Refine
Document Type
- Article (8)
Language
- English (8)
Has Fulltext
- yes (8)
Is part of the Bibliography
- no (8)
Keywords
- Machine learning (2)
- Prostate cancer (2)
- Artificial intelligence (1)
- Breast cancer (1)
- Cancer detection and diagnosis (1)
- Capecitabine (1)
- Clinical trial (1)
- EGFL7 (1)
- EP300 (1)
- EWSR1 (1)
Institute
- Medizin (8)
- Informatik und Mathematik (2)
Purpose: The PELICAN trial evaluates for the first time efficacy and safety of pegylated liposomal doxorubicin (PLD) versus capecitabine as first-line treatment of metastatic breast cancer (MBC).
Methods: This randomized, phase III, open-label, multicenter trial enrolled first-line MBC patients who were ineligible for endocrine or trastuzumab therapy. Cumulative adjuvant anthracyclines of 360 mg/m2 doxorubicin or equivalent were allowed. Left ventricular ejection fraction of >50 % was required. Patients received PLD 50 mg/m2 every 28 days or capecitabine 1250 mg/m2 twice daily for 14 days every 21 days. The primary endpoint was time-to-disease progression (TTP).
Results: 210 patients were randomized (n = 105, PLD and n = 105, capecitabine). Adjuvant anthracyclines were given to 37 % (PLD) and 36 % (capecitabine) of patients. No significant difference was observed in TTP [HR = 1.21 (95 % confidence interval, 0.838–1.750)]. Median TTP was 6.0 months for both PLD and capecitabine. Comparing patients with or without prior anthracyclines, no significant difference in TTP was observed in the PLD arm (log-rank P = 0.64). For PLD versus capecitabine, respectively, overall survival (median, 23.3 months vs. 26.8 months) and time-to-treatment failure (median, 4.6 months vs. 3.7 months) were not statistically significantly different. Compared to PLD, patients on capecitabine experienced more serious adverse events (P = 0.015) and more cardiac events among patients who had prior anthracycline exposure (18 vs. 8 %; P = 0.31).
Conclusion: Both PLD and capecitabine are effective first-line agents for MBC.
DNA methylation-based prediction of response to immune checkpoint inhibition in metastatic melanoma
(2021)
Background: Therapies based on targeting immune checkpoints have revolutionized the treatment of metastatic melanoma in recent years. Still, biomarkers predicting long-term therapy responses are lacking. Methods: A novel approach of reference-free deconvolution of large-scale DNA methylation data enabled us to develop a machine learning classifier based on CpG sites, specific for latent methylation components (LMC), that allowed for patient allocation to prognostic clusters. DNA methylation data were processed using reference-free analyses (MeDeCom) and reference-based computational tumor deconvolution (MethylCIBERSORT, LUMP). Results: We provide evidence that DNA methylation signatures of tumor tissue from cutaneous metastases are predictive for therapy response to immune checkpoint inhibition in patients with stage IV metastatic melanoma. Conclusions: These results demonstrate that LMC-based segregation of large-scale DNA methylation data is a promising tool for classifier development and treatment response estimation in cancer patients under targeted immunotherapy.
Background: Identification of families at risk for ovarian cancer offers the opportunity to consider prophylactic surgery thus reducing ovarian cancer mortality. So far, identification of potentially affected families in Germany was solely performed via family history and numbers of affected family members with breast or ovarian cancer. However, neither the prevalence of deleterious variants in BRCA1/2 in ovarian cancer in Germany nor the reliability of family history as trigger for genetic counselling has ever been evaluated.
Methods: Prospective counseling and germline testing of consecutive patients with primary diagnosis or with platinum-sensitive relapse of an invasive epithelial ovarian cancer. Testing included 25 candidate and established risk genes. Among these 25 genes, 16 genes (ATM, BRCA1, BRCA2, CDH1, CHEK2, MLH1, MSH2, MSH6, NBN, PMS2, PTEN, PALB2, RAD51C, RAD51D, STK11, TP53) were defined as established cancer risk genes. A positive family history was defined as at least one relative with breast cancer or ovarian cancer or breast cancer in personal history.
Results: In total, we analyzed 523 patients: 281 patients with primary diagnosis of ovarian cancer and 242 patients with relapsed disease. Median age at primary diagnosis was 58 years (range 16–93) and 406 patients (77.6%) had a high-grade serous ovarian cancer. In total, 27.9% of the patients showed at least one deleterious variant in all 25 investigated genes and 26.4% in the defined 16 risk genes. Deleterious variants were most prevalent in the BRCA1 (15.5%), BRCA2 (5.5%), RAD51C (2.5%) and PALB2 (1.1%) genes. The prevalence of deleterious variants did not differ significantly between patients at primary diagnosis and relapse. The prevalence of deleterious variants in BRCA1/2 (and in all 16 risk genes) in patients <60 years was 30.2% (33.2%) versus 10.6% (18.9%) in patients ≥60 years. Family history was positive in 43% of all patients. Patients with a positive family history had a prevalence of deleterious variants of 31.6% (36.0%) versus 11.4% (17.6%) and histologic subtype of high grade serous ovarian cancer versus other showed a prevalence of deleterious variants of 23.2% (29.1%) and 10.2% (14.8%), respectively. Testing only for BRCA1/2 would miss in our series more than 5% of the patients with a deleterious variant in established risk genes.
Conclusions: 26.4% of all patients harbor at least one deleterious variant in established risk genes. The threshold of 10% mutation rate which is accepted for reimbursement by health care providers in Germany was observed in all subgroups analyzed and neither age at primary diagnosis nor histo-type or family history sufficiently enough could identify a subgroup not eligible for genetic counselling and testing. Genetic testing should therefore be offered to every patient with invasive epithelial ovarian cancer and limiting testing to BRCA1/2 seems to be not sufficient.
Objective: Advanced or recurrent endometrial cancer (EC) no longer amenable to surgery or radiotherapy is a life-threatening disease with limited therapeutic options left. Eighty percent of ECs express receptors for luteinizing hormone-releasing hormone (LHRH), which can be targeted by AEZS-108 (zoptarelin doxorubicin acetate). This phase 2 trial was performed to assess the efficacy and safety of AEZS-108 in this group of patients.
Methods: Patients had FIGO (Fédération Internationale de Gynécologie et d'Obstétrique) III or IV or recurrent EC, LHRH receptor-positive tumor status, and at least had 1 measurable lesion (Response Evaluation Criteria in Solid Tumors). Prior anthracycline therapy was not allowed. Patients received AEZS-108 as a 2-hour infusion on day 1 of a 21-day cycle. The treatment was continued for a maximum of 6 to 8 cycles. The primary end point was the response rate determined by the Response Evaluation Criteria in Solid Tumors.
Results: From April 2008 to November 2009, 44 patients were included in the study at 8 centers in Germany (AGO) and 3 centers in Bulgaria. Forty-three of these patients were eligible. Two (5%) patients had a complete remission, and 8 (18%) achieved a partial remission. Stable disease for at least 6 weeks was observed in 44%. The median time to progression was 7 months, and the median overall survival was 15 months. The most frequently reported grade 3 or 4 adverse effects were neutropenia (12%) and leucopenia (9%).
Conclusions: AEZS-108, an LHRH-agonist coupled to doxorubicin, has significant activity and low toxicity in women with advanced or recurrent LHRH receptor-positive EC, supporting the principle of receptor-mediated targeted chemotherapy.
EGFL7 enhances surface expression of integrin α5β1 to promote angiogenesis in malignant brain tumors
(2018)
Glioblastoma (GBM) is a typically lethal type of brain tumor with a median survival of 15 months postdiagnosis. This negative prognosis prompted the exploration of alternative treatment options. In particular, the reliance of GBM on angiogenesis triggered the development of anti‐VEGF (vascular endothelial growth factor) blocking antibodies such as bevacizumab. Although its application in human GBM only increased progression‐free periods but did not improve overall survival, physicians and researchers still utilize this treatment option due to the lack of adequate alternatives. In an attempt to improve the efficacy of anti‐VEGF treatment, we explored the role of the egfl7 gene in malignant glioma. We found that the encoded extracellular matrix protein epidermal growth factor‐like protein 7 (EGFL7) was secreted by glioma blood vessels but not glioma cells themselves, while no major role could be assigned to the parasitic miRNAs miR‐126/126*. EGFL7 expression promoted glioma growth in experimental glioma models in vivo and stimulated tumor vascularization. Mechanistically, this was mediated by an upregulation of integrin α5β1 on the cellular surface of endothelial cells, which enhanced fibronectin‐induced angiogenic sprouting. Glioma blood vessels that formed in vivo were more mature as determined by pericyte and smooth muscle cell coverage. Furthermore, these vessels were less leaky as measured by magnetic resonance imaging of extravasating contrast agent. EGFL7‐inhibition using a specific blocking antibody reduced the vascularization of experimental gliomas and increased the life span of treated animals, in particular in combination with anti‐VEGF and the chemotherapeutic agent temozolomide. Data allow for the conclusion that this combinatorial regimen may serve as a novel treatment option for GBM.
Ependymomas encompass a heterogeneous group of central nervous system (CNS) neoplasms that occur along the entire neuroaxis. In recent years, extensive (epi-)genomic profiling efforts have identified several molecular groups of ependymoma that are characterized by distinct molecular alterations and/or patterns. Based on unsupervised visualization of a large cohort of genome-wide DNA methylation data, we identified a highly distinct group of pediatric-type tumors (n = 40) forming a cluster separate from all established CNS tumor types, of which a high proportion were histopathologically diagnosed as ependymoma. RNA sequencing revealed recurrent fusions involving the pleomorphic adenoma gene-like 1 (PLAGL1) gene in 19 of 20 of the samples analyzed, with the most common fusion being EWSR1:PLAGL1 (n = 13). Five tumors showed a PLAGL1:FOXO1 fusion and one a PLAGL1:EP300 fusion. High transcript levels of PLAGL1 were noted in these tumors, with concurrent overexpression of the imprinted genes H19 and IGF2, which are regulated by PLAGL1. Histopathological review of cases with sufficient material (n = 16) demonstrated a broad morphological spectrum of tumors with predominant ependymoma-like features. Immunohistochemically, tumors were GFAP positive and OLIG2- and SOX10 negative. In 3/16 of the cases, a dot-like positivity for EMA was detected. All tumors in our series were located in the supratentorial compartment. Median age of the patients at the time of diagnosis was 6.2 years. Median progression-free survival was 35 months (for 11 patients with data available). In summary, our findings suggest the existence of a novel group of supratentorial neuroepithelial tumors that are characterized by recurrent PLAGL1 fusions and enriched for pediatric patients.
Background: Prostate cancer is a major health concern in aging men. Paralleling an aging society, prostate cancer prevalence increases emphasizing the need for efcient diagnostic algorithms.
Methods: Retrospectively, 106 prostate tissue samples from 48 patients (mean age,
66 ± 6.6 years) were included in the study. Patients sufered from prostate cancer (n = 38) or benign prostatic hyperplasia (n = 10) and were treated with radical prostatectomy or Holmium laser enucleation of the prostate, respectively. We constructed tissue microarrays (TMAs) comprising representative malignant (n = 38) and benign (n = 68) tissue cores. TMAs were processed to histological slides, stained, digitized and assessed for the applicability of machine learning strategies and open–source tools in diagnosis of prostate cancer. We applied the software QuPath to extract features for shape, stain intensity, and texture of TMA cores for three stainings, H&E, ERG, and PIN-4. Three machine learning algorithms, neural network (NN), support vector machines (SVM), and random forest (RF), were trained and cross-validated with 100 Monte Carlo random splits into 70% training set and 30% test set. We determined AUC values for single color channels, with and without optimization of hyperparameters by exhaustive grid search. We applied recursive feature elimination to feature sets of multiple color transforms.
Results: Mean AUC was above 0.80. PIN-4 stainings yielded higher AUC than H&E and
ERG. For PIN-4 with the color transform saturation, NN, RF, and SVM revealed AUC of 0.93 ± 0.04, 0.91 ± 0.06, and 0.92 ± 0.05, respectively. Optimization of hyperparameters improved the AUC only slightly by 0.01. For H&E, feature selection resulted in no increase of AUC but to an increase of 0.02–0.06 for ERG and PIN-4.
Conclusions: Automated pipelines may be able to discriminate with high accuracy between malignant and benign tissue. We found PIN-4 staining best suited for classifcation. Further bioinformatic analysis of larger data sets would be crucial to evaluate the reliability of automated classifcation methods for clinical practice and to evaluate potential discrimination of aggressiveness of cancer to pave the way to automatic precision medicine.
Objectives: To analyze the performance of radiological assessment categories and quantitative computational analysis of apparent diffusion coefficient (ADC) maps using variant machine learning algorithms to differentiate clinically significant versus insignificant prostate cancer (PCa). Methods: Retrospectively, 73 patients were included in the study. The patients (mean age, 66.3 ± 7.6 years) were examined with multiparametric MRI (mpMRI) prior to radical prostatectomy (n = 33) or targeted biopsy (n = 40). The index lesion was annotated in MRI ADC and the equivalent histologic slides according to the highest Gleason Grade Group (GrG). Volumes of interest (VOIs) were determined for each lesion and normal-appearing peripheral zone. VOIs were processed by radiomic analysis. For the classification of lesions according to their clinical significance (GrG ≥ 3), principal component (PC) analysis, univariate analysis (UA) with consecutive support vector machines, neural networks, and random forest analysis were performed. Results: PC analysis discriminated between benign and malignant prostate tissue. PC evaluation yielded no stratification of PCa lesions according to their clinical significance, but UA revealed differences in clinical assessment categories and radiomic features. We trained three classification models with fifteen feature subsets. We identified a subset of shape features which improved the diagnostic accuracy of the clinical assessment categories (maximum increase in diagnostic accuracy ΔAUC = + 0.05, p < 0.001) while also identifying combinations of features and models which reduced overall accuracy. Conclusions: The impact of radiomic features to differentiate PCa lesions according to their clinical significance remains controversial. It depends on feature selection and the employed machine learning algorithms. It can result in improvement or reduction of diagnostic performance.