All-cause mortality as the primary endpoint for the GRAIL/National Health Service England multi-cancer screening trial

A randomized trial of the GRAIL GalleriTM multi-cancer screening test is being planned for the National Health Service in England, and will have 140,000 healthy participants aged 50–79: 70,000 exposed to screening and 70,000 unexposed. The test reportedly detects 50 different cancers and is expected to reduce all-cancer mortality by approximately 25%. Given this effect size—and that cancer deaths constitute a large fraction of all deaths—the trial is sufficiently large to test the effect on all-cause mortality. Because most patients believe cancer screening “saves lives”, the GRAIL/National Health Service collaboration could set the evaluation standard for multi-cancer screening.

Creating a training set for artificial intelligence from initial segmentations of airways

Airways segmentation is important for research about pulmonary disease but require a large amount of time by trained specialists. We used an openly available software to improve airways segmentations obtained from an artificial intelligence (AI) tool and retrained the tool to get a better performance. Fifteen initial airway segmentations from low-dose chest computed tomography scans were obtained with a 3D-Unet AI tool previously trained on Danish Lung Cancer Screening Trial and Erasmus-MC Sophia datasets. Segmentations were manually corrected in 3D Slicer. The corrected airway segmentations were used to retrain the 3D-Unet. Airway measurements were automatically obtained and included count, airway length and luminal diameter per generation from the segmentations. Correcting segmentations required 2–4 h per scan. Manually corrected segmentations had more branches (p < 0.001), longer airways (p < 0.001) and smaller luminal diameters (p = 0.004) than initial segmentations. Segmentations from retrained 3D-Unets trended towards more branches and longer airways compared to the initial segmentations. The largest changes were seen in airways from 6th generation onwards. Manual correction results in significantly improved segmentations and is potentially a useful and time-efficient method to improve the AI tool performance on a specific hospital or research dataset.

Dietary consumption of tea and the risk of prostate cancer in the Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial

Tea contains polyphenols such as flavonoids, anthocyanidins, flavanols and phenolic acids which in laboratory studies have reported to promote antioxidant enzyme formation, reduces excess inflammation, slow cancer cell proliferation and promote apoptosis. Evidence from epidemiological studies, on the effect of tea consumption on CaP incidence has been conflicting. We analysed data from 25 097 men within the intervention arm of the 155000 participant Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial screening trial. Histologically confirmed cases of prostate cancer were reported in 3,088 men (12.3%) during the median 11.5 year follow up. Tea consumption was assessed with a food frequency questionnaire. Baseline characteristics were compared between groups using Chi-square and Kruskal-Wallis tests. Cox regression models were used to assess associations between tea intake and CaP incidence. There was no statistical difference between the risk of CaP between men who never drank tea to those who drank tea at any quantity. Amongst tea drinkers, those in the highest third of consumption group had a small but significantly lower risk compared to those in the lowest third (11.2% v 13.2% HR 1.16; 95% CI 1.05-1.29, p=0.004). This pattern persisted with adjustments for demographics and lifestyle. In conclusion, among tea drinkers, there was a small positive association between drinking tea and a reduced risk of prostate cancer. It does not support starting to drink tea, if men previously did not, to reduce the risk. Further research is needed to establish whether tea is justified for future prospective nutritional intervention studies investigating CaP prevention.

Lung-RADS usado no rastreamento do câncer de pulmão: a doença granulomatosa faz diferença?

Objetivo: o objetivo é analisar os resultados do rastreamento do câncer de pulmão por tomografia computadorizada de baixa dose (LDCT) por meio do Sistema de Relatórios e Dados de Triagem de TC de Pulmão (Lung-RADS) em um centro de câncer brasileiro. Materiais e métodos: revisamos retrospectivamente os prontuários de pacientes submetidos ao programa de rastreamento de câncer de pulmão de LDCT basal no A.C. Camargo Cancer Center. Os critérios de inclusão e exclusão foram iguais aos do National Lung Cancer Screening Trial (NLST). Os critérios para achados de imagem foram aqueles classificados de acordo com as categorias de avaliação do Lung Imaging Reporting and Data System (Lung-RADS). Resultados: Dos 287 indivíduos avaliados neste estudo, 72,1% apresentaram TC de triagem negativa (categorias 1 ou 2 do Lung-RADS), o restante teve TC de triagem positiva, considerando 5,6% na categoria 4A de Lung-RADS, 2,1% na categoria 4B e 1,0% na categoria 4X. Os principais achados foram avaliados em 218 (75,9%) indivíduos, com nódulos sólidos (64,8%), parcialmente sólidos (2,7%) e não sólidos (8,3%). A maioria dos pacientes (59,1%) apresentaram nódulos sólidos menores que 6 mm. Os resultados histológicos confirmaram câncer de pulmão em 2 casos (prevalência de 0,7% de todos os pacientes triados). Conclusões: A prevalência de câncer de pulmão em nossa amostra foi compatível com a literatura. No entanto, tivemos uma prevalência maior das categorias 3 e 4A do Lung-RADS do que o esperado. Isso pode estar associado à maior incidência de doenças granulomatosas, principalmente tuberculose, na população brasileira.

Automatic airway segmentation from computed tomography using robust and efficient 3-D convolutional neural networks

This paper presents a fully automatic and end-to-end optimised airway segmentation method for thoracic computed tomography, based on the U-Net architecture. We use a simple and low-memory 3D U-Net as backbone, which allows the method to process large 3D image patches, often comprising full lungs, in a single pass through the network. This makes the method simple, robust and efficient. We validated the proposed method on three datasets with very different characteristics and various airway abnormalities: (1) a dataset of pediatric patients including subjects with cystic fibrosis, (2) a subset of the Danish Lung Cancer Screening Trial, including subjects with chronic obstructive pulmonary disease, and (3) the EXACT’09 public dataset. We compared our method with other state-of-the-art airway segmentation methods, including relevant learning-based methods in the literature evaluated on the EXACT’09 data. We show that our method can extract highly complete airway trees with few false positive errors, on scans from both healthy and diseased subjects, and also that the method generalizes well across different datasets. On the EXACT’09 test set, our method achieved the second highest sensitivity score among all methods that reported good specificity.