Receiver operating characteristic (ROC) curves are useful tools to evaluate classifiers in biomedical and bioinformatics applications. However, conclusions are often reached through inconsistent use or insufficient statistical analysis. To support researchers in their ROC curves analysis we developed pROC, a package for R and S+ that contains a set of tools displaying, analyzing, smoothing and comparing ROC curves in a user-friendly, object-oriented and flexible interface. With data previously imported into the R or S+ environment, the pROC package builds ROC curves and includes functions for computing confidence intervals, statistical tests for comparing total or partial area under the curve or the operating points of different classifiers, and methods for smoothing ROC curves. Intermediary and final results are visualised in user-friendly interfaces. A case study based on published clinical and biomarker data shows how to perform a typical ROC analysis with pROC. pROC is a package for R and S+ specifically dedicated to ROC analysis. It proposes multiple statistical tests to compare ROC curves, and in particular partial areas under the curve, allowing proper ROC interpretation. pROC is available in two versions: in the R programming language or with a graphical user interface in the S+ statistical software. It is accessible at http://expasy.org/tools/pROC/
 under the GNU General Public License. It is also distributed through the CRAN and CSAN public repositories, facilitating its installation.

/pdf/proc-an-open-source-package-for-r-and-s-to-analyze-and-p7qu4wozaa.pdf

pROC: an open-source package for R and S+ to analyze and compare ROC curves

Essential Medical Statistics

https://www.thelancet.com/pdfs/journals/lancet/PIIS0140-6736(14)60584-5.pdf

Effect of treatment delay, age, and stroke severity on the effects of intravenous thrombolysis with alteplase for acute ischaemic stroke: a meta-analysis of individual patient data from randomised trials

Effect of treatment delay, age, and stroke severity on the effects of intravenous thrombolysis with alteplase for acute ischaemic stroke: a meta-analysis of individual patient data from randomised trials | NOVA. The University of Newcastle's Digital Repository

The development of decision support systems for pathology and their deployment in clinical practice have been hindered by the need for large manually annotated datasets. To overcome this problem, we present a multiple instance learning-based deep learning system that uses only the reported diagnoses as labels for training, thereby avoiding expensive and time-consuming pixel-wise manual annotations. We evaluated this framework at scale on a dataset of 44,732 whole slide images from 15,187 patients without any form of data curation. Tests on prostate cancer, basal cell carcinoma and breast cancer metastases to axillary lymph nodes resulted in areas under the curve above 0.98 for all cancer types. Its clinical application would allow pathologists to exclude 65–75% of slides while retaining 100% sensitivity. Our results show that this system has the ability to train accurate classification models at unprecedented scale, laying the foundation for the deployment of computational decision support systems in clinical practice.

Clinical-grade computational pathology using weakly supervised deep learning on whole slide images.

Further consideration of the incidence of cancers around the nuclear installations at Sellafield and Dounreay

The recent study of O’Neill and her colleagues increases the evidence for a raised risk of childhood leukaemia (and some other childhood cancers) with increased birthweight. With funding from Children with Cancer UK, we have been archiving electronically the records of the Oxford Survey of Childhood Cancers (OSCC), a large case-control study of childhood cancer mortality in Great Britain during 1953-81, which finally included 15,276 case-control pairs. Although data were collected for many potential risk factors, such as maternal age and maternal illnesses during pregnancy, birthweight was not systematically collected in the OSCC, although particular studies using OSCC data obtained birthweights for deaths during 1971–76 and 1977–81. One aspect of the OSCC findings that we plan to re-examine is the association between the risk of childhood leukaemia and other cancers and an antenatal X-ray examination, and it is our intention to investigate the influence of birthweight on this association, since it is feasible that heavier fetuses may be more frequently the subject of diagnostic radiography. However, we wish to draw attention to the earlier findings for birthweight reported from the USA by Monson and MacMahon, which may have been overlooked in various reviews of the evidence. They extended the original case-cohort study of MacMahon, and identified a total of 1104 deaths from cancer during 1947–67 at less than 10 years of age among a cohort of almost 1.5 million births in 42 large maternity hospitals in the north-eastern USA during 1947–60. Medical records were obtained for the children affected with cancer and for a sample of 14 294 (1%) of the birth cohort. The principal objective of the study was the examination of the association between the risk of childhood leukaemia or other cancer and an antenatal X-ray examination, but various other data on possible confounding factors were also obtained, including birthweight. Table 1 shows the findings of the study for birthweight, for 1104 cancer deaths (597 leukaemia, 507 other cancers) at less than 10 years of age. There is a clear monotonically increasing trend of leukaemia mortality rate with categories of increasing birthweight, such that the rate more than doubles between the least and greatest birthweight categories. Using the birthweight distribution for the 1% sample of births and the mortality rates given in Table 1 to infer the actual numbers of deaths from leukaemia and other cancers in the different birthweight categories, we fitted a logistic model to the proportions; from this we estimate that for leukaemia there is a linear trend in log odds ratio equivalent to a factor increase of 1.38 per kg birthweight (95% confidence interval 1.15, 1.64). Although not so clear from Table 1, the trend is similar for the grouping of all other cancers, with an estimated increase of 1.36 per kg (1.13, 1.65). The trend for leukaemia is statistically compatible with those reported by O’Neill et al., although the differences between the studies (mortality vs incidence, age at death/diagnosis, continuous vs categorial analysis, etc.) should be noted. Also shown in Table 1 is how the frequency of antenatal X-ray examination varies with birthweight in the 1% sample of the birth cohort: there is a significant (P1⁄4 0.02) positive association between birthweight and fetal X-raying frequency.

Childhood cancer—the role of birthweight and antenatal radiography

Committee on Medical Aspects of Radiation in the Environment (COMARE). Seventeenth report. Further considerations of the incidence of cancers around the nuclear installations at Sellafield and Dounreay.

Childhood cancer and intrauterine irradiation

In this report, the Committee on Medical Aspects of Radiation in the Environment (COMARE) presents a comprehensive review of the radium contamination in the area around Dalgety Bay. This report covers the history of the site, the type and extent of the contamination, the recent investigations and the cancer epidemiology for the area. The report also considers the implications for other similarly contaminated sites.

John F. Bithell

Papers

Further consideration of the incidence of cancers around the nuclear installations at Sellafield and Dounreay

Childhood cancer—the role of birthweight and antenatal radiography

Committee on Medical Aspects of Radiation in the Environment (COMARE). Seventeenth report. Further considerations of the incidence of cancers around the nuclear installations at Sellafield and Dounreay.

Childhood cancer and intrauterine irradiation

Radium contamination in the area around Dalgety Bay