Refine
Year of publication
Document Type
- Article (59)
- Conference Proceeding (2)
- Preprint (1)
Language
- English (62)
Has Fulltext
- yes (62)
Is part of the Bibliography
- no (62) (remove)
Keywords
- glioblastoma (16)
- glioma (4)
- hypoxia (4)
- meningioma (4)
- Glioblastoma (3)
- MRI (3)
- Brain metastasis (2)
- DNA methylation (2)
- EMT (2)
- Fasting (2)
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.
Background: Epilepsy surgery is an established treatment for drug-resistant focal epilepsy (DRFE) that results in seizure freedom in about 60% of patients. Correctly identifying an epileptogenic lesion in magnetic resonance imaging (MRI) is challenging but highly relevant since it improves the likelihood of being referred for presurgical diagnosis. The epileptogenic lesion’s etiology directly relates to the surgical intervention’s indication and outcome. Therefore, it is vital to correctly identify epileptogenic lesions and their etiology presurgically.
Methods: We compared the final histopathological diagnoses of all patients with DRFE undergoing epilepsy surgery at our center between 2015 and 2021 with their MRI diagnoses before and after presurgical diagnosis at our epilepsy center, including MRI evaluations by expert epilepsy neuroradiologists. Additionally, we analyzed the outcome of different subgroups.
Results: This study included 132 patients. The discordance between histopathology and MRI diagnoses significantly decreased from 61.3% for non-expert MRI evaluations (NEMRIs) to 22.1% for epilepsy center MRI evaluations (ECMRIs; p < 0.0001). The MRI-sensitivity improved significantly from 68.6% for NEMRIs to 97.7% for ECMRIs (p < 0.0001). Identifying focal cortical dysplasia (FCD) and amygdala dysplasia was the most challenging for both subgroups. 65.5% of patients with negative NEMRI were seizure-free 12 months postoperatively, no patient with negative ECMRI achieved seizure-freedom. The mean duration of epilepsy until surgical intervention was 13.6 years in patients with an initial negative NEMRI and 9.5 years in patients with a recognized lesion in NEMRI.
Conclusions: This study provides evidence that for patients with DRFE—especially those with initial negative findings in a non-expert MRI—an early consultation at an epilepsy center, including an ECMRI, is important for identifying candidates for epilepsy surgery. NEMRI-negative findings preoperatively do not preclude seizure freedom postoperatively. Therefore, patients with DRFE that remain MRI-negative after initial NEMRI should be referred to an epilepsy center for presurgical evaluation. Nonreferral based on NEMRI negativity may harm such patients and delay surgical intervention. However, ECMRI-negative patients have a reduced chance of becoming seizure-free after epilepsy surgery. Further improvements in MRI technique and evaluation are needed and should be directed towards improving sensitivity for FCDs and amygdala dysplasias.