Cancer histology reflects underlying molecular processes and disease progression and contains rich phenotypic information that is predictive of patient outcomes. In this study, we show a ...computational approach for learning patient outcomes from digital pathology images using deep learning to combine the power of adaptive machine learning algorithms with traditional survival models. We illustrate how these survival convolutional neural networks (SCNNs) can integrate information from both histology images and genomic biomarkers into a single unified framework to predict time-to-event outcomes and show prediction accuracy that surpasses the current clinical paradigm for predicting the overall survival of patients diagnosed with glioma. We use statistical sampling techniques to address challenges in learning survival from histology images, including tumor heterogeneity and the need for large training cohorts. We also provide insights into the prediction mechanisms of SCNNs, using heat map visualization to show that SCNNs recognize important structures, like microvascular proliferation, that are related to prognosis and that are used by pathologists in grading. These results highlight the emerging role of deep learning in precision medicine and suggest an expanding utility for computational analysis of histology in the future practice of pathology.
Abstract
The CBTRUS Statistical Report: Pediatric Brain Tumor Foundation Childhood and Adolescent Primary Brain and Other Central Nervous System Tumors Diagnosed in the United States in 2014–2018 ...comprehensively describes the current population-based incidence of primary malignant and non-malignant brain and other CNS tumors in children and adolescents ages 0–19 years, collected and reported by central cancer registries covering approximately 100% of the United States population. Overall, brain and other CNS tumors are the most common solid tumor, the most common cancer, and the most common cause of cancer death in children and adolescents ages 0–19 years. This report aims to serve as a useful resource for researchers, clinicians, patients, and families.
Glioma is the most commonly occurring malignant brain tumor in the United States, and its incidence varies by age, sex, and race or ethnicity. Survival after brain tumor diagnosis has been shown to ...vary by these factors.
To quantify the differences in incidence and survival rates of glioma in adults by race or ethnicity.
This population-based study obtained incidence data from the Central Brain Tumor Registry of the United States and survival data from Surveillance, Epidemiology, and End Results registries, covering the period January 1, 2000, to December 31, 2014. Average annual age-adjusted incidence rates with 95% CIs were generated by glioma histologic groups, race, Hispanic ethnicity, sex, and age groups. One-year and 5-year relative survival rates were generated by glioma histologic groups, race, Hispanic ethnicity, and insurance status. The analysis included 244 808 patients with glioma diagnosed in adults aged 18 years or older. Data were collected from January 1, 2000, to December 31, 2014. Data analysis took place from December 11, 2017, to January 31, 2018.
Overall, 244 808 patients with glioma were analyzed. Of these, 150 631 (61.5%) were glioblastomas, 46 002 (18.8%) were non-glioblastoma astrocytomas, 26 068 (10.7%) were oligodendroglial tumors, 8816 (3.6%) were ependymomas, and 13 291 (5.4%) were other glioma diagnoses in adults. The data set included 137 733 males (56.3%) and 107 075 (43.7%) females. There were 204 580 non-Hispanic whites (83.6%), 17 321 Hispanic whites (7.08%), 14 566 blacks (6.0%), 1070 American Indians or Alaska Natives (0.4%), and 5947 Asians or Pacific Islanders (2.4%). Incidences of glioblastoma, non-glioblastoma astrocytoma, and oligodendroglial tumors were higher among non-Hispanic whites than among Hispanic whites (30% lower overall), blacks (52% lower overall), American Indians or Alaska Natives (58% lower overall), or Asians or Pacific Islanders (52% lower overall). Most tumors were more common in males than in females across all race or ethnicity groups, with the great difference in glioblastoma where the incidence was 60% higher overall in males. Most tumors (193 329 79.9%) occurred in those aged 45 years or older, with differences in incidence by race or ethnicity appearing in all age groups. Survival after diagnosis of glioma of different subtypes was generally comparable among Hispanic whites, blacks, and Asians or Pacific Islanders but was lower among non-Hispanic whites for many tumor types, including glioblastoma, irrespective of treatment type.
Incidence of glioma and 1-year and 5-year survival rates after diagnosis vary significantly by race or ethnicity, with non-Hispanic whites having higher incidence and lower survival rates compared with individuals of other racial or ethnic groups. These findings can inform future discovery of risk factors and reveal unaddressed health disparities.
Glioblastoma multiforme (GBM) is the most common and aggressive primary central nervous system malignancy with a median survival of 15 months. The average incidence rate of GBM is 3.19/100,000 ...population, and the median age of diagnosis is 64 years. Incidence is higher in men and individuals of white race and non-Hispanic ethnicity. Many genetic and environmental factors have been studied in GBM, but the majority are sporadic, and no risk factor accounting for a large proportion of GBMs has been identified. However, several favorable clinical prognostic factors are identified, including younger age at diagnosis, cerebellar location, high performance status, and maximal tumor resection. GBMs comprise of primary and secondary subtypes, which evolve through different genetic pathways, affect patients at different ages, and have differences in outcomes. We report the current epidemiology of GBM with new data from the Central Brain Tumor Registry of the United States 2006 to 2010 as well as demonstrate and discuss trends in incidence and survival. We also provide a concise review on molecular markers in GBM that have helped distinguish biologically similar subtypes of GBM and have prognostic and predictive value.
A high tumour mutational burden (hypermutation) is observed in some gliomas
; however, the mechanisms by which hypermutation develops and whether it predicts the response to immunotherapy are poorly ...understood. Here we comprehensively analyse the molecular determinants of mutational burden and signatures in 10,294 gliomas. We delineate two main pathways to hypermutation: a de novo pathway associated with constitutional defects in DNA polymerase and mismatch repair (MMR) genes, and a more common post-treatment pathway, associated with acquired resistance driven by MMR defects in chemotherapy-sensitive gliomas that recur after treatment with the chemotherapy drug temozolomide. Experimentally, the mutational signature of post-treatment hypermutated gliomas was recapitulated by temozolomide-induced damage in cells with MMR deficiency. MMR-deficient gliomas were characterized by a lack of prominent T cell infiltrates, extensive intratumoral heterogeneity, poor patient survival and a low rate of response to PD-1 blockade. Moreover, although bulk analyses did not detect microsatellite instability in MMR-deficient gliomas, single-cell whole-genome sequencing analysis of post-treatment hypermutated glioma cells identified microsatellite mutations. These results show that chemotherapy can drive the acquisition of hypermutated populations without promoting a response to PD-1 blockade and supports the diagnostic use of mutational burden and signatures in cancer.
Population-based estimates of the incidence of brain metastases are not generally available. The purpose of this study was to calculate population-based incidence proportions (IPs) of brain ...metastases from single primary lung, melanoma, breast, renal, or colorectal cancer.
Patients diagnosed with single primary lung, melanoma, breast, renal, or colorectal cancer (1973 to 2001) in the Metropolitan Detroit Cancer Surveillance System (MDCSS) were used for analysis. IP of brain metastases by primary site and variable of interest (race, sex, age at diagnosis of primary cancer, and Surveillance, Epidemiology, and End Results SEER stage of primary cancer) was calculated with 95% CIs.
Total IP percentage (IP%) of brain metastases was 9.6% for all primary sites combined, and highest for lung (19.9%), followed by melanoma (6.9%), renal (6.5%), breast (5.1%), and colorectal (1.8%) cancers. Racial differences were seen with African Americans demonstrating higher IP% of brain metastases compared with other racial groups for most primary sites. IP% was significantly higher for female patients with lung cancer, and significantly higher for male patients with melanoma. The highest IP% of brain metastases occurred at different ages at diagnoses: age 40 to 49 years for primary lung cancer; age 50 to 59 years for primary melanoma, renal, or colorectal cancers; and age 20 to 39 for primary breast cancer. IP% significantly increased as SEER stage of primary cancer advanced for all primary sites.
Total IP% of brain metastases was lower than previously reported, and it varied by primary site, race, sex, age at diagnosis of primary cancer, and SEER stage of primary cancer.