INTRODUCTION: Risk-stratified screening (RSS) scheduling may facilitate more effective use of same-day diagnostic testing for potentially abnormal mammograms, thereby reducing the need for follow-up appointments (recall). Our simulation study assessed the potential impact of RSS scheduling on patients recommended for same-day diagnostics. METHODS: We used a discrete event simulation to model workflow at a high-volume breast imaging center, incorporating artificial intelligence (AI)-triaged same-day diagnostic workups after screening mammograms. The RSS design sequences patients in the daily screening schedule using cancer risk categories developed from Tyrer-Cuzick and deep learning model scores. We compared recall variance, required hours of operation to accommodate all patients, and patient wait times using traditional (random) and RSS schedules. RESULTS: The baseline simulation included 60 daily patients, with an average of 42% receiving screening mammograms and 11% (about three patients) being recommended for diagnostic workups. Compared with traditional scheduling, RSS scheduling reduces recall variance by up to 30% (1.98 versus 2.82, P < .05). With same-day diagnostics, RSS scheduling had a modest impact, increasing the number of patients served within normal operating hours by up to 1.3% (55.4 versus 54.7, P < .05), decreasing necessary operational hours by 12 min (10.3 versus 10.5 hours, P < .05), and increasing patient waiting times by an average of 2.4 min (0.24 versus 0.20 hours, P < .05). CONCLUSION: Our simulation study suggests that RSS scheduling could reduce recall variance. This approach might enable same-day diagnostics using AI triage by accommodating patients within normal operating hours.

Background

Increasing focus on potentially unnecessary diagnosis and treatment of certain breast cancers prompted our investigation of whether clinical and mammographic features predictive of invasive breast cancer versus ductal carcinoma in situ (DCIS) differ by age.

Methods

We analyzed 1,475 malignant breast biopsies, 1,063 invasive and 412 DCIS, from 35,871 prospectively collected consecutive diagnostic mammograms interpreted at University of California, San Francisco between 1/6/1997 and 6/29/2007. We constructed three logistic regression models to predict the probability of invasive cancer versus DCIS for the following groups: women ≥ 65 (older group), women 50-64 (middle age group), and women < 50 (younger group). We identified significant predictors and measured the performance in all models using area under the receiver operating characteristic curve (AUC).

Results

The models for older and the middle age groups performed significantly better than the model for younger group (AUC = 0.848 vs, 0.778; p = 0.049 and AUC = 0.851 vs, 0.778; p = 0.022, respectively). Palpability and principal mammographic finding were significant predictors in distinguishing invasive from DCIS in all age groups. Family history of breast cancer, mass shape and mass margins were significant positive predictors of invasive cancer in the older group whereas calcification distribution was a negative predictor of invasive cancer (i.e. predicted DCIS). In the middle age group--mass margins, and in the younger group--mass size were positive predictors of invasive cancer.

Conclusions

Clinical and mammographic finding features predict invasive breast cancer versus DCIS better in older women than younger women. Specific predictive variables differ based on age.