Deep learning automatic semantic segmentation of glioblastoma multiforme regions on multimodal magnetic resonance images.

OBJECTIVES: In patients having naïve glioblastoma multiforme (GBM), this study aims to assess the efficacy of Deep Learning algorithms in automating the segmentation of brain magnetic resonance (MR) images to accurately determine 3D masks for 4 distinct regions: enhanced tumor, peritumoral edema, non-enhanced/necrotic tumor, and total tumor. MATERIAL AND METHODS: A 3D U-Net neural network algorithm was developed for semantic segmentation of GBM. The training dataset was manually delineated by a group of expert neuroradiologists on MR images from the Brain Tumor Segmentation Challenge 2021 (BraTS2021) image repository, as ground truth labels for diverse glioma (GBM and low-grade glioma) subregions across four MR sequences (T1w, T1w-contrast enhanced, T2w, and FLAIR) in 1251 patients. The in-house test was performed on 50 GBM patients from our cohort (PerProGlio project). By exploring various hyperparameters, the network's performance was optimized, and the most optimal parameter configuration was identified. The assessment of the optimized network's performance utilized Dice scores, precision, and sensitivity metrics. RESULTS: Our adaptation of the 3D U-net with additional residual blocks demonstrated reliable performance on both the BraTS2021 dataset and the in-house PerProGlio cohort, employing only T1w-ce sequences for enhancement and non-enhanced/necrotic tumor models and T1w-ce + T2w + FLAIR for peritumoral edema and total tumor. The mean Dice scores (training and test) were 0.89 and 0.75; 0.75 and 0.64; 0.79 and 0.71; and 0.60 and 0.55, for total tumor, edema, enhanced tumor, and non-enhanced/necrotic tumor, respectively. CONCLUSIONS: The results underscore the high precision with which our network can effectively segment GBM tumors and their distinct subregions. The level of accuracy achieved agrees with the coefficients recorded in previous GBM studies. In particular, our approach allows model specialization for each of the different tumor subregions employing only those MR sequences that provide value for segmentation.

BCAS1 defines a heterogeneous cell population in diffuse gliomas.

Oligodendrocyte precursor markers have become of great interest to identify new diagnostic and therapeutic targets for diffuse gliomas, since state-of-the-art studies point towards immature oligodendrocytes as a possible source of gliomagenesis. Brain enriched myelin associated protein 1 (BCAS1) is a novel marker of immature oligodendrocytes and was proposed to contribute to tumorigenesis in non-central nervous system tumors. However, BCAS1 role in diffuse glioma is still underexplored. This study analyzes the expression of BCAS1 in different tumor samples from patients with diffuse gliomas (17 oligodendrogliomas; 8 astrocytomas; 60 glioblastomas) and uncovers the molecular and ultrastructural features of BCAS1+ cells by immunostaining and electron microscopy. Our results show that BCAS1+ cells exhibit stellate or spherical morphology with similar ultrastructural features. Stellate and spherical cells were detected as isolated cells in all studied gliomas. Nevertheless, only stellate cells were found to be proliferative and formed tightly packed nodules with a highly proliferative rate in oligodendrogliomas. Our findings provide a comprehensive characterization of the BCAS1+ cell population within diffuse gliomas. The observed proliferative capacity and distribution of BCAS1+ stellate cells, particularly in oligodendrogliomas, highlight BCAS1 as an interesting marker, warranting further investigation into its role in tumor malignancy.

Proteometabolomics of initial and recurrent glioblastoma highlights an increased immune cell signature with altered lipid metabolism.

BACKGROUND: There is an urgent need to better understand the mechanisms associated with the development, progression, and onset of recurrence after initial surgery in glioblastoma (GBM). The use of integrative phenotype-focused -omics technologies such as proteomics and lipidomics provides an unbiased approach to explore the molecular evolution of the tumor and its associated environment. METHODS: We assembled a cohort of patient-matched initial (iGBM) and recurrent (rGBM) specimens of resected GBM. Proteome and metabolome composition were determined by mass spectrometry-based techniques. We performed neutrophil-GBM cell coculture experiments to evaluate the behavior of rGBM-enriched proteins in the tumor microenvironment. ELISA-based quantitation of candidate proteins was performed to test the association of their plasma concentrations in iGBM with the onset of recurrence. RESULTS: Proteomic profiles reflect increased immune cell infiltration and extracellular matrix reorganization in rGBM. ASAH1, SYMN, and GPNMB were highly enriched proteins in rGBM. Lipidomics indicates the downregulation of ceramides in rGBM. Cell analyses suggest a role for ASAH1 in neutrophils and its localization in extracellular traps. Plasma concentrations of ASAH1 and SYNM show an association with time to recurrence. CONCLUSIONS: We describe the potential importance of ASAH1 in tumor progression and development of rGBM via metabolic rearrangement and showcase the feedback from the tumor microenvironment to plasma proteome profiles. We report the potential of ASAH1 and SYNM as plasma markers of rGBM progression. The published datasets can be considered as a resource for further functional and biomarker studies involving additional -omics technologies.

Local detection of microvessels in IDH-wildtype glioblastoma using relative cerebral blood volume: an imaging marker useful for astrocytoma grade 4 classification.

BACKGROUND: The microvessels area (MVA), derived from microvascular proliferation, is a biomarker useful for high-grade glioma classification. Nevertheless, its measurement is costly, labor-intense, and invasive. Finding radiologic correlations with MVA could provide a complementary non-invasive approach without an extra cost and labor intensity and from the first stage. This study aims to correlate imaging markers, such as relative cerebral blood volume (rCBV), and local MVA in IDH-wildtype glioblastoma, and to propose this imaging marker as useful for astrocytoma grade 4 classification. METHODS: Data from 73 tissue blocks belonging to 17 IDH-wildtype glioblastomas and 7 blocks from 2 IDH-mutant astrocytomas were compiled from the Ivy GAP database. MRI processing and rCBV quantification were carried out using ONCOhabitats methodology. Histologic and MRI co-registration was done manually with experts' supervision, achieving an accuracy of 88.8% of overlay. Spearman's correlation was used to analyze the association between rCBV and microvessel area. Mann-Whitney test was used to study differences of rCBV between blocks with presence or absence of microvessels in IDH-wildtype glioblastoma, as well as to find differences with IDH-mutant astrocytoma samples. RESULTS: Significant positive correlations were found between rCBV and microvessel area in the IDH-wildtype blocks (p < 0.001), as well as significant differences in rCBV were found between blocks with microvascular proliferation and blocks without it (p < 0.0001). In addition, significant differences in rCBV were found between IDH-wildtype glioblastoma and IDH-mutant astrocytoma samples, being 2-2.5 times higher rCBV values in IDH-wildtype glioblastoma samples. CONCLUSIONS: The proposed rCBV marker, calculated from diagnostic MRIs, can detect in IDH-wildtype glioblastoma those regions with microvessels from those without it, and it is significantly correlated with local microvessels area. In addition, the proposed rCBV marker can differentiate the IDH mutation status, providing a complementary non-invasive method for high-grade glioma classification.

Nests of dividing neuroblasts sustain interneuron production for the developing human brain.

The human cortex contains inhibitory interneurons derived from the medial ganglionic eminence (MGE), a germinal zone in the embryonic ventral forebrain. How this germinal zone generates sufficient interneurons for the human brain remains unclear. We found that the human MGE (hMGE) contains nests of proliferative neuroblasts with ultrastructural and transcriptomic features that distinguish them from other progenitors in the hMGE. When dissociated hMGE cells are transplanted into the neonatal mouse brain, they reform into nests containing proliferating neuroblasts that generate young neurons that migrate extensively into the mouse forebrain and mature into different subtypes of functional interneurons. Together, these results indicate that the nest organization and sustained proliferation of neuroblasts in the hMGE provide a mechanism for the extended production of interneurons for the human forebrain.

Evaluation of the optical criteria for sessile serrated lesions of the colon: A prospective study on a colorectal cancer screening population.

Background and study aims We aimed to describe the presence and combination of Hazewinkel's optical diagnosis (OD) criteria for sessile serrated lesions (SSL), determining which lesion characteristics increase the probability of a correct OD, with a focus on diminutive lesions. Patients and methods This was a prospective study describing the presence of Hazewinkel's OD criteria for SSL in lesions found in consecutive CRC screening colonoscopies. The presence of each OD criterion and their diagnostic combinations in SSL, related to the lesion's NBI International Colorectal Endoscopic (NICE) classification category, size, and location, were described. The presence of two or more optical criteria was considered diagnostic of SSL. The OD was compared to pathology as the gold standard. Results Seventy-nine SSLs (5.6 %) were diagnosed. Cloud-like appearance was the most prevalent OD criterion (35, 44.3 %). OD criteria were more frequently identified in NICE type 1, ≥ 10 mm, and proximal lesions. Only 26 SLLs fulfilled the OD criteria (sensitivity 32.9 %, 95 % CI 29.1 %-36.7 %). The sensitivity for diminutive SSL was 14.7 %, (95 % CI 11.9 %-17.6 %). Eighty-five lesions were optically diagnosed as SSL. However, only in 26 SSL was this the definitive diagnosis (positive predictive value 30.6 %, 95 % CI 26.9 %-34.3 %). Size > 5 mm and proximal location increased the probability of a correct diagnosis. The overall accuracy of the optical criteria was 92.0 % (95 % CI, 89.8 %-94.2 %). Conclusions The Hazewinkel's optical criteria are not reliable for a positive diagnosis of SSL, particularly for diminutive lesions.

[Gastric gastrointestinal stromal tumour and schwannoma: An unusual presentation in a patient without neurofibromatosis]. / Tumor del estroma gastrointestinal y schwannoma gástricos: una presentación inusual en un paciente sin neurofibromatosis.

The majority of gastric neoplasms are of epithelial type. Stromal tumours are the next most frequent and are most commonly gastrointestinal stromal tumours, followed by leiomyoma and schwannoma. We present an exceptional case of a patient with a gastrointestinal stromal tumour with suspicion of residual gastric disease, which was diagnosed post-operatively as a schwannoma.

MicroRNA profile in very young women with breast cancer.

BACKGROUND: Breast cancer is rarely diagnosed in very young women (35 years old or younger), and it often presents with distinct clinical-pathological features related to a more aggressive phenotype and worse prognosis when diagnosed at this early age. A pending question is whether breast cancer in very young women arises from the deregulation of different underlying mechanisms, something that will make this disease an entity differentiated from breast cancer diagnosed in older patients. METHODS: We performed a comprehensive study of miRNA expression using miRNA Affymetrix2.0 array on paraffin-embedded tumour tissue of 42 breast cancer patients 35 years old or younger, 17 patients between 45 and 65 years old and 29 older than 65 years. Data were statistically analyzed by t-test and a hierarchical clustering via average linkage method was conducted. Results were validated by qRT-PCR. Putative targeted pathways were obtained using DIANA miRPath online software. RESULTS: The results show a differential and unique miRNA expression profile of 121 miRNAs (p-value <0.05), 96 of those with a FDR-value <0.05. Hierarchical clustering grouped the samples according to their age, but not by subtype nor by tumour characteristics. We were able to validate by qRT-PCR differences in the expression of 6 miRNAs: miR-1228*, miR-3196, miR-1275, miR-92b, miR-139 and miR-1207. Moreover, all of the miRNAs maintained the expression trend. The validated miRNAs pointed out pathways related to cell motility, invasion and proliferation. CONCLUSIONS: The study suggests that breast cancer in very young women appears as a distinct molecular signature. To our knowledge, this is the first time that a validated microRNA profile, distinctive to breast cancer in very young women, has been presented. The miRNA signature may be relevant to open an important field of research in order to elucidate the underlying mechanism in this particular disease, which in a more clinical setting, could potentially help to identify therapeutic targets in this particular set of patients.

Concordance of genomic alterations between primary and recurrent breast cancer.

There is growing interest in delivering genomically informed cancer therapy. Our aim was to determine the concordance of genomic alterations between primary and recurrent breast cancer. Targeted next-generation sequencing was performed on formalin-fixed paraffin-embedded (FFPE) samples, profiling 3,320 exons of 182 cancer-related genes plus 37 introns from 14 genes often rearranged in cancer. Point mutations, indels, copy-number alterations (CNA), and select rearrangements were assessed in 74 tumors from 43 patients (36 primary and 38 recurrence/metastases). Alterations potentially targetable with established or investigational therapeutics were considered "actionable." Alterations were detected in 55 genes (mean 3.95 alterations/sample, range 1-12), including mutations in PIK3CA, TP53, ARID1A, PTEN, AKT1, NF1, FBXW7, and FGFR3 and amplifications in MCL1, CCND1, FGFR1, MYC, IGF1R, MDM2, MDM4, AKT3, CDK4, and AKT2. In 33 matched primary and recurrent tumors, 97 of 112 (86.6%) somatic mutations were concordant. Of identified CNAs, 136 of 159 (85.5%) were concordant: 37 (23.3%) were concordant, but below the reporting threshold in one of the matched samples, and 23 (14.5%) discordant. There was an increased frequency of CDK4/MDM2 amplifications in recurrences, as well as gains and losses of other actionable alterations. Forty of 43 (93%) patients had actionable alterations that could inform targeted treatment options. In conclusion, deep genomic profiling of cancer-related genes reveals potentially actionable alterations in most patients with breast cancer. Overall there was high concordance between primary and recurrent tumors. Analysis of recurrent tumors before treatment may provide additional insights, as both gains and losses of targets are observed.

Emergence of constitutively active estrogen receptor-α mutations in pretreated advanced estrogen receptor-positive breast cancer.

PURPOSE: We undertook this study to determine the prevalence of estrogen receptor (ER) α (ESR1) mutations throughout the natural history of hormone-dependent breast cancer and to delineate the functional roles of the most commonly detected alterations. EXPERIMENTAL DESIGN: We studied a total of 249 tumor specimens from 208 patients. The specimens include 134 ER-positive (ER(+)/HER2(-)) and, as controls, 115 ER-negative (ER(-)) tumors. The ER(+) samples consist of 58 primary breast cancers and 76 metastatic samples. All tumors were sequenced to high unique coverage using next-generation sequencing targeting the coding sequence of the estrogen receptor and an additional 182 cancer-related genes. RESULTS: Recurring somatic mutations in codons 537 and 538 within the ligand-binding domain of ER were detected in ER(+) metastatic disease. Overall, the frequency of these mutations was 12% [9/76; 95% confidence interval (CI), 6%-21%] in metastatic tumors and in a subgroup of patients who received an average of 7 lines of treatment the frequency was 20% (5/25; 95% CI, 7%-41%). These mutations were not detected in primary or treatment-naïve ER(+) cancer or in any stage of ER(-) disease. Functional studies in cell line models demonstrate that these mutations render estrogen receptor constitutive activity and confer partial resistance to currently available endocrine treatments. CONCLUSIONS: In this study, we show evidence for the temporal selection of functional ESR1 mutations as potential drivers of endocrine resistance during the progression of ER(+) breast cancer.

High stability of microRNAs in tissue samples of compromised quality.

Degradation of tissue samples limits performing RNA-based molecular studies, but little is known about the potential usefulness of samples of compromised quality for studies focused on miRNAs. In this work we analyze a series of cryopreserved tissue samples (n = 14), frozen samples that underwent a severe thawing process (n = 10), and their paired formalin-fixed paraffin-embedded (FFPE) tissue samples (n = 24) from patients with breast cancer obtained during primary surgical resection and collected in 2011. Quality and integrity analyses of the total and small fraction of RNA were carried out. Recovery of specific RNA molecules (miRNAs hsa-miR-21, hsa-miR-125b, and hsa-miR-191; snoRNA RNU6B; and mRNAs GAPDH and HPRT1) was also analyzed by quantitative RT-PCR. Our results suggest that visualisation of the small RNA electrophoretic profiles obtained using the Agilent 2100 bioanalyzer makes it possible to differentiate between the three groups of samples (optimally frozen, thawed, and FFPE). We demonstrate that specific miRNA molecules can be similarly recovered from different tissue sample sources, which supports their high degree of stability. We conclude that miRNAs are robustly detected irrespective of the quality of the tissue sample. In this regard, a word of caution should be raised before degraded samples are discarded: although prior quality assessment of the biological material to be analyzed is recommended, our work demonstrates that degraded tissue samples are also suitable for miRNA studies.

High ERK protein expression levels correlate with shorter survival in triple-negative breast cancer patients.

The mitogen-activated protein kinase (MAPK) signaling pathway is known to be activated in triple-negative breast cancer (TNBC). Extracellular signal-related kinase (ERK), a member of the MAPK pathway, promotes cell proliferation, angiogenesis, cell differentiation, and cell survival. To assess the prognostic impact of ERK in TNBC patients, relative quantities of ERK (ERK-2 and pMAPK) and direct targets of the ERK pathway (MAPK/ERK kinase 1, phospho-enriched protein in astrocytes [PEA]-15, phosphorylated (p)PEA-15, tuberous sclerosis protein 2, p70S6 kinase, and p27) were measured using reverse-phase protein arrays in tumor tissue from patients with TNBC (n = 97) and non-TNBC (n = 223). Protein levels in patients with TNBC were correlated with clinical and tumor characteristics and outcome. The median age of patients with TNBC was 55 years (range, 27-86 years). Disease stage was I in 21%, II in 60%, and III in 20% of the patients. In a multivariate analysis, among patients with TNBC, those with ERK-2-overexpressing tumors had a lower overall survival rate than those with low ERK-2-expressing tumors (hazard ratio [HR], 2.76; 95% confidence interval [CI], 1.19-6.41). However, high pMAPK levels were associated with a significantly higher relapse-free survival rate (HR, 0.66; 95% CI, 0.46-0.95). In conclusion, ERK-2 and pMAPK are valuable prognostic markers in TNBC. Further studies are justified to elucidate ERK's role in TNBC tumorigenicity and metastasis.

Prognostic impact of phosphorylated HER-2 in HER-2+ primary breast cancer.

PURPOSE: Tyrosine 1248 is one of the autophosphorylation sites of human epidermal growth factor receptor (HER)-2. We determined the prognostic value of the expression level of tyrosine 1248-phosphorylated HER-2 (pHER-2) in patients with HER-2(+) primary breast cancer using a reverse-phase protein array. PATIENTS AND METHODS: The optimal cutoff value of pHER-2 for segregating disease-free survival (DFS) was determined by receiver operating characteristic (ROC) curve analysis. Five-year DFS for pHER-2 expression level was estimated with the Kaplan-Meier method using both derivation (n = 162) and validation (n = 227) cohorts. RESULTS: Of the 162 patients in the derivation cohort, 26 had high HER-2 expression levels. The area under the ROC curve for pHER-2 level and DFS was 0.662. Nineteen of the 162 patients (11.7%) had high pHER-2 expression levels (pHER-2(high)); 143 patients (88.3%) had low pHER-2 expression levels (pHER-2(low)). Among the 26 patients with high HER-2 expression levels, the 17 pHER-2(high) patients had a significantly lower 5-year DFS rate than the nine pHER-2(low) patients (23.5% versus 77.8%). On multivariate analysis, only pHER-2(high) independently predicted DFS in the Cox proportional hazards model. In the validation cohort, among 61 patients with high HER-2 expression, the difference in 5-year DFS rates between pHER-2(high) (n = 7) and pHER-2(low) (n = 54) patients was marginal (57.1% versus 81.5%). CONCLUSION: In patients with HER-2(+) primary breast cancer, pHER-2(high) patients had a lower 5-year DFS rate than pHER-2(low) patients. Quantification of pHER-2 expression level may provide prognostic information beyond the current standard HER-2 status.

A genomic predictor of response and survival following taxane-anthracycline chemotherapy for invasive breast cancer.

CONTEXT: Prediction of high probability of survival from standard cancer treatments is fundamental for individualized cancer treatment strategies. OBJECTIVE: To develop a predictor of response and survival from chemotherapy for newly diagnosed invasive breast cancer. DESIGN, SETTING, AND PATIENTS: Prospective multicenter study conducted from June 2000 to March 2010 at the M. D. Anderson Cancer Center to develop and test genomic predictors for neoadjuvant chemotherapy. Patients were those with newly diagnosed ERBB2 (HER2 or HER2/neu)-negative breast cancer treated with chemotherapy containing sequential taxane and anthracycline-based regimens (then endocrine therapy if estrogen receptor [ER]-positive). Different predictive signatures for resistance and response to preoperative (neoadjuvant) chemotherapy (stratified according to ER status) were developed from gene expression microarrays of newly diagnosed breast cancer (310 patients). Breast cancer treatment sensitivity was then predicted using the combination of signatures for (1) sensitivity to endocrine therapy, (2) chemoresistance, and (3) chemosensitivity, with independent validation (198 patients) and comparison with other reported genomic predictors of chemotherapy response. MAIN OUTCOME MEASURES: Distant relapse-free survival (DRFS) if predicted treatment sensitive and absolute risk reduction ([ARR], difference in DRFS between 2 predicted groups) at median follow-up (3 years). RESULTS: Patients in the independent validation cohort (99% clinical stage II-III) who were predicted to be treatment sensitive (28%) had 56% (95% CI, 31%-78%) probability of excellent pathologic response and DRFS of 92% (95% CI, 85%-100%), with an ARR of 18% (95% CI, 6%-28%). Survival was predicted in ER-positive (30% predicted sensitive; DRFS, 97% [95% CI, 91%-100%]; ARR, 11% [95% CI, 0.1%-21%]) and ER-negative (26% predicted sensitive; DRFS, 83% [95% CI, 68%-100%]; ARR, 26% [95% CI, 4%-48%]) subsets and was significant in multivariate analysis. Other genomic predictors showed paradoxically worse survival for patients predicted to be responsive to chemotherapy. CONCLUSION: A genomic predictor combining ER status, predicted chemoresistance, predicted chemosensitivity, and predicted endocrine sensitivity identified patients with high probability of survival following taxane and anthracycline chemotherapy.

PI3K pathway mutations and PTEN levels in primary and metastatic breast cancer.

The purpose of this work was to determine whether there are differences in PIK3CA mutation status and PTEN protein expression between primary and matched metastatic breast tumors as this could influence patient management. Paraffin sections of 50 µm were used for DNA extraction and slides of 3 µm for immunohistochemistry (IHC) and FISH. Estrogen receptor, progesterone receptor, and HER2 IHC were repeated in a central laboratory for both primary tumors and metastases. PTEN levels were assessed by IHC and phosphoinositide 3-kinase (PI3K) pathway mutations were detected by a mass spectroscopy-based approach. Median age was 48 years (range: 30-83 years). Tumor subtype included 72% hormone receptor positive/HER2 negative, 20% HER2-positive, and less than 7.8% triple receptor negative. Tissues were available for PTEN IHC in 46 primary tumors and 52 metastases. PTEN was lost in 14 (30%) primary tumors and 13 (25%) metastases. There were five cases of PTEN loss and eight cases of PTEN gain from primary tumors to metastases (26% discordance). Adequate DNA was obtained from 46 primary tumors and from 50 metastases for PIK3CA analysis. PIK3CA mutations were detected in 19 (40%) of primary tumors and 21 (42%) of metastases. There were five cases of PIK3CA mutation loss and four cases of mutation gain (18% discordance). There was an increase of the level of PIK3CA mutations in four cases and decrease in one case from primary tumors to metastases. There is a high level of discordance in PTEN level, PIK3CA mutations, and receptor status between primary tumors and metastases that may influence patient selection and response to PI3K-targeted therapies.