For virtually every patient with colorectal cancer (CRC), hematoxylin–eosin (HE)–stained tissue slides are available. These images contain quantitative information, which is not routinely used to objectively extract prognostic biomarkers. In the present study, we investigated whether deep convolutional neural networks (CNNs) can extract prognosticators directly from these widely available images. We hand-delineated single-tissue regions in 86 CRC tissue slides, yielding more than 100,000 HE image patches, and used these to train a CNN by transfer learning, reaching a nine-class accuracy of >94% in an independent data set of 7,180 images from 25 CRC patients. With this tool, we performed automated tissue decomposition of representative multitissue HE images from 862 HE slides in 500 stage I–IV CRC patients in the The Cancer Genome Atlas (TCGA) cohort, a large international multicenter collection of CRC tissue. Based on the output neuron activations in the CNN, we calculated a “deep stroma score,” which was an independent prognostic factor for overall survival (OS) in a multivariable Cox proportional hazard model (hazard ratio [HR] with 95% confidence interval [CI]: 1.99 [1.27–3.12], p = 0.0028), while in the same cohort, manual quantification of stromal areas and a gene expression signature of cancer-associated fibroblasts (CAFs) were only prognostic in specific tumor stages. We validated these findings in an independent cohort of 409 stage I–IV CRC patients from the “Darmkrebs: Chancen der Verhütung durch Screening” (DACHS) study who were recruited between 2003 and 2007 in multiple institutions in Germany. Again, the score was an independent prognostic factor for OS (HR 1.63 [1.14–2.33], p = 0.008), CRC-specific OS (HR 2.29 [1.5–3.48], p = 0.0004), and relapse-free survival (RFS; HR 1.92 [1.34–2.76], p = 0.0004). A prospective validation is required before this biomarker can be implemented in clinical workflows. In our retrospective study, we show that a CNN can assess the human tumor microenvironment and predict prognosis directly from histopathological images.

The aim of this study is to review accelerometer wear methods and correlations between accelerometry and physical activity questionnaire data, depending on participant characteristics. We included 57 articles about physical activity measurement by accelerometry and questionnaires. Criteria were to have at least 100 participants of at least 18 years of age with manuscripts available in English. Accelerometer wear methods were compared. Spearman and Pearson correlation coefficients between questionnaires and accelerometers and differences between genders, age categories, and body mass index (BMI) categories were assessed. In most investigations, requested wear time was seven days during waking hours and devices were mostly attached on hips with waist belts. A minimum of four valid days with wear time of at least ten hours per day was required in most studies. Correlations (r = Pearson, ρ = Spearman) of total questionnaire scores against accelerometer measures across individual studies ranged from r = 0.08 to ρ = 0.58 (P < 0.001) for men and from r = −0.02 to r = 0.49 (P < 0.01) for women. Correlations for total physical activity among participants with ages ≤65 ranged from r = 0.04 to ρ = 0.47 (P < 0.001) and from r = 0.16 (P = 0.02) to r = 0.53 (P < 0.01) among the elderly (≥65 years). Few studies investigated stratification by BMI, with varying cut points and inconsistent results. Accelerometers appear to provide slightly more consistent results in relation to self-reported physical activity among men. Nevertheless, due to overall limited consistency, different aspects measured by each method, and differences in the dimensions studied, it is advised that studies use both questionnaires and accelerometers to gain the most complete physical activity information.

Various studies suggest that vitamin D may reduce breast cancer risk. Most studies assessed the effects of dietary intake only, although endogenous production is an important source of vitamin D. Therefore, the measurement of serum 25-hydroxyvitamin D [25(OH)D] better indicates overall vitamin D status. To assess the association of 25(OH)D serum concentrations with post-menopausal breast cancer risk, we used a population-based case–control study in Germany, which recruited incident breast cancer patients aged 50–74 between 2002 and 2005. Information on sociodemographic and breast cancer risk factors was collected by personal interview. For this analysis, we included 1394 cases and 1365 controls, matched on year of birth and time of blood collection. Conditional logistic regression was used to calculate odds ratios (ORs) for breast cancer adjusted for potential confounders. Serum 25(OH)D concentration was significantly inversely associated with post-menopausal breast cancer risk. Compared with the lowest category (<30 nM), OR [95% confidence intervals (CI)] for the higher categories of 25(OH)D (30–45, 45–60, 60–75 and ≥75 nM) were 0.57 (0.45–0.73), 0.49 (0.38–0.64), 0.43 (0.32–0.57) and 0.31 (0.24–0.42), respectively ( P_trend < 0.0001). Analysis using fractional polynomials indicated a non-linear association. The association was stronger in women never using menopausal hormone therapy (HT) compared with past and current users ( P_interaction < 0.0001). Our findings strongly suggest a protective effect for post-menopausal breast cancer through a better vitamin D supply as characterized by serum 25(OH)D measurement, with a stronger inverse association in women with low serum 25(OH)D concentrations (<50 nM).

Background: Multiparity, young age at first childbirth, and breast-feeding are associated with a reduced risk of breast cancer in the general population. The breast cancer predisposition gene, BRCA1, regulates normal cell differentiation. Because mammary gland cells divide and differentiate during pregnancy, reproductive factors may influence breast cancer risk in BRCA1/2 mutation carriers differently than they do in noncarriers. Methods: We performed a retrospective cohort study of 1601 women in the International BRCA1/2 Carrier Cohort Study cohort, all of whom carried a mutation in BRCA1 or BRCA2. Information on reproductive factors was obtained from a questionnaire. At the time of interview 853 subjects were classified with breast cancer. Data were analyzed by using a weighted cohort approach. All statistical tests were two-sided. Results: There was no statistically significant difference in the risk of breast cancer between parous and nulliparous women. Among parous women, an increasing number of full-term pregnancies was associated with a statistically significant decrease in the risk of breast cancer ( P_trend = .008); risk was reduced by 14% (95% confidence interval [CI] = 6% to 22%) for each additional birth. This association was the same for carriers of mutations in either BRCA1 or BRCA2 and was restricted to women older than 40 years. In BRCA2 mutation carriers, first childbirth at later ages was associated with an increased risk of breast cancer compared with first childbirth before age 20 years (20–24 years, hazard ratio [HR] = 2.33 [95% CI = 0.93 to 5.83]; 25–29 years, HR = 2.68 [95% CI = 1.02 to 7.07]; ≥30 years, HR = 1.97 [95% CI = 0.67 to 5.81]), whereas in BRCA1 mutation carriers, first childbirth at age 30 years or later was associated with a reduced risk of breast cancer compared with first childbirth before age 20 years (HR = 0.58 [95% CI = 0.36 to 0.94]). Neither history of interrupted pregnancies (induced abortions or miscarriage) nor history of breast-feeding was statistically significantly associated with the risk of breast cancer. Conclusions: BRCA1 and BRCA2 mutation carriers older than 40 years show a similar reduction in breast cancer risk with increasing parity as non-carriers.