The transition from screen-film to digital mammography may have altered diagnostic evaluation of women following a positive screening examination. This study compared the use and timeliness of diagnostic imaging and biopsy for women screened with screen-film or digital mammography. Data were obtained from 35,321 positive screening mammograms on 32,087 women aged 40-89 years, from 22 breast cancer surveillance consortium facilities in 2005-2008. Diagnostic pathways were classified by their inclusion of diagnostic mammography, ultrasound, magnetic resonance imaging, and biopsy. We compared time to resolution and frequency of diagnostic pathways by patient characteristics, screening exam modality, and radiology facility. Between-facility differences were evaluated by computing the proportion of mammograms receiving follow-up with a particular pathway for each facility and examining variation in these proportions across facilities. Multinomial logistic regression adjusting for age, calendar year, and facility compared odds of follow-up with each pathway. The median time to resolution of a positive screening mammogram was 10 days. Compared to screen-film mammograms, digital mammograms were more frequently followed by only a single diagnostic mammogram (46 vs. 36 %). Pathways following digital screening mammography were also less likely to include biopsy (16 vs. 20 %). However, in adjusted analyses, most differences were not statistically significant (p = 0.857 for mammography only; p = 0.03 for biopsy). Substantial variability in diagnostic pathway frequency was seen across facilities. For instance, the frequency of evaluation with diagnostic mammography alone ranged from 23 to 55 % across facilities. Differences in evaluation of positive digital and screen-film screening mammograms were minor, and appeared to be largely attributable to substantial variation between radiology facilities. To guide health systems in their efforts to eliminate practices that do not contribute to effective care, we need further research to identify the causes of this variation and the best evidence-based approach for follow-up.

Accurate lesion segmentation is critical in stroke rehabilitation research for the quantification of lesion burden and accurate image processing. Current automated lesion segmentation methods for T1-weighted (T1w) MRIs, commonly used in stroke research, lack accuracy and reliability. Manual segmentation remains the gold standard, but it is time-consuming, subjective, and requires neuroanatomical expertise. We previously released an open-source dataset of stroke T1w MRIs and manually-segmented lesion masks (ATLAS v1.2, N = 304) to encourage the development of better algorithms. However, many methods developed with ATLAS v1.2 report low accuracy, are not publicly accessible or are improperly validated, limiting their utility to the field. Here we present ATLAS v2.0 (N = 1271), a larger dataset of T1w MRIs and manually segmented lesion masks that includes training (n = 655), test (hidden masks, n = 300), and generalizability (hidden MRIs and masks, n = 316) datasets. Algorithm development using this larger sample should lead to more robust solutions; the hidden datasets allow for unbiased performance evaluation via segmentation challenges. We anticipate that ATLAS v2.0 will lead to improved algorithms, facilitating large-scale stroke research.