Validation of a natural language processing algorithm to identify adenomas and measure adenoma detection rates across a health system: a population-level study.

BACKGROUND AND AIMS: Measuring adenoma detection rates (ADRs) at the population level is challenging because pathology reports are often reported in an unstructured format; further, there is significant variation in reporting methods across institutions. Natural language processing (NLP) can be used to extract relevant information from text-based records. We aimed to develop and validate an NLP algorithm to identify colorectal adenomas that could be used to report ADR at the population level in Ontario, Canada. METHODS: The sampling frame included pathology reports from all colonoscopies performed in Ontario in 2015 and 2016. Two random samples of 450 and 1000 reports were selected as the training and validation sets, respectively. Expert clinicians reviewed and classified reports as adenoma or other. The training set was used to develop an NLP algorithm (to identify adenomas) that was evaluated using the validation set. The NLP algorithm test characteristics were calculated using expert review as the reference. We used the algorithm to measure ADR for all endoscopists in Ontario in 2019. RESULTS: The 1450 pathology reports were derived from 62 laboratories, 266 pathologists, and 532 endoscopists. In the training set, the NLP algorithm for any adenoma had a sensitivity of 99.60% (95% confidence interval (CI), 97.77-99.99), specificity of 99.01% (95% CI, 96.49-99.88), positive predictive value of 99.19% (95% CI, 97.12-99.90), and F1 score of .99. Similar results were obtained for the validation set. The median ADR was 33% (interquartile range, 26%-40%). CONCLUSIONS: When we used a population-based sample from Ontario, our NLP algorithm was highly accurate and was used at the system level to measure ADR.

The Development of a Data Collection and Browser Fingerprinting System.

The urgent need to protect user privacy and security has emerged as the World Wide Web has become an increasingly necessary part of daily life. Browser fingerprinting is a very interesting topic in the industry of technology security. New technology will always raise new security issues and browser fingerprinting will undoubtedly follow the same process. It has become one of the most popular topics in online privacy because, to date, there is still no exact solution as to how to stop it entirely. The majority of solutions just aim to reduce the likelihood of obtaining a browser fingerprint. Research on browser fingerprinting is unquestionably required since it is essential to educate users, developers, policymakers, and law enforcement about it so that they can make strategic choices based on knowledge. Browser fingerprinting must be recognised in order to defend against privacy problems. A browser fingerprint is described as data gathered by the receiving server to identify a distant device, and it is different from cookies. Websites frequently utilize browser fingerprinting to obtain information about the type and version of the browser, as well as the operating system, and other current settings. It has been known that even when cookies are disabled, fingerprints can be used to fully or partially identify users or devices. In this communication paper, a new insight into the challenge of browser fingerprint is encouraged as a new venture. Thus, the initial way to truly understand the browser fingerprint is the need to collect browser fingerprints. In this work, the process of data collection for browser fingerprinting through scripting, to offer a complete all-in-one fingerprinting test suite, has been thoughtfully divided into appropriate sections and grouped with key information to be carried out. The objective is to gather fingerprint data with no personal identification information and make it an open source of raw datasets in the industry for any future research purposes. To our best knowledge, there are no open datasets made available for browser fingerprints in the research field. The dataset will be widely accessible by anyone interested in obtaining those data. The dataset collected will be very raw and will be in the form of a text file. Thus, the main contribution of this work is to share an open dataset of browser fingerprints along with its collection methodology.

Diffuse intrinsic pontine glioma: is MRI surveillance improved by region of interest volumetry?

BACKGROUND: Paediatric diffuse intrinsic pontine glioma (DIPG) is noteworthy for its fibrillary infiltration through neuroparenchyma and its resultant irregular shape. Conventional volumetry methods aim to approximate such irregular tumours to a regular ellipse, which could be less accurate when assessing treatment response on surveillance MRI. Region-of-interest (ROI) volumetry methods, using manually traced tumour profiles on contiguous imaging slices and subsequent computer-aided calculations, may prove more reliable. OBJECTIVE: To evaluate whether the reliability of MRI surveillance of DIPGs can be improved by the use of ROI-based volumetry. MATERIALS AND METHODS: We investigated the use of ROI- and ellipsoid-based methods of volumetry for paediatric DIPGs in a retrospective review of 22 MRI examinations. We assessed the inter- and intraobserver variability of the two methods when performed by four observers. RESULTS: ROI- and ellipsoid-based methods strongly correlated for all four observers. The ROI-based volumes showed slightly better agreement both between and within observers than the ellipsoid-based volumes (inter-[intra-]observer agreement 89.8% [92.3%] and 83.1% [88.2%], respectively). Bland-Altman plots show tighter limits of agreement for the ROI-based method. CONCLUSION: Both methods are reproducible and transferrable among observers. ROI-based volumetry appears to perform better with greater intra- and interobserver agreement for complex-shaped DIPG.

Successful treatment of metastatic relapse of medulloblastoma in childhood with single session stereotactic radiosurgery: a report of 3 cases.

Stereotactic radiosurgery (SRS) is an increasingly used treatment modality in adults, but its use and effectiveness in pediatric brain tumors is still uncertain. We describe 3 patients with metastatic relapse of medulloblastoma, who were treated with SRS, and achieved prolonged, progression-free survival. Tolerability of the treatment was excellent with no adverse effects reported. This work adds to the growing evidence that SRS may have an important role to play in the treatment of pediatric brain tumors.

Cryptic kin discrimination during communal lactation in mice favours cooperation between relatives.

Breeding females can cooperate by rearing their offspring communally, sharing synergistic benefits of offspring care but risking exploitation by partners. In lactating mammals, communal rearing occurs mostly among close relatives. Inclusive fitness theory predicts enhanced cooperation between related partners and greater willingness to compensate for any partner under-investment, while females are less likely to bias investment towards own offspring. We use a dual isotopic tracer approach to track individual milk allocation when familiar pairs of sisters or unrelated house mice reared offspring communally. Closely related pairs show lower energy demand and pups experience better access to non-maternal milk. Lactational investment is more skewed between sister partners but females pay greater energetic costs per own offspring reared with an unrelated partner. The choice of close kin as cooperative partners is strongly favoured by these direct as well as indirect benefits, providing a driver to maintain female kin groups for communal breeding.

Evolution of a Systematic Approach to Smoking Cessation in Ontario's Regional Cancer Centres.

Smoking cessation after a cancer diagnosis can significantly improve a person's prognosis, treatment efficacy and safety, and quality of life. In 2012, Cancer Care Ontario (now part of Ontario Health) introduced a Framework for Smoking Cessation, to be implemented for new ambulatory cancer patients at the province's 14 Regional Cancer Centres (RCCs). Over time, the program has evolved to become more efficient, use data for robust performance management, and broaden its focus to include new patient populations and additional data collection. In 2017, the framework was revised from a 5As to a 3As brief intervention model, along with an opt-out approach to referrals. The revised model was based on emerging evidence, feedback from stakeholders, and an interim program evaluation. Results showed an initial increase in referrals to cessation services. Two indicators (tobacco use screening and acceptance of a referral) are routinely monitored as part of Ontario Health's system-wide performance management approach, which has been identified as a key driver of change among RCCs. Due to the COVID-19 pandemic, many RCCs reported a decrease in these indicators. RCCs that were able to maintain a high level of smoking cessation activities during the pandemic offer valuable lessons, including the opportunity to swiftly leverage virtual care. Future directions for the program include capturing data on cessation outcomes and expanding the intervention to new populations. A focus on system recovery from COVID-19 will be paramount. Smoking cessation must remain a core element of high-quality cancer care, so that patients achieve the best possible health benefits from their treatments.

Arginine-vasopressin infusion in a child with cranial diabetes insipidus during hyperhydration therapy with chemotherapy: a therapeutic challenge.

SUMMARY: An 11-year-old girl presented with acute lower limb weakness, dehydration, hypernatraemia and secondary rhabdomyolysis on a background of an 8-month history of polyuria. Radiological investigations revealed a suprasellar tumour which was diagnosed on biopsy as a non-metastatic germinoma. Further endocrinological investigations confirmed panhypopituitarism and she commenced desmopressin, hydrocortisone and thyroxine. Her chemotherapeutic regime consisted of etoposide, carboplatin and ifosfamide, the latter of which required 4 litres of hyperhydration therapy daily. During the first course of ifosfamide, titration of oral desmopressin was trialled but this resulted in erratic sodium control leading to disorientation. Based on limited literature, we then trialled an arginine-vasopressin (AVP) infusion. A sliding scale was developed to adjust the AVP dose, with an aim to achieve a urine output of 3-4 mL/kg/h. During the second course of ifosamide, AVP infusion was commenced at the outset and tighter control of urine output and sodium levels was achieved. In conclusion, we found that an AVP infusion during hyperhydration therapy was required to achieve eunatraemia in a patient with cranial diabetes insipidus. Developing an AVP sliding scale requires individual variation; further reports/case series are required to underpin practice. LEARNING POINTS: Certain chemotherapeutic regimens require large fluid volumes of hyperhydration therapy which can result in significant complications secondary to rapid serum sodium shifts in patients with diabetes insipidus. The use of a continuous AVP infusion and titrating with a sliding scale is more effective than oral desmopressin in regulating plasma sodium and fluid balance during hyperhydration therapy. No adverse effects were found in our patient using a continuous AVP infusion. Adjustment of the AVP infusion rate depends on urine output, fluid balance, plasma sodium levels and sensitivity/response of the child to titrated AVP doses.

Comparison of self-reported late effects with medical records among survivors of childhood cancer.

BACKGROUND: Survival rates following childhood cancer have increased, but survivors experience significant late effects. Long-term follow-up is recommended but imposes an increasing burden on health services. We report prevalence of morbidity in a cohort of survivors from South Yorkshire based on: (i) case-note analysis and (ii) self-reported late effects (parent-reported for under-16s). METHODS: Treatment information was taken from case-notes. Comparisons were made between late effects described in notes and reported by 108 survivors aged >16 years, and 45 parents of survivors (12-15 years). FINDINGS: Of 892 patients diagnosed with childhood cancer and some benign conditions registered on hospital databases from January 1990 to December 2005, 337 (37.8%) met eligibility criteria. Ninety-one survivors (16) (84.3%, confidence interval [CI]: 76.0-90.6) reported one or more late effects (mean=3.5; CI: 3.0-4.1), significantly higher than the number of late effects documented in medical notes (mean=0.7; CI: 0.5-0.9; t=-11.26, p<0.001). Thirty-five parents (77.8%, CI: 65.1-90.4) reported late effects for their children (mean=2.7; CI: 2.0-3.4), again higher than medical notes (mean=0.7; CI: 0.4-1.1; t=7.18, p<0.001). More than 30 specialties were involved in survivor care (mean=1.5; CI: 1.4-1.6; range 0-6). Those with more late effects saw more specialties (r=0.51, p<0.001). INTERPRETATION: We confirm the wide range of late effects experienced by survivors of child cancer, significantly greater than those recorded in medical notes, and requiring care from a range of specialties. Decisions about follow-up need to take account of patient-reported morbidity and concerns.