstruction dataset, providing multi-modal, high-resolution, and high-precision data collected by real-vehicle platform in diverse driving conditions. It covers common road types containing approximately 16,000 pairs of stereo images, point clouds, and ground-truth depth/disparity maps, with accurate data processing pipelines to ensure its quality. Preliminary evaluations reveal the effectiveness of our dataset and the challenge of the task, underscoring substantial opportunities of it as a valuable resource for advancing computer vision techniques. The reconstructed road structure and texture contribute to the analysis and prediction of vehicle responses for motion planning and control systems.

Volumetric parameters of 68[Ga]Ga-DOTA-TATE PET/CT in the prediction of response to treatment with long-acting somatostatin analogues in patients with well-differentiated NETs

Volumetric parameters of 68[Ga]Ga-DOTA-TATE PET/CT in the prediction of response to treatment with long-acting somatostatin analogues in patients with well-differentiated NETs

Clinical Response Predictive Model for Omalizumab in Moderate-to-Severe Asthma Patients

Objective. Our study aimed to develop a predictive model for evaluating the clinical response of omalizumab treatment in moderate-to-severe asthma patients. Methods. This single-center, prospective study collected patients who meet the diagnostic criteria for moderate-to-severe bronchial asthma set by the National Asthma Prevention and Treatment Group in 2016 in the first hospital affiliated with Soochow University. Patients recruited were treated with omalizumab once per four weeks; at the beginning of each injection, blood eosinophils and the level of total serum IgE (IU/mL) were tested. After four injections of omalizumab, asthma control test (ACT), the 15-item Mini Asthma Quality of Life Questionnaire (MiniAQLQ), global treatment effectiveness (GETE), and lung function of all patients were evaluated in the 16th week. We used the selection operator method to build a logistic model and evaluated the clinical response of omalizumab in these patients. Results. This study included 108 moderate-to-severe patients (aged 39.86 ± 14.59 years). Eighty-nine patients finished treatment for 16 weeks, and 74 patients (83.1%) had an excellent or good response. The serum level of total IgE increased significantly after injection of omalizumab, while blood eosinophils count decreased significantly from baseline. Using the GETE as a clinical outcome, several clinical variables were significant predictors of clinical response. The corrected AUC and Brier scores were 0.872 and 0.111, which showed good discrimination. Significant variables included age, weight, family allergic history, acute exacerbations, the ratio of total serum IgE level at the 4th week to the baseline level, forced expiratory volume in one second/forced vital capacity (FEV1/FVC), and commodities of rhinitis. Using the improvement in maximal expiratory flow 25% of the measured value to the predicted value (MEF25%pre) as clinical outcome, the significant variables included weight, duration of asthma, use of oral corticosteroids (OCS), total serum IgE level at the 4th week, and history of rhinitis. Its corrected AUC and Brier scores were 0.674 and 0.225 after internal validation. Conclusion. Omalizumab treatment remarkably improved asthma control and pulmonary function in Chinese patients with moderate-to-severe asthma. The response prediction model we developed provides convenient approaches to help identify better clinical response patients to omalizumab treatment.

A clinicopathological-imaging nomogram for the prediction of pathological complete response in breast cancer cases administered neoadjuvant therapy

ObjectiveTo construct a user-friendly nomogram with MRI and clinicopathological parameters for the prediction of pathological complete response (pCR) after neoadjuvant therapy (NAT) in patients with breast cancer (BC). MethodsWe retrospectively enrolled consecutive female patients pathologically confirmed with breast cancer who received NAT followed by surgery between January 2018 and December 2022 as the development cohort. Additionally, we prospectively collected eligible candidates between January 2023 and December 2023 as an external validation group at our institution. Pretreatment MRI features and clinicopathological variables were collected, and the pre- and post-treatment background parenchymal enhancement (BPE) and the changes in BPE on two MRIs were compared between patients who achieved pCR and those who did not. Multivariable logistic regression analysis was used to identify independent variables associated with pCR in the development cohort. These independent variables were combined into a predictive nomogram for which performance was assessed using the area under the receiver operating characteristic curve (AUC), calibration plot, decision curve analysis, and external validation. ResultsIn the development cohort, there were a total of 276 female patients with a mean age of 48.3 ± 8.7 years, while in the validation cohort, there were 87 female patients with a mean age of 49.0 ± 9.5 years. Independent prognostic factors of pCR included small tumor size, HER2(+), high Ki-67 index,high signal enhancement ratio (SER), low minimum value of apparent diffusion coefficient (ADCmin), and significantly decreased BPE after NAT(change of BPE). The nomogram, which incorporates the above parameters, demonstrated excellent predictive performance in both the development and external validation cohorts, with AUC values of 0.900 and 0.850, respectively. Additionally, the nomogram showed excellent calibration capacities, as indicated by Hosmer-Lemeshow test p values of 0.508 and 0.423 in the two cohorts. Furthermore, the nomogram provided greater net benefits compared to the default simple schemes in both cohorts. ConclusionA nomogram constructed using tumor size, HER2 status, Ki-67 index, SER, ADCmin, and changes in pre- and post-NAT BPE demonstrated strong predictive performance, calibration ability, and greater net benefits for predicting pCR in patients with BC after NAT. This suggests that the user-friendly nomogram could be a valuable imaging biomarker for identifying suitable candidates for NAT.

Systematic literature review to inform the EULAR recommendations for the use of imaging in crystal-induced arthropathies in clinical practice

ObjectiveTo summarise current data regarding the use of imaging in crystal-induced arthropathies (CiAs) informing a European Alliance of Associations for Rheumatology task force.MethodsWe performed four systematic searches in Embase, Medline...

Sensitivity analysis of gyratory compactor variables to mimic roller compacted concrete pavement performance

ABSTRACT The compaction parameters suggested for Superpave gyratory compactor by ASTM C1800 are primarily derived from the asphalt mix (ASTM D6925). Since the rheological behaviour of Roller Compacted Concrete Pavement (RCCP) is significantly different from asphalt mixes, conventional Superpave compaction parameters cannot be directly adopted for RCCP. Hence, the current study focuses on formulating compaction parameters for the design of RCCP to resemble the field behaviour. For this task, response surface methodology was employed to investigate the effect of gyratory compactor parameters such as internal angle (0.82–3°), compaction pressure (600–1000 kPa), and gyration levels (60–120) on the performance of RCCP specimens. The accuracy of the response prediction model was verified with ANOVA analysis. Further, multiobjective optimization was performed and validated with experimental findings. For validation of laboratory findings, a field test section was constructed using a duplex roller (dual drum vibratory roller). The results showed that the ASTM C1800 suggested gyratory parameters resulted in lower fresh density and strength properties of 4% and 7–45%, respectively, than the field specimens. However, the specimens fabricated at 1.92° internal angle, 120 gyrations, 1000 kPa, and 1 aspect ratio in the gyratory compactor could provide comparable field properties.

Abstract PO1-15-12: Circulating tumor DNA (ctDNA) for prediction of pathologic complete response (pCR) after 12 weeks of pembrolizumab + trastuzumab + pertuzumab in HER2-enriched early breast cancer: WSG-Keyriched-1 trial

Abstract Background ctDNA testing is emerging as an important biomarker in early breast cancer (eBC). However, its value in prediction of tumor response to de-escalated, chemotherapy-free neoadjuvant therapy (NAT) remains underexplored. In WSG-Keyriched-1 (NCT03988036), a single-arm phase 2 trial, we investigated for the first time a chemotherapy-free NAT with dual HER2 blockade and pembrolizumab in HER2- enriched HER2+ eBC. In this pre-specified translational analysis, we evaluated whether ctDNA measurement could predict pCR. Methods 48 patients (pts) with HER2 2+ (ISH+) or 3+ eBC (stage I-III) and HER2-enriched subtype by PAM50 were included. Pts received pembrolizumab (200 mg), trastuzumab biosimilar (loading dose (LD) 8 mg/kg bodyweight (BW), maintenance dose (MD) 6 mg/kg BW), and pertuzumab (LD 840 mg/kg BW, MD 420 mg/kg BW) q3w for 12 weeks. Primary objective was pCR (ypT0/is ypN0). ctDNA was analyzed in 92 plasma samples collected from 31 pts at baseline (BL) and week 3 of NAT and 30 pts at end of treatment (EOT). Sequencing libraries with unique molecular identifiers were constructed from cell-free DNA and hybridization panels with ≤50 patient-specific somatic mutations from tumor sequencing were used for enrichment. Libraries were sequenced to an ultra-high depth of 100,000×. Sequencing data was analyzed using a combination of public pipelines (megSAP and umiVar2 for mapping and deduplication) and custom in-house scripts (to extract the allele counts for patient-specific variants). To reduce the error rate, only corrected reads with ≥4 duplicates were used. p-value for ctDNA detection was calculated using the total variant count, depth and sample-specific error. Association between ctDNA with pCR and other clinical parameters were assessed with Chi-square or Fisher’s exact test and univariable logistic regression. Additionally, bivariable logistic regressions for pCR were performed with ctDNA and either cT, cN, or grade. Results ctDNA was detected (ctDNA+) in 58.0% (n=18/31) of pts at BL, 9.7% (n=3/31) at week 3, and 10% (n=3/30) at EOT. ctDNA was cleared by week 3 in 83.3% of pts (n=15/18). Compared with ctDNA-negative cases (ctDNA-) at BL, those ctDNA+ more often had cT2-3 disease (94.4%, n=17/18, vs 38.5%, n=5/13; p=.001) and lymph node involvement (61.1%, n=11/18, vs 0%, n=0/13; p< .001). 72.2% (n=13/18) of ctDNA+ and 61.5% (n=8/13) of ctDNA- cases at BL were grade 3 (p=.53). Each of the 3 pts who remained ctDNA+ at week 3 was node-positive; all pts remaining ctDNA- at week 3 were node-negative. pCR rate was 38.9% (n=7/18) in ctDNA+ vs 76.9% (n=10/13) in ctDNA- pts at BL (p=.067), 0% (n=0/3) in ctDNA+ vs 60.7% (n=17/28) in ctDNA- at week 3 (p=.081), and 33.3% (n=1/3) in ctDNA+ vs 59.3% (n=16/27) in ctDNA- at EOT (p=.565). pCR rate was highest in pts who remained ctDNA- throughout week 3 (76.9%, n=10/13), compared to pts with ctDNA cleared by week 3 (46.7%, n=7/15), and pts remaining ctDNA+ (0%, n=0/3, p=.024). ctDNA at BL was predictive for pCR in univariable analysis (odds ratio, OR 0.22, 95%CI 0.04- 0.93, area under curve, AUC=0.69). Bivariable models including ctDNA and grade yielded the highest accuracy for analyses with ctDNA at BL (AUC=0.77), week 3 (AUC=0.79), and with ctDNA change from BL to week 3 (AUC=0.87). Conclusions Absence of ctDNA prior to and during NAT, as well as early clearance of ctDNA associates with a higher pCR rate in eBC pts with HER2-enriched intrinsic subtype treated with chemotherapy-free pembrolizumab plus dual HER2 blockade. These results are in line with prior data with more intensive therapy from I-SPY 2 and NeoALTTO trials in HER2+ eBC and indicate that ctDNA analysis during NAT appears feasible for real- time monitoring of response to treatment and could be used to guide escalation/de-escalation strategies. Further trials with a larger number of pts and pre-designed follow-up are needed to confirm the role of ctDNA for prediction of tumor response and relapse. Citation Format: Monika Graeser, Sherko Kuemmel, Oleg Gluz, Christopher Schroeder, Leon Schuetz, Olga Kelemen, Stephan Ossowski, Katarzyna Jozwiak, Mattea Reinisch, Athina Kostara, Iris Scheffen, Kerstin Lüdtke-Heckenkamp, Felix Hilpert, Angela Kentsch, Carsten Ziske, Reinhard Depenbusch, Michael Braun, Jens-Uwe Blohmer, Matthias Christgen, Stephan Bartels, Hans-Heinrich Kreipe, Enrico Pelz, Ulrike Nitz, Peter Schmid, Nadia Harbeck, Andreas Hartkopf. Circulating tumor DNA (ctDNA) for prediction of pathologic complete response (pCR) after 12 weeks of pembrolizumab + trastuzumab + pertuzumab in HER2-enriched early breast cancer: WSG-Keyriched-1 trial [abstract]. In: Proceedings of the 2023 San Antonio Breast Cancer Symposium; 2023 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2024;84(9 Suppl):Abstract nr PO1-15-12.

Abstract PO3-27-01: Prediction of pathological complete response to neoadjuvant treatment in loco-regional advancedbreast cancer by using advanced PET/CT metabolic parameters

Abstract Prediction of pathological complete response to neoadjuvant treatment in loco-regional advanced breast cancer by using advanced PET/CT metabolic parameters. Abstract: Background: The standard treatment for locally advanced breast cancer (LABC) is neoadjuvant systemic therapy followed by surgery. The usefulness of FDG PET/CT for evaluating the treatment efficacy in breast cancer is well established; however, the role of quantitative functional PET/CT parameters, such as the metabolic tumor volume (MTV) and total lesion glycolysis (TLG), remains inconclusive. The aim of this study was to determine the correlation between quantitative functional parameters derived from PET/CT in LABC patients at staging and the response to neoadjuvant therapy. Methods: Seventy-eight consecutive patients (aged 25-82 y) with confirmed LABC (stages II-III) who underwent PET/CT as a part of their staging assessment and received neoadjuvant therapy followed by surgery were retrospectively included in this study. All patients underwent an FDG PET/CT at staging; and values of SUV max, SUV mean, MTV, and TLG were determined from the PET/CT studies. The patients were divided into two groups, based on the completeness of histopathologic response to neoadjuvant therapy, both in the primary tumor and axillary lymph nodes. Clinical, imaging and demographic data were collected from patient’s medical record. Results: A complete pathologic response (pCR) was achieved in 33/78 (42%) patients. Among patients with ER/PR positive, Her2-neu negative tumors, in 23/27 (85%) patients only partial pathologic response was reached; 91% of patients with ER/PR negative, Her2-neu positive and 43% of Triple-negative patients had complete pathological? response. For the primary breast tumor response, a higher SUV max, 10.7±5.34 for pCR vs. 7.24±4.62 for non-pCR group was measured, p-value=0.003; threshold value of 6.625. A higher SUV mean in the primary tumor resulted in a significantly higher rate of pCR, 5.95±2.36 vs 4,4±2.16, p=0.004, with a threshold value of 5.795. Neither MTV, not TLG were predictive parameters of the primary tumor response. For lymph nodes response, SUV max and SUV mean had no role in the prediction of pCR. Statistically significant difference however, was found between pCR and non-pCR in MTV values in the involved lymph nodes, with a median =2.13 cm³ [0-7.85] in pCR group vs 5.43 cm³ [3.11-15.3] in non-pCR, p=0.01 (threshold value 2.375) as well as in TLG measurements, with a median of 4.64 [0-19.8] vs. 12.7 [3.21-23.1], p=0.047, (25-75 IQR). Conclusion: Primary breast tumor SUV max/mean values have a significant correlation with the treatment response and threshold values were found. Low MTV and TLG measurements in lymph nodes may predict pCR. Neither MTV nor TLG had were found to have a significant role in the prediction of the primary tumor response. Citation Format: Maisun Shamya, Olga Kagna, Zohar Keidar, Tala Palchan-Hazan, Salem Billan, Roxolyana Abdah-Bortnyak. Prediction of pathological complete response to neoadjuvant treatment in loco-regional advancedbreast cancer by using advanced PET/CT metabolic parameters [abstract]. In: Proceedings of the 2023 San Antonio Breast Cancer Symposium; 2023 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2024;84(9 Suppl):Abstract nr PO3-27-01.

Abstract PS09-01: Individual patient data meta-analysis of clinical and translational biomarkers for prediction of pathological complete response (pCR) after de-escalated therapy in HER2+ breast cancer in four trials of the West German Study Group

Abstract Background Introduction of anti-HER2 therapies substantially improved outcomes for HER2-positive early breast cancer (HER2+ eBC). However, a considerable proportion of patients (pts) may still be overtreated with systemic chemotherapy (CTx) combinations. Establishing safe de-escalation strategies to avoid toxicities requires precise patient selection. Therefore, we set out to determine predictors for efficacy and survival in the unique setting of four de-escalation trials investigating short (12-week) neoadjuvant treatments (NAT) in HER2+ eBC. We investigated the prognostic ability of clinical and translational biomarkers for pCR to identify pts with the best prognosis after de-escalated systemic CTx-free NAT. Methods 756 pts who received de-escalated NAT were analyzed: trastuzumab + pertuzumab (T + P, n=92) and T + P + paclitaxel (pac, n=42) in ADAPT-HR-/HER2+ (NCT01817452); trastuzumab emtansine (T-DM1, n=118), T-DM1 + endocrine therapy (ET, n=125), and T + ET (n=129) in ADAPT-HR+/HER2+ (NCT01779206), T + P + pac (n=107) and T + P + ET (n=100) in TP-II (NCT03272477); and T + P + pembrolizumab (n=43) in Keyriched-1 (NCT03988036) in HER2-enriched (HER2-E) subtype by PAM50. The primary endpoint of each trial was pCR (ypT0/is ypN0). Survival was a secondary endpoint in all trials but Keyriched-1; survival data for TP-II are pending. Baseline gene expression was analyzed by BC360 assay; intrinsic molecular subtypes were determined using the PAM50 predictor. Stromal tumor infiltrating lymphocytes (sTILs) were evaluated at baseline and at week 3 of NAT. sTILs were categorized using a 30% threshold and the low cellularity category ( < 500 invasive tumor cells) at week 3. Prognostic markers for pCR were identified with univariate and multivariable logistic regression models with backstep algorithm excluding variables with p≥0.1; considered variables included age, sTILs (baseline, 3-weeks), cT, cN, grade, hormone receptor and HER2 status, intrinsic subtype, and standardized ERBB2 and ESR1 gene expression. These analyses were performed separately for NAT involving systemic CTx (i.e. containing pac, n=149) and systemic CTx-free NAT (n=607). Results Overall pCR rate was 39.1% (CTx: 66.4%; CTx-free: 32.5%). Multivariable analysis identified predictors of pCR after CTx-free NAT: low cellularity at week 3 (vs < 30% sTILs: OR 3.21, 95%CI 1.85-5.57), ERBB2 (OR 1.90, 95%CI 1.41-2.55), HER2 3+ (vs 1+/2+ and ISH+: OR 8.01, 95%CI 1.65-38.99), LumA/LumB/Basal subtype (vs HER2-E: OR 0.57, 95%CI 0.34-0.97), cT2 (vs 1: OR 0.54, 95%CI 0.33-0.89), and cN1-3 (vs 0, OR 0.46, 95%CI 0.27-0.78). In CTx-containing NAT, only ERBB2 (OR 2.08, 95%CI 1.28-3.40) and additionally ESR1 (OR 0.38, 95%CI 0.23-0.61) were prognostic for pCR. Updated analysis including expanded gene expression data from ADAPT-HR+/HER2+ and a prognostic score for pCR after CTx-free NAT developed using machine learning methods will be presented at the meeting. Conclusions This pooled analysis demonstrated a 66% pCR rate after a short (12-week) de-escalated NAT in HER2+ eBC. In one-third of patients (with mainly stage I cancer, higher HER2 expression by immunohistochemistry and gene expression analysis, and low cellularity at 3 weeks), pCR can be achieved without systemic CTx. Identified biomarkers could pave the way for developing new patient selection strategies to spare CTx-associated acute and late toxicities in a clinically meaningful number of patients. Citation Format: Monika Graeser, Oleg Gluz, Christine zu Eulenburg, Sherko Küemmel, Raquel von Schumann, Matthias Christgen, Rachel Wuerstlein, Enrico Pelz, Hans-Heinrich Kreipe, Peter Schmid, Marc Thill, Michael Braun, Jochem Potenberg, Claudia Schumacher, Joke Tio, Andreas Hartkopf, Marianne Just, Christian Schem, Kerstin Lüdtke-Heckenkamp, Eva-Maria Grischke, Felix Hilpert, Angela Kentsch, Ronald Kates, Ulrike Nitz, Nadia Harbeck. Individual patient data meta-analysis of clinical and translational biomarkers for prediction of pathological complete response (pCR) after de-escalated therapy in HER2+ breast cancer in four trials of the West German Study Group [abstract]. In: Proceedings of the 2023 San Antonio Breast Cancer Symposium; 2023 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2024;84(9 Suppl):Abstract nr PS09-01.

Apriori prediction of chemotherapy response in locally advanced breast cancer patients using CT imaging and deep learning: transformer versus transfer learning

ObjectiveNeoadjuvant chemotherapy (NAC) is a key element of treatment for locally advanced breast cancer (LABC). Predicting the response to NAC for patients with Locally Advanced Breast Cancer (LABC) before treatment initiation could be beneficial to optimize therapy, ensuring the administration of effective treatments. The objective of the work here was to develop a predictive model to predict tumor response to NAC for LABC using deep learning networks and computed tomography (CT).Materials and methodsSeveral deep learning approaches were investigated including ViT transformer and VGG16, VGG19, ResNet-50, Res-Net-101, Res-Net-152, InceptionV3 and Xception transfer learning networks. These deep learning networks were applied on CT images to assess the response to NAC. Performance was evaluated based on balanced_accuracy, accuracy, sensitivity and specificity classification metrics. A ViT transformer was applied to utilize the attention mechanism in order to increase the weight of important part image which leads to better discrimination between classes.ResultsAmongst the 117 LABC patients studied, 82 (70%) had clinical-pathological response and 35 (30%) had no response to NAC. The ViT transformer obtained the best performance range (accuracy = 71 ± 3% to accuracy = 77 ± 4%, specificity = 86 ± 6% to specificity = 76 ± 3%, sensitivity = 56 ± 4% to sensitivity = 52 ± 4%, and balanced_accuracy=69 ± 3% to balanced_accuracy=69 ± 3%) depending on the split ratio of train-data and test-data. Xception network obtained the second best results (accuracy = 72 ± 4% to accuracy = 65 ± 4, specificity = 81 ± 6% to specificity = 73 ± 3%, sensitivity = 55 ± 4% to sensitivity = 52 ± 5%, and balanced_accuracy = 66 ± 5% to balanced_accuracy = 60 ± 4%). The worst results were obtained using VGG-16 transfer learning network.ConclusionDeep learning networks in conjunction with CT imaging are able to predict the tumor response to NAC for patients with LABC prior to start. A ViT transformer could obtain the best performance, which demonstrated the importance of attention mechanism.

Abstract PS05-08: Restriction Spectrum Imaging MRI for automated evaluation of response to neoadjuvant therapy in breast cancer

Abstract Introduction: Dynamic contrast-enhanced MRI (DCE) is currently used to evaluate neoadjuvant therapy response of breast cancer (1). However, DCE requires expert radiologist readers to assess the change in longest tumor dimension during therapy, as well as administration of Gadolinium contrast agents. One MRI modality that does not require contrast agents is diffusion-weighted MRI (DWI), a method that detects the microscopic diffusion of water molecules. However, the commonly used DWI method apparent diffusion coefficient (ADC) is not fully optimised in the breast (2). The purpose of the current study was to evaluate the recent DWI method Restriction Spectrum Imaging (RSI) (2) to automatically monitor breast tumor size during neoadjuvant therapy. Methods: Twenty-seven women underwent 3T MRI at four time points during therapy at University of California San Diego; 17 received all four scans (see Table 1 for patient details). Inclusion criteria included biopsy-proven unilateral invasive breast cancer ≥2.5 cm (defined on clinical examination/imaging) with indication for neoadjuvant therapy. The therapy used was primarily paclitaxel (+/-experimental agent) followed by anthracycline. The MRI protocol included Gadolinium DCE and DWI (b-values 0, 500, 1500, 4000 s/mm2); TE/TR = 82/9000 ms. ADC was calculated using b-values < 1000 s/mm2 while signal from all available b-values were fitted to the previously-developed three-component RSI model (2). The tumor size by RSI was assessed against manual DCE tumor size and mean ADC values. Prediction of therapy response during therapy and residual tumor post-therapy were assessed using non-pathological complete response (non-pCR) as endpoint. pCR was defined as ypT0/is or ypN0. Results: Ten patients experienced pCR. Prediction of non-pCR by ROC AUC at the early-therapy time point was 0.65 for RSI, 0.64 for DCE and 0.45 for ADC (Table 2). Prediction of post-therapy residual tumor is given in Table 3. Discussion: The novel RSI cancer tissue classifier predicted response to neoadjuvant therapy after only 19 days. RSI could also identify 71% of cases with residual tumor at surgery with 90% specificity post-therapy. RSI performance was similar to performance by standard MRI by manual tumor measurement on DCE. In contrast to standard-of-care MRI by using DCE and ADC that requires manual user input, the RSI classifier is automatic. This suggests that RSI may aid to cost-efficiently evaluate neoadjuvant therapy of breast cancer, with the aim to help guide clinical decision-making and enable tailored therapy regimens.