Predictive model for a second hip fracture occurrence using natural language processing and machine learning on electronic health records

Hip fractures (HFx) are associated with a higher morbidity and mortality rates, leading to a significant reduction in life quality and in limitation of patient´s mobility. The present study aimed to obtain real-world evidence on the clinical characteristics of patients with an initial and a second hip fracture (HFx) and develop a predictive model for second HFx using artificial intelligence. Electronic health records from one hospital centre in Spain from January 2011 to December 2019 were analysed using EHRead® technology, based on natural language processing and machine learning. A total of 1,960 patients with HFx were finally included during the study period after meeting all inclusion and exclusion criteria. From this total, 1835 (93.6%) patients were included in the HFx subgroup, while 124 (6.4%) were admitted to the second HFx (2HFx) subgroup. The mean age of the participants was 84 years and 75.5% were female. Most of comorbidities were more frequently identified in the HFx group, including hypertension (72.0% vs. 67.2%), cognitive impairment (33.0% vs. 31.2%), diabetes mellitus (28.7% vs. 24.8%), heart failure (27.6% vs. 22.4%) and chronic kidney disease (26.9% vs. 16.0%). Based on clinical criteria, 26 features were selected as potential prediction factors. From there, 16 demographics and clinical characteristics such as comorbidities, medications, measures of disabilities for ambulation and type of refracture were selected for development of a competitive risk model. Specifically, those predictors with different associated risk ratios, sorted from higher to lower risk relevance were visual deficit, malnutrition, walking assistance, hypothyroidism, female sex, osteoporosis treatment, pertrochanteric fracture, dementia, age at index, osteoporosis, renal failure, stroke, COPD, heart disease, anaemia, and asthma. This model showed good performance (dependent AUC: 0.69; apparent performance: 0.75) and could help the identification of patients with higher risk of developing a second HFx, allowing preventive measures. This study expands the current available information of HFx patients in Spain and identifies factors that exhibit potential in predicting a second HFx among older patients.

Characterization and spatial distribution of the immune cell infiltrate in triple-negative breast cancer: a novel classification based on plasma cells and CD8+ T cells

Stromal tumor-infiltrating lymphocytes (sTILs) are a robust prognostic and predictive biomarker in triple-negative breast carcinoma. However, the sTIL compartment comprises different cell populations. The aim of the study is to characterize the distribution of T cells (CD3+ and CD8+), B cells, and plasma cells and explore their association with outcome in the surgical specimen of 62 patients. Furthermore, programmed death ligand 1 expression and the presence of tertiary lymphoid structures (TLSs) are explored. Patients with higher sTILs achieve better progression-free survival (PFS) (P = .0013), and tumors have more plasma cells in the infiltrate. Specifically, higher counts of T cells (both CD3+ and CD8+) have better PFS (P = .002 and P = .0086, respectively) as it is observed in tumors with higher infiltration of CD8+ T cells in the tumor core (P = .035). Higher infiltration by B cells and plasma cells shows a positive tendency toward increased PFS (P = .06 and P = .058). Programmed death ligand 1 (SP142) is positive in 56% of tumors. Tumors with at least 1 TLS (42%) show higher CD8+ T cell infiltration in the tumor core and the sTIL value doubles compared to tumors devoid of TLSs [sTIL mean: 36 ± 11% and 18 ± 5% (CI [Confidence Interval]: 95%), respectively]. Our study demonstrates that the characterization of the immune cell infiltration is as relevant as its distribution. Moreover, the importance of considering different immune cell types for classification is emphasized. Therefore, a new classification of triple-negative breast carcinoma immune infiltration with CD8+ T cell and plasma cell densities in the tumor core and infiltrative margin is proposed.

Pre-existing tumor host immunity characterization in resected non-small cell lung cancer

INTRODUCTION

Neoadjuvant and adjuvant immune checkpoint blockade (ICB) have recently become standard of care in resectable non-small cell lung cancer (NSCLC). Yet, biomarkers that inform patients who benefit from this approach remain largely unknown. Here, we interrogated the tumor immune microenvironment (TIME) in early-stage NSCLC patients that underwent up-front surgery.

METHODS

A total of 185 treatment-naïve patients with early-stage NSCLC, that underwent up-front surgical treatment between 2006 and 2018 at Hospital del Mar were included. 124 lung adenocarcinomas (LUADs), and 61 squamous cell carcinoma (LUSCs) were included in a tissue microarray. Immunohistochemistry for CD3, CD4, CD8, CD68, CD80, CD103, FOXP3, PD-1, PD-L1, PD-L2 and HLA class II were evaluated by digital image analysis (QuPath software). TIME was categorized into four groups using PD-L1 expression in tumor cells (<1 % or ≥1 %) and tumor resident memory (CD103⁺) immune cells (using the median as cut-off). We explored the association between different TIME dimensions and patient's clinicopathological features and outcomes.

RESULTS

We found increased levels of T cell markers (CD3⁺, CD4⁺, CD8⁺ cells), functional immune markers (FOXP3⁺ cells) as well as, higher HLA-II tumor membrane expression in LUADs compared to LUSCs (p < 0.05 for all). In contrast, LUSCs displayed higher percentage of intratumor macrophages (CD68⁺ cells) as well as, higher PD-L1 and PD-L2 tumor membrane expression (p < 0.05 for all). Unsupervised analysis revealed three different tumor subsets characterized by membrane tumor expression of PD-L1, PD-L2 and HLA-class II. Enrichment of T cells (CD3⁺, CD8⁺ cells), regulatory T cells (FOXP3⁺ cells) and macrophages (CD68⁺ cells) was observed in the CD103⁺/PD-L1⁺ group (p < 0.05 for all). Multivariate analysis showed that infiltration by CD103⁺ immune cells was associated with improved OS (p = 0.009).

CONCLUSIONS

TIME analysis in resected NSCLC highlighted differences by histology, PD-L1 expression and molecular subgroups. Biomarker studies using IHC might aid to individually tailor adjuvant treatment in early-stage NSCLC.

Liquid Biopsy Detects Early Molecular Response and Predicts Benefit to First-Line Chemotherapy plus Cetuximab in Metastatic Colorectal Cancer: PLATFORM-B Study

PURPOSE

Chemotherapy plus anti-EGFR is standard first-line therapy in RAS wild-type (wt) metastatic colorectal cancer (mCRC), but biomarkers of early response are clinically needed. We aimed to define the utility of ctDNA to assess early response in patients with mCRC receiving first-line anti-EGFR therapy.

EXPERIMENTAL DESIGN

Prospective multicentric study of tissue patients with RAS wt mCRC treated with first-line chemotherapy plus cetuximab undergoing sequential liquid biopsies. Baseline and early (C3) ctDNA were analyzed by NGS. Trunk mutations were assessed as surrogate marker of total tumor burden. RAS/BRAF/MEK/EGFR-ECD were considered mutations of resistance. ctDNA results were correlated with clinical outcome.

RESULTS

One hundred patients were included. ctDNA was detected in 72% of patients at baseline and 34% at C3. Decrease in ctDNA trunk mutations correlated with progression-free survival (PFS; HR, 0.23; P = 0.001). RAS/BRAF were the only resistant mutations detected at C3. An increase in the relative fraction of RAS/BRAF at C3 was followed by an expansion of the RAS clone until PD, and was associated with shorter PFS (HR, 10.5; P < 0.001). The best predictor of response was the combined analysis of trunk and resistant mutations at C3. Accordingly, patients with "early molecular response" (decrease in trunk and decrease in resistant mutations) had better response (77.5% vs. 25%, P = 0.008) and longer PFS (HR, 0.18; P < 0.001) compared with patients with "early molecular progression" (increase in trunk and/or increase in resistant mutations).

CONCLUSIONS

ctDNA detects early molecular response and predicts benefit to chemotherapy plus cetuximab. A comprehensive NGS-based approach is recommended to integrate information on total disease burden and resistant mutations. See related commentary by Eluri et al., p. 302.

The FLARE score, circulating neutrophils, and association with COVID-19 outcomes in patients with solid tumors

2551

Background: Inflammation and neutrophils play a central role in severe Covid-19 disease. In previous data, we showed that the FLARE score, combining both tumor and Covid-19-induced proinflammatory status (proinflam-status), predicts early mortality in cancer patients (pts) with Covid-19 infection. We aimed to assess the impact of this score in a larger cohort and characterize the immunophenotype (IF) of circulating neutrophils. Methods: Multicenter retrospective cohort (RC) of pts with cancer and Covid-19 infection across 14 international centers. Circulating inflammatory markers were collected at two timepoints: baseline (-15 to -45d before Covid-19 diagnosis) and Covid-19 diagnosis. Tumor-induced proinflam-status was defined by high dNLR (neutrophils/(leucocytes-neutrophils)>3) at baseline. Covid-19-induced proinflam-status was defined by +100% increase of dNLR between both timepoints. We built the FLARE score combining both Tumor and Infection-induced inflammation: T+/I+ (poor), if both proinflam-status; T+/I- (T-only), if inflammation only due to tumor; T-/I+ (I-only), if inflammation only due to Covid; T-/I- (favorable), if no proinflam-status. The IF of circulating neutrophils by flow cytometry was determined in a unicenter prospective cohort (PC) of pts with cancer during Covid-19 infection and in healthy volunteers (HV). Primary endpoint was 30-day mortality. Results: 524 pts were enrolled in the RC with a median follow-up of 84d (95%CI 78-90). Median age was 69 (range 35-98), 52% were male and 78% had baseline PS <1.Thoracic cancers were the most common (26%). 70% had active disease, 51% advanced stage and 57% were under systemic therapy. dNLR was high in 25% at baseline vs 55% at Covid-19 diagnosis. The median dNLR increase between both timepoints was +70% (IQR: 0-349%); 42% had +100% increase of dNLR. Pts distribution and mortality across FLARE groups is resumed in the Table. Overall mortality rate was 26%. In multivariate analysis, including gender, stage and PS, the FLARE poor group was independently associated with 30-day mortality [OR 5.27; 1.37-20.3]. 44 pts were enrolled in the PC. Median circulating neutrophils were higher in pts with cancer (n=10, 56.7% [IQR: 39-78.4%]) vs HV (n=6, 35.8% [IQR: 25.6-21%]), and particularly higher in pts with cancer and severe Covid-19 infection (n=7, 88.6% [IQR: 80.9-94%] (p=0.003). A more comprehensive characterization of the IF of circulating neutrophils, including Lox1/CD62/CD64, will be presented at ASCO. Conclusions: The FLARE score, combining tumor and Covid-19-induced proinflam-status, can identify the population at higher risk for mortality. A better characterization of circulating neutrophils may help improve the prediction of Covid-19 outcomes in pts with cancer. [Table: see text]

Predicting recurrence of venous thromboembolism in anticoagulated cancer patients using real-world data and machine learning

e18742

Background: The clinical predictors of venous thromboembolism (VTE) recurrence in patients with cancer are not well known. Our objective was to develop a predictive model for risk of VTE recurrence in anticoagulant-treated cancer patients within the first 6 months following VTE diagnosis. Methods: Using the EHRead technology, based on Natural Language Processing (NLP) and machine learning (ML), the unstructured clinical data in EHRs from 9 Spanish hospitals between 2014 and 2018 was extracted and analyzed. The study population, comprising all adult anticoagulated cancer patients with VTE, was downsampled to prevent bias and class imbalance. A total of 94 patient characteristics were explored, and Random Forest (RF) feature selection was performed to identify the most relevant predictors for VTE recurrence. Multiple algorithms were used to train different prediction models, which were subsequently validated in a hold-out dataset. The model with the best performance metrics (i.e., ROC-AUC) was selected as the final model. Results: From a source population of 2,893,208 patients, 21,227 anticoagulant-treated patients with VTE and active cancer (53.9% male, median age of 70 years) were identified. Across the study period, yearly incidence of VTE remained relatively stable, ranging from 2.7 to 3.9%. The most common type of VTE was deep vein thrombosis (68.2% of patients), followed by pulmonary embolism (28.4%). Regarding primary cancer locations, the most frequent were colorectal (10.1%) and lung cancer (8.5%). Of all trained and validated models, the RF approach yielded the best performance, with a ROC-AUC = 0.72. The following predictors of VTE recurrence were identified: pulmonary embolism, deep vein thrombosis, metastasis, adenocarcinoma, hemoglobin values, serum creatinine values, platelet count, leukocyte count, family history of VTE, and patients’ age. Conclusions: Using NLP and ML, we were able to use the real-world data in EHRs to build a predictive model of VTE recurrence in cancer patients based on individual clinical features. These results may improve the clinical management of VTE recurrence in this population.

Quantification and progress over time of specific antibodies against SARS-CoV-2 in breast milk of lactating women vaccinated with BNT162b2 Pfizer-BioNTech COVID-19 vaccine (LacCOVID)

Background

Several observational studies demonstrated the passage of postvaccine antibodies through breast milk in women vaccinated against coronavirus disease 2019 (COVID-19), mostly with messenger RNA (mRNA)–based vaccines, but lacked long-term data.

Methods

A 6-month prospective cohort study was performed to determine severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) vaccine–induced antibody levels in the breast milk of 33 lactating healthcare workers at different timepoints after mRNA BNT162b2 Pfizer-BioNTech COVID-19 vaccination. Moreover, we examined the correlation of SARS-CoV-2 antibody levels between serum and breast milk, adverse events related to vaccination, and rate of SARS-CoV-2 infections.

Results

Mothers’ median age was 38 (interquartile range [IQR], 36–39) years and 15 (IQR, 10–22) months for infants. Median (IQR) SARS-CoV-2 immunoglobulin G (IgG) spike protein subunit S1 (S1) vaccine–induced levels at different timepoints for serum–milk pairs were 519 (234–937) to 1 (0–2.9) arbitrary units (AU)/mL at 2 weeks after first dose and 18 644 (9923–29 264) to 78 (33.7–128), 12 478 (6870–20 801) to 50.4 (24.3–104), 4094 (2413–8480) to 19.9 (10.8–51.9), 1350 (831–2298) to 8.9 (7.8–31.5) AU/mL at 2, 4, 12 and 24 weeks after second dose, respectively. We observed a positive correlation of antibody levels between serum and breast milk, no serious adverse events related to vaccination, and 2 (6%) COVID-19 vaccine breakthrough infections.

Conclusions

Women vaccinated with Pfizer-BioNTech transmit antibodies into breast milk with a positive correlation with serum levels. Both decreased over time in a 6-month follow-up.

Comprehensive Analysis of Metabolic Isozyme Targets in Cancer

Metabolic reprogramming is a hallmark of malignant transformation, and loss of isozyme diversity (LID) contributes to this process. Isozymes are distinct proteins that catalyze the same enzymatic reaction but can have different kinetic characteristics, subcellular localization, and tissue specificity. Cancer-dominant isozymes that catalyze rate-limiting reactions in critical metabolic processes represent potential therapeutic targets. Here, we examined the isozyme expression patterns of 1,319 enzymatic reactions in 14 cancer types and their matching normal tissues using The Cancer Genome Atlas mRNA expression data to identify isozymes that become cancer-dominant. Of the reactions analyzed, 357 demonstrated LID in at least one cancer type. Assessment of the expression patterns in over 600 cell lines in the Cancer Cell Line Encyclopedia showed that these reactions reflect cellular changes instead of differences in tissue composition; 50% of the LID-affected isozymes showed cancer-dominant expression in the corresponding cell lines. The functional importance of the cancer-dominant isozymes was assessed in genome-wide CRISPR and RNAi loss-of-function screens: 17% were critical for cell proliferation, indicating their potential as therapeutic targets. Lists of prioritized novel metabolic targets were developed for 14 cancer types; the most broadly shared and functionally validated target was acetyl-CoA carboxylase 1 (ACC1). Small molecule inhibition of ACC reduced breast cancer viability in vitro and suppressed tumor growth in cell line- and patient-derived xenografts in vivo. Evaluation of the effects of drug treatment revealed significant metabolic and transcriptional perturbations. Overall, this systematic analysis of isozyme expression patterns elucidates an important aspect of cancer metabolic plasticity and reveals putative metabolic vulnerabilities.

SIGNIFICANCE

This study exploits the loss of metabolic isozyme diversity common in cancer and reveals a rich pool of potential therapeutic targets that will allow the repurposing of existing inhibitors for anticancer therapy. See related commentary by Kehinde and Parker, p. 1695.

Exploring the Association of Cancer and Depression in Electronic Health Records: Combining Encoded Diagnosis and Mining Free Text Clinical Notes (Preprint)

Background

A cancer diagnosis is a source of psychological and emotional stress, which are often maintained for sustained periods of time that may lead to depressive disorders. Depression is one of the most common psychological conditions in patients with cancer. According to the Global Cancer Observatory, breast and colorectal cancers are the most prevalent cancers in both sexes and across all age groups in Spain.

Objective

This study aimed to compare the prevalence of depression in patients before and after the diagnosis of breast or colorectal cancer, as well as to assess the usefulness of the analysis of free-text clinical notes in 2 languages (Spanish or Catalan) for detecting depression in combination with encoded diagnoses.

Methods

We carried out an analysis of the electronic health records from a general hospital by considering the different sources of clinical information related to depression in patients with breast and colorectal cancer. This analysis included ICD-9-CM (International Classification of Diseases, Ninth Revision, Clinical Modification) diagnosis codes and unstructured information extracted by mining free-text clinical notes via natural language processing tools based on Systematized Nomenclature of Medicine Clinical Terms that mentions symptoms and drugs used for the treatment of depression.

Results

We observed that the percentage of patients diagnosed with depressive disorders significantly increased after cancer diagnosis in the 2 types of cancer considered—breast and colorectal cancers. We managed to identify a higher number of patients with depression by mining free-text clinical notes than the group selected exclusively on ICD-9-CM codes, increasing the number of patients diagnosed with depression by 34.8% (441/1269). In addition, the number of patients with depression who received chemotherapy was higher than those who did not receive this treatment, with significant differences (P<.001).

Conclusions

This study provides new clinical evidence of the depression-cancer comorbidity and supports the use of natural language processing for extracting and analyzing free-text clinical notes from electronic health records, contributing to the identification of additional clinical data that complements those provided by coded data to improve the management of these patients.

mTOR inhibition and trastuzumab-emtansine (T-DM1) in HER2-positive breast cancer

In patients with trastuzumab-resistant HER2-positive breast cancer, the combination of everolimus (mTORC1 inhibitor) with trastuzumab failed to show a clinically significant benefit. However, the combination of mTOR inhibition and the antibody-drug conjugate (ADC) trastuzumab-emtansine (T-DM1) remains unexplored. We tested T-DM1 plus everolimus in a broad panel of HER2-positive breast cancer cell lines. The combination was superior to T-DM1 alone in four cell lines (HCC1954, SKBR3, EFM192A, and MDA-MB-36) and in two cultures from primary tumor cells derived from HER2-positive patient-derived xenografts (PDX), but not in BT474 cells. In the trastuzumab-resistant HCC1954 cell line, we characterized the effects of the combination using TAK-228 (mTORC1 and 2 inhibitor) and knockdown of the different mTOR complex components. T-DM1 did not affect mTOR downstream signaling nor induct autophagy. Importantly, mTOR inhibition increased intracellular T-DM1 levels, leading to increased lysosomal accumulation of the compound. The increased efficacy of mTOR inhibition plus T-DM1 was abrogated by lysosome inhibitors (chloroquine and bafilomycin A1). Our experiments suggest that BT474 are less sensitive to T-DM1 due to lack of optimal lysosomal processing and intrinsic resistance to the DM1 moiety. Finally, we performed several in vivo experiments that corroborated the superior activity of T-DM1 and everolimus in HCC1954 and PDX-derived mouse models. In summary, everolimus in combination with T-DM1 showed strong antitumor effects in HER2-positive breast cancer, both in vitro and in vivo. This effect might be related, at least partially, to mTOR-dependent lysosomal processing of T-DM1, a finding that might apply to other ADCs that require lysosomal processing. Implications: Inhibition of mTOR increases the anti-tumor activity of T-DM1, supporting that the combination of mTOR inhibitors and antibody-drug conjugates warrants clinical evaluation in patients with HER2-positive breast cancer.

LCOR mediates interferon-independent tumor immunogenicity and responsiveness to immune-checkpoint blockade in triple-negative breast cancer

Ligand-dependent corepressor (LCOR) mediates normal and malignant breast stem cell differentiation. Cancer stem cells (CSCs) generate phenotypic heterogeneity and drive therapy resistance, yet their role in immunotherapy is poorly understood. Here we show that immune-checkpoint blockade (ICB) therapy selects for LCORlow CSCs with reduced antigen processing/presentation machinery (APM) driving immune escape and ICB resistance in triple-negative breast cancer (TNBC). We unveil an unexpected function of LCOR as a master transcriptional activator of APM genes binding to IFN-stimulated response elements (ISREs) in an IFN signaling-independent manner. Through genetic modification of LCOR expression, we demonstrate its central role in modulation of tumor immunogenicity and ICB responsiveness. In TNBC, LCOR associates with ICB clinical response. Importantly, extracellular vesicle (EV) Lcor-messenger RNA therapy in combination with anti-PD-L1 overcame resistance and eradicated breast cancer metastasis in preclinical models. Collectively, these data support LCOR as a promising target for enhancement of ICB efficacy in TNBC, by boosting of tumor APM independently of IFN.

Preclinical and Clinical Characterization of Fibroblast-derived Neuregulin-1 on Trastuzumab and Pertuzumab Activity in HER2-positive Breast Cancer

PURPOSE

To characterize expression of neuregulin-1 (NRG1), an HER3 ligand, in HER2-positive breast cancer and its relation with the efficacy of trastuzumab with or without pertuzumab.

EXPERIMENTAL DESIGN

Characterization of NRG1 expression in tumor cell lines, in tumor specimens, and in cancer-associated fibroblasts (CAFs). Patient-derived CAFs were used to investigate NRG1 impact on the activity of trastuzumab with or without pertuzumab in HER2-positive breast cancer cells. The relationship between NRG1 expression and pathologic response to anti-HER2-based neoadjuvant therapy was assessed in a retrospective patient cohort and in the NeoSphere trial.

RESULTS

NRG1 was expressed in HER2-positive breast cancer-derived fibroblasts at significantly higher levels than in cancer cells. NRG1 and the conditioned media (CM) from CAFs phosphorylated HER3 and AKT in cancer cells and mediated trastuzumab resistance. Stable genetic depletion of NRG1 from CAFs overcame trastuzumab resistance. Pertuzumab effectively suppressed trastuzumab resistance mediated by either NRG1 or CAF's CM. NRG1 engaged an epithelial-to-mesenchymal transition that was prevented by trastuzumab and pertuzumab. In clinical samples, stromal and/or tumor cell expression of NRG1 determined by immunohistochemistry was uncommon (13.2%) yet significantly linked with residual disease following trastuzumab-based neoadjuvant therapy. In the NeoSphere trial, the magnitude of the difference of pathologic complete response rates favoring the pertuzumab arm was higher in the NRG1-high group.

CONCLUSIONS

CAF-derived NRG1 mediates trastuzumab resistance through HER3/AKT, which might be reverted by pertuzumab. In patients with HER2-positive breast cancer, high expression of NRG1 was associated to poor response to trastuzumab, but not in combination with pertuzumab.

Abstract 423: Pooled analysis of matching score and patient outcome in I-PREDICT and WINTHER studies

Introduction: Comprehensive genomic profiling (CGP) of tumors is available at many specialized cancer centers, potentially allowing oncologists to individualize patient therapy based on tumor molecular features. However, clinicians still lack a systematic approach to deal with the wealth of information provided by CGP. The proportion of plausible drug-matched alterations over the total number of alterations detected by CGP (Matching Score, MS) has been postulated to predict the likelihood of patient benefit. We performed a systematic analysis of MS and its association with the outcome of patients treated according to tumor DNA alterations within two recent precision oncology trials.

Materials and Methods: We downloaded the supplementary data from WINTHER (NCT01856296) and I-PREDICT (NCT02534675) studies. We included patients who received treatment for one or more single-gene alterations, excluding eight patients that received immunotherapy due to high tumor mutational burden, PD-L1 positivity or microsatellite instability. We used MS cut-offs of 0.25 (as in WINTHER) and 0.5 (as in I-PREDICT) and also MS as a continuous variable to study its association with clinical variables such as age, sex, tumor type, and number of previous treatment lines (Kruskall-Wallis and Wilcoxon-rank tests), as well as disease control rate (DCR), progression-free survival (PFS) and overall survival (OS) (Cox proportional hazards models).

Results: The pooled analysis included 144 patients. Median age was 60 years (range 21-86) and 64% had received ≥ 2 lines of therapy. Most patients had colorectal (37%), lung (14%), or breast (11%) tumors, and 58% were treated with 2 or more drugs. MS was not associated with any clinical variable nor with DCR. MS &gt; 0.25 (n=98; 68%) was associated with longer PFS (median 3.5 vs. 1.9 months, HR [95%CI]: 0.60 [0.41-0.87]) but not with OS. In contrast, MS &gt; 0.5 (n=41; 28%) was associated with longer OS (median 14.1 vs. 6.5 months, HR [95%CI]: 0.50 [0.29-0.54]) but not with PFS. As a continuous variable, increasing MS values were associated with better OS (HR [95% CI]: 0.43 [0.20-0.95]) and with a non-significant trend towards better PFS (HR [95% CI]: 0.54 [0.21-1.05]). In a multivariable model adjusted for Age and study, PFS/OS probabilities improved consistently with increasing MS values from 0 to 0.5. With MS values &gt; 0.5, OS probabilities kept increasing while PFS probabilities plateaued.

Conclusions: The significant and consistent relationship between MS and OS (in contrast with PFS) may be explained by patients with higher MS values having more targeted therapies available for successive treatment lines, although we cannot exclude the existence of other confounding prognostic variables. We note that, as better drugs are developed and new knowledge is incorporated into the clinic, the value of CGP will probably increase over time, improving the number of patients with actionable alterations and their outcome.

Citation Format: David Casadevall, Michal Marczyk, Xavier Monzonis, Lajos Pusztai, Joan Albanell. Pooled analysis of matching score and patient outcome in I-PREDICT and WINTHER studies [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 423.

Clinical Impact of Presurgery Circulating Tumor DNA after Total Neoadjuvant Treatment in Locally Advanced Rectal Cancer: A Biomarker Study from the GEMCAD 1402 Trial

PURPOSE

Total neoadjuvant treatment (TNT) is a valid strategy for patients with high-risk locally advanced rectal cancer (LARC). Biomarkers of response to TNT are an unmet clinical need. We aimed to determine the value of circulating tumor DNA (ctDNA) to predict tumor response, recurrence, and survival in patients with LARC treated with TNT.

EXPERIMENTAL DESIGN

The GEMCAD 1402 was a phase II randomized, multicentric clinical trial that randomized 180 patients with LARC to modified schedule of fluorouracil, leucovorin, and oxaliplatin (mFOLFOX6) +/- aflibercept, followed by chemoradiation and surgery. Plasma samples were collected at baseline and after TNT within 48 hours before surgery (presurgery). An ultrasensitive assay that integrates genomic and epigenomic cancer signatures was used to assess ctDNA status. ctDNA results were correlated with variables of local tumor response in the surgery sample, local/systemic recurrence, and survival.

RESULTS

A total of 144 paired plasma samples from 72 patients were included. ctDNA was detectable in 83% of patients at baseline and in 15% following TNT (presurgery). No association was found between ctDNA status and pathologic response. Detectable presurgery ctDNA was significantly associated with systemic recurrence, shorter disease-free survival (HR, 4; P = 0.033), and shorter overall survival (HR, 23; P < 0.0001).

CONCLUSIONS

In patients with LARC treated with TNT, presurgery ctDNA detected minimal metastatic disease identifying patients at high risk of distant recurrence and death. This study sets the basis for prospective clinical trials that use liquid biopsy to personalize the therapeutic approach following TNT.

Abstract P4-10-17: Baseline and pharmacodynamic changes of circulating exosomal microRNAs predict early versus late progression to palbociclib plus endocrine therapy in patients with metastatic breast cancer. A sub-analysis of the PARSIFAL-1 trial

BACKGROUND: Palbociclib in combination with endocrine therapy (ET) is the first- or second-line standard of care for patients (pts) with hormone receptor (HR)-positive (+)/human epidermal growth factor receptor 2 (HER2)-negative (-) metastatic breast cancer (MBC). No clinically validated markers of long-term benefit from palbociclib have been established and the clinically relevant resistance mechanisms to cyclin-dependent kinases 4 and 6 inhibitors remain undefined. In the present study, we explored candidate circulating exosomal microRNAs (miRNAs) as putative predictors of benefit and/or resistance to palbociclib in combination with ET in pts included in the PARSIFAL-1 trial (ClinicalTrials.gov identifier: NCT02491983). MATERIALS AND METHODS: PARSIFAL-1 was a randomized, open-label, phase II trial aimed at evaluating the efficacy and safety of palbociclib plus either fulvestrant or letrozole in HR+/HER2- MBC pts. For the study of exosomal miRNAs, forty-five consenting pts were selected based on primary endocrine resistance according to the ABC-4 criteria. Nine pts who progressed within the first six months after treatment initiation in the absence of an objective response were considered Resistant and thirty-six pts who progressed more than six months after starting therapy were classified as Sensitive. Exosomes were isolated from plasma samples collected at study entry (baseline) and after 12 weeks of treatment initiation. Differences in miRNA expression between Resistant and Sensitive pts as well as miRNA pharmacodynamic changes between baseline and 12-week samples were assessed. Ribonucleic acid (RNA) was isolated using the miRNeasy plasma kit and the library preparation was done using the QIAseq miRNA library kit according to manufacturer’s instructions. All experiments were conducted at QIAGEN Genomic Services in Germany. Differential expression of miRNA between different conditions was studied using EdgeR statistical software package from Bioconductor. Estimated p-values for significantly differentially expressed miRNAs were adjusted using Benjamini-Hochberg’s False Discovery Rate (FDR). Differences in expression of miRNA with an FDR below 0.05 were considered significant. RESULTS: Sequencing of exosomal RNA and preparation of miRNA libraries were successful for all included samples, with good technical performance. On average, 2.8 million Unique Molecular Index-corrected reads were obtained for each sample and the average percentage of mappable reads was 34.5%. Overall, expression of miRNAs was higher in Resistant compared with Sensitive pts. Eight miRNAs were called as differentially expressed between the two groups. Four miRNAs (miR-1246; miR-375; miR-193a-5p; miR-181d-5p) were differentially expressed at baseline, three (miR-196a-5p, miR-200a-3p, miR-320d) were differentially expressed at 12 weeks, and expression of one miRNA (miR-141-3p) was consistently higher across both time points in Resistant pts. Significant pharmacodynamic changes in miRNA expression were observed both in Resistant and Sensitive pts. Seven miRNAs (miR-224-5p, miR-16-5p, let-7a-5p, miR-381-3p, miR-200c-3p, miR-493-3p, let-7b-5p) in Sensitive pts and three miRNAs (miR-223-3p, miR-126-3p, miR-320b) in Resistant pts were differentially expressed between baseline and 12-week samples. CONCLUSIONS: Circulating exosomal miRNA profiling is feasible in liquid biopsies from MBC pts. Differential expression of selected miRNAs at baseline or their pharmacodynamic modulation may predict benefit from palbociclib combined with ET in pts with HR+/HER2- MBC. Validation of the most promising miRNAs by custom quantitative PCR is warranted.

Citation Format: David Casadevall, Joan Albanell Mestres, Federico Rojo, Beatríz Bellosillo, Abel González, Violeta Serra, Meritxell Bellet, Miguel Angel Gil, Miquel Angel Pujana, Angelo Gámez, Enrique Espinosa, Peter Schmid, Joseph Gligorov, Frederik Marmé, Carlos L. Arteaga, Leonardo Mina, Andrea Malfettone, Miguel Sampayo, José Manuel Pérez-García, Javier Cortés, Antonio Llombart-Cussac. Baseline and pharmacodynamic changes of circulating exosomal microRNAs predict early versus late progression to palbociclib plus endocrine therapy in patients with metastatic breast cancer. A sub-analysis of the PARSIFAL-1 trial [abstract]. In: Proceedings of the 2019 San Antonio Breast Cancer Symposium; 2019 Dec 10-14; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2020;80(4 Suppl):Abstract nr P4-10-17.

Circulating Tumor Cells: Applications for Early Breast Cancer

Breast cancer is the most common malignancy among women. Most of breast cancer patients are diagnosed in early stages and will be treated with curative intent. Despite this, some patients will relapse. The identification of patients at high risk remains an important challenge. CTCs can be useful to identify this patients, to assess tumor dynamics and to monitoring therapy. There is definitive evidence on the prognostic role of CTCs in early breast cancer (eBC) but its clinical utility in daily practice is still lacking. We have to take into consideration that the studies published to date mainly evaluated the presence of CTC based on the expression of epithelial surface markers. Future studies need to overcome this limitation and important advances in technical methods can assess CTCs and capture the heterogeneity of the tumor landscape. It is also tempting to speculate that CTCs may also provide complementary information on the interplay of tumor cells with the immune system. The combination of different methods to detect tumoral disease by liquid biopsy may provide new ways to personalize in an unprecedented manner the management of patients with eBC.

Dynamic clonal remodelling in breast cancer metastases is associated with subtype conversion

BACKGROUND

Changes in the clinical subtype (CS) and intrinsic subtype (IS) between breast cancer (BC) metastases and corresponding primary tumours have been reported. However, their relationship with tumour genomic changes remains poorly characterised. Here, we analysed the association between genomic remodelling and subtype conversion in paired primary and metastatic BC samples.

METHODS

A total of 57 paired primary and metastatic tumours from GEICAM/2009-03 (ConvertHER, NCT01377363) study participants with centrally assessed CS (n = 57) and IS (n = 46) were analysed. Targeted capture and next-generation sequencing of 202 genes on formalin-fixed paraffin-embedded samples was performed. The cancer cell fraction (CCF) of mutations in primary and metastatic pairs was estimated as a surrogate of tumour clonal architecture. Changes in mutation CCF between matched primary and metastatic tumours were analysed in the presence or absence of subtype conversion.

FINDINGS

CS conversion occurred in 24.6% and IS conversion occurred in 36.9% of metastases. Primary tumours and metastases had a median of 11 (range, 3-29) and 9 (range, 1-38) mutations, respectively (P = 0.05). Overall, mutations in metastases showed a higher estimated CCF than in primary tumours (median CCF, 0.51 and 0.47, respectively; P = 0.042), consistent with increased clonal homogeneity. The increase in mutation CCF was significant in CS-converted (P = 0.04) but not in IS-converted (P = 0.48) metastases. Clonal remodelling was highest in metastases from hormone receptor-positive and human epidermal growth factor 2 (HER2)-positive tumours (P = 0.006).

CONCLUSIONS

Mutations in BC metastases showed significantly higher estimated CCF than primary tumours. CCF changes were more prominent in metastases with CS conversion. Our findings suggest that changes in BC subtypes are linked to clonal remodelling during BC evolution.

Immune microenvironment of triple-negative breast cancer in African-American and Caucasian women

PURPOSE

African-American (AA) patients with triple-negative breast cancer (TNBC) are less likely to achieve pathologic complete response from neoadjuvant chemotherapy and have poorer prognosis than Caucasian patients with TNBC, suggesting potential biological differences by race. Immune infiltration is the most consistent predictive marker for chemotherapy response and improved prognosis in TNBC. In this study, we test the hypothesis that the immune microenvironment differs between AA and Caucasian patients.

METHODS

RNA-seq expression data were obtained from The Cancer Genome Atlas (TCGA) database for 162 AA and 697 Caucasian breast cancers. Estrogen receptor (ER)-positive, human epidermal growth factor receptor-2 (HER2)-positive, and TNBC subtypes were included in the analyses. Tumor infiltrating lymphocyte (TIL) counts, immunomodulatory scores, and molecular subtypes were obtained from prior publications for a subset of the TNBC cases. Differences in immune cell distributions and immune functions, measured through gene expression and TIL counts, as well as neoantigen, somatic mutation, amplification, and deletion loads, were compared by race and tumor subtype.

RESULTS

Immune metagene analysis demonstrated marginal immune attenuation in AA TNBC relative to Caucasian TNBC that did not reach statistical significance. The distributions of immune cell populations, lymphocyte infiltration, molecular subtypes, and genomic aberrations between AA and Caucasian subtypes were also not significantly different. The MHC1 metagene demonstrated increased expression in AA ER-positive cancers relative to Caucasian ER-positive cancers.

CONCLUSIONS

This study suggests that the immunological differences between AA and Caucasian breast cancers represented by TCGA data are subtle, if they exist at all. We observed no consistent racial differences in immune gene expression or TIL counts in TNBC by race. However, this study cannot rule out small differences in immune cell subtype distribution and activity status that may not be apparent in bulk RNA analysis.

The prognostic role of epigenetic dysregulation in bladder cancer: A systematic review

BACKGROUND

Despite adequate treatment and follow-up, around one fifth of patients with localized bladder cancer will present with disease progression. Adequate prognostic biomarkers are lacking to define patients who are at risk. Mutations in chromatin remodeling genes are more frequently found in bladder cancer than in any other solid tumor. However, the prognostic relevance of epigenetic dysregulation has not been established and may offer an opportunity for biomarker discovery.

METHODS

Looking for prognostic epigenetic factors, we performed a comprehensive PubMed search using keywords such as "bladder cancer", "chromatin remodeling", "gene methylation" and "epigenetics". We only included studies reporting on the association of epigenetic markers with prognostic outcomes such as recurrence, progression or survival.

RESULTS

Of 1113 results, 87 studies met the inclusion criteria, which represented a total of 85 epigenetic markers with potential prognostic relevance. No prospective studies were identified. Seventy-three percent (64/87) of the studies involved mixed cohorts of muscle invasive and non-muscle invasive bladder cancer. Promoter methylation of genes with putative prognostic value affected cellular processes such as cell cycle, apoptosis, cell-adhesion or migration, as well as critical pathways such as MAP-kinase or Wnt. Alteration of chromatin regulatory elements suggest a prognostic relevance alterations leading to a predominantly silenced chromatin state.

CONCLUSIONS

The prognostic impact of epigenetic alterations in bladder cancer is still unclear. Prospective evaluation of methylation marks and chromatin remodeling gene alterations using consistent methods and criteria is warranted.

CD274 (PDL1) and JAK2 genomic amplifications in pulmonary squamous-cell and adenocarcinoma patients

AIMS

CD274 (PDL1) and JAK2 (9p24.1) gene amplifications have been recently described in pulmonary carcinomas in association with programmed death-ligand 1 (PD-L1) expression. Furthermore, PTEN loss has been explored preclinically in relation to PD-L1 expression. Our aim was to determine whether these genomic alterations affect PD-L1 expression levels in non-small-cell lung cancer.

METHODS AND RESULTS

PD-L1 and PTEN expression determined by immunohistochemistry (IHC), and CD274, JAK2 and PTEN copy number alterations (CNAs) determined by fluorescence in-situ hybridisation, were studied in 171 pulmonary carcinoma specimens. PD-L1 expression was positive in 40 cases (23.3%), and CD274 amplification was present in 14 tumours (8.8%). Concordance between both events was found in 12 of 14 amplified cases (P = 0.0001). We found nine JAK2-amplified cases (5.7%), seven with PD-L1 expression (P = 0.0006). Moreover, six of the seven cases had JAK2 and CD274 coamplification (9p24.1 genomic amplification). Remarkably, the average PD-L1 IHC score was higher in these amplified cases (230 versus 80; P = 0.001). Non-statistical associations were observed between PD-L1 expression and PTEN loss and PTEN deletions.

CONCLUSIONS

We describe a subset of patients (8.2%) who had 9p24.1 amplifications resulting in high expression of PD-L1. Our results provide evidence for genomic up-regulation of PD-L1 expression in non-small-cell lung cancer.

Heterogeneity of Tumor and Immune Cell PD-L1 Expression and Lymphocyte Counts in Surgical NSCLC Samples

BACKGROUND

Immune-checkpoint inhibitors against programmed cell death protein 1 (PD-1)/programmed death-ligand 1 (PD-L1) have shown remarkable therapeutic activity in non-small-cell lung cancer (NSCLC). However, biomarker-based patient selection remains a challenge. Our aim was to assess the heterogeneity of various immune markers between different tumor areas of surgically resected NSCLC specimens.

MATERIALS AND METHODS

We included 94 adenocarcinoma (ADC) and 50 squamous cell carcinoma (SCC) specimens. Two distinct tumor areas of each tumor sample were selected and incorporated into tissue microarrays. PD-L1 expression in tumor cells (TCs) and immune cells (ICs) was assessed using clone SP142 (Ventana). PDL1 gene amplification was assessed using fluorescence in situ hybridization. CD3 and CD8 densities were quantified using digital image-based analysis. Heterogeneity was assessed using kappa agreement index (KI) and intraclass correlation coefficient.

RESULTS

Prevalence of PD-L1 expression was 16.8% in TCs and 27.8% in ICs. Eleven tumors (7.6%) showed PDL1 amplification. In ADC, KI of PD-L1 TC and IC expression between cores was 0.465 and 0.260, compared with 0.274 and 0.124 in SCC, respectively. Higher concordance was observed for PDL1 amplification (KI, 0.647 in ADC and KI, 1 in SCC). Eleven (61.1%) of 18 amplified cores showed PD-L1 staining in < 5% of TCs. Intraclass correlation coefficients for CD3 and CD8 were 0.293 and 0.186 in ADC and 0.489 and 0.610 in SCC samples, respectively.

CONCLUSIONS

We found significant heterogeneity of PD-L1 expression in both ADC and SCC samples, especially in the IC compartment. Heterogeneous expression of PD-L1 could misclassify patients for PD-1/PD-L1-directed therapies.

Dabrafenib in an elderly patient with metastatic melanoma and BRAF V600R mutation: a case report

Background

Approximately 50 % of malignant melanomas harbor activating point mutations in the BRAF gene. Typically, these mutations result in the substitution of the amino acid valine at codon 600 of the gene, and 90–95 % of mutations are either BRAF^V600E or BRAF^V600K. Specific BRAF inhibitors such as dabrafenib and vemurafenib are the mainstays of treatment in patients with metastatic BRAF-mutant malignant melanomas. The third most common BRAF mutation is V600R, which also leads to increased BRAF signaling. Although evidence exists about the activity of dabrafenib and vemurafenib in patients with the BRAF^V600R mutation, these patients have been systematically excluded from recent trials with targeted therapies.

Case presentation

Here, we report the positive results in terms of survival and quality of life obtained with dabrafenib in an 80-year-old Caucasian male patient with a Charlson Comorbidity Index of 8 diagnosed with metastatic malignant melanoma harboring the BRAF^V600R mutation. Our patient was treated with dabrafenib for 7 months with minimal toxicity. We also report exploratory analyses of circulating tumor DNA during targeted treatment. Interestingly, the mutation was not detected after starting treatment and became detectable before radiological disease progression.

Conclusions

Our report suggests that (1) a relevant benefit can be obtained with a BRAF inhibitor in real-world patients with a malignant melanoma harboring a BRAF^V600R mutation, and that (2) circulating tumor DNA detection might be of help in assessing tumor burden in everyday clinical practice. The results reported here should encourage the inclusion of patients with BRAF^V600R-mutated malignant melanomas in future prospective clinical trials with BRAF inhibitors.

MET expression and copy number heterogeneity in nonsquamous non-small cell lung cancer (nsNSCLC)

OBJECTIVE

We aimed to assess MET intratumoral heterogeneity and its potential impact on biomarker-based patient selection as well as potential surrogate biomarkers of MET activation.

METHODS

Our study included 120 patients with non-squamous Non-small-cell Lung Cancer (nsNSCLC), of which 47 were incorporated in tissue microarrays (TMA). Four morphologically distinct tumor areas were selected to assess MET heterogeneity. MET positivity by immunohistochemistry (IHC) was defined as an above-median H-score and by +2/+3 staining intensity in >50% of tumor cells (Metmab criteria). MET FISH positivity was defined by MET/CEP7 ratio ≥ 2.0 and/or MET ≥ 5.0. MET staining pattern (cytoplasmic vs. membranous) and mesenchymal markers were investigated as surrogates of MET activation.

RESULTS

Median MET H-score was 140 (range 0-400) and 47.8% of patients were MET positive by Metmab criteria. Eight cases (6.8%) were MET FISH positive and showed higher H-scores (p = 0.021). MET positivity by IHC changed in up to 40% of cases among different tumor areas, and MET amplification in 25-50%. Cytoplasmic MET staining and positivity for vimentin predicted poor survival (p = 0.042 and 0.047, respectively).

CONCLUSIONS

MET status is highly heterogeneous among different nsNSCLC tumor areas, hindering adequate patient selection for MET-targeted therapies. MET cytoplasmic staining and vimentin might represent surrogate markers for MET activation.

Is there a role for epidermal growth factor receptor tyrosine kinase inhibitors in epidermal growth factor receptor wild-type non-small cell lung cancer?

Non-small cell lung cancer (NSCLC) is the most common type of lung cancer with a world-wide annual incidence of around 1.3 million. The majority of patients are diagnosed with advanced disease and survival remains poor. However, relevant advances have occurred in recent years through the identification of biomarkers that predict for benefit of therapeutic agents. This is exemplified by the efficacy of epidermal growth factor receptor (EGFR) tyrosine kinase inhibitors for the treatment of EGFR mutant patients. These drugs have also shown efficacy in unselected populations but this point remains controversial. Here we have reviewed the clinical data that demonstrate a small but consistent subgroup of EGFR wild-type patients with NSCLC that obtain a clinical benefit from these drugs. Moreover, we review the biological rationale that may explain this benefit observed in the clinical setting.