Search

Purpose Surgical excision is currently recommended for all occurrences of atypical ductal hyperplasia (ADH) found on core needle biopsies for malignancy diagnoses and treatment of lesions. The excision of all ADH lesions may lead to overtreatment, which results in invasive surgeries for benign lesions in many women. A machine learning method to predict ADH upgrade may help clinicians and patients decide whether combined active surveillance and hormonal therapy is a reasonable alternative to surgical excision. Methods The following six machine learning models were developed to predict ADH upgrade from core needle biopsy: gradient-boosting trees, random forest, radial support vector machine (SVM), weighted K-nearest neighbors (KNN), logistic elastic net, and logistic regression. The study cohort consisted of 128 lesions from 124 women at a tertiary academic care center in New Hampshire who had ADH on core needle biopsy and who underwent an associated surgical excision from 2011 to 2017. Results The best-performing models were gradient-boosting trees (area under the curve [AUC], 68%; accuracy, 78%) and random forest (AUC, 67%; accuracy, 77%). The top five most important features that determined ADH upgrade were age at biopsy, lesion size, number of biopsies, needle gauge, and personal and family history of breast cancer. Using the random forest model, 98% of all malignancies would have been diagnosed through surgical biopsies, whereas 16% of unnecessary surgeries on benign lesions could have been avoided (ie, 87% sensitivity at 45% specificity). Conclusion These results add to the growing body of support for machine learning models as useful aids for clinicians and patients in decisions about the clinical management of ADH.

The cell-division cycle of Caulobacter crescentus depends on periodic activation and deactivation of the essential response regulator CtrA. Although CtrA is critical for transcription during some parts of the cell cycle, its activity must be eliminated before chromosome replication because CtrA also blocks the initiation of DNA replication. CtrA activity is down-regulated both by dephosphorylation and by proteolysis, mediated by the ubiquitous ATP-dependent protease ClpXP. Here we demonstrate that proteins needed for rapid CtrA proteolysis in vivo form a phosphorylation-dependent and cyclic diguanylate (cdG)-dependent adaptor complex that accelerates CtrA degradation in vitro by ClpXP. The adaptor complex includes CpdR, a single-domain response regulator; PopA, a cdG-binding protein; and RcdA, a protein whose activity cannot be predicted. When CpdR is unphosphorylated and when PopA is bound to cdG, they work together with RcdA in an all-or-none manner to reduce the Km of CtrA proteolysis 10-fold. We further identified a set of amino acids in the receiver domain of CtrA that modulate its adaptor-mediated degradation in vitro and in vivo. Complex formation between PopA and CtrA depends on these amino acids, which reside on alpha-helix 1 of the CtrA receiver domain, and on cdG binding by PopA. These results reveal that each accessory factor plays an essential biochemical role in the regulated proteolysis of CtrA and demonstrate, to our knowledge, the first example of a multiprotein, cdG-dependent proteolytic adaptor.