Evaluating the Age-Based Recommendations for Long-Term Follow-Up in Breast Cancer.

BACKGROUND: After 5 years of annual follow-up following breast cancer, Dutch guidelines are age based: annual follow-up for women <60 years, 60-75 years biennial, and none for >75 years. We determined how the risk of recurrence corresponds to these consensus-based recommendations and to the risk of primary breast cancer in the general screening population. SUBJECTS, MATERIALS, AND METHODS: Women with early-stage breast cancer in 2003/2005 were selected from the Netherlands Cancer Registry (n = 18,568). Cumulative incidence functions were estimated for follow-up years 5-10 for locoregional recurrences (LRRs) and second primary tumors (SPs). Risks were compared with the screening population without history of breast cancer. Alternative cutoffs for age were determined by log-rank tests. RESULTS: The cumulative risk for LRR/SP was lower in women <60 years (5.9%, 95% confidence interval [CI] 5.3-6.6) who are under annual follow-up than for women 60-75 (6.3%, 95% CI 5.6-7.1) receiving biennial visits. All risks were higher than the 5-year risk of a primary tumor in the screening population (ranging from 1.4% to 1.9%). Age cutoffs <50, 50-69, and > 69 revealed better risk differentiation and would provide more risk-based schedules. Still, other factors, including systemic treatments, had an even greater impact on recurrence risks. CONCLUSION: The current consensus-based recommendations use suboptimal age cutoffs. The proposed alternative cutoffs will lead to a more balanced risk-based follow-up and thereby more efficient allocation of resources. However, more factors should be taken into account for truly individualizing follow-up based on risk for recurrence. IMPLICATIONS FOR PRACTICE: The current age-based recommendations for breast cancer follow-up after 5 years are suboptimal and do not reflect the actual risk of recurrent disease. This results in situations in which women with higher risks actually receive less follow-up than those with a lower risk of recurrence. Alternative cutoffs could be a start toward risk-based follow-up and thereby more efficient allocation of resources. However, age, or any single risk factor, is not able to capture the risk differences and therefore is not sufficient for determining follow-up. More risk factors should be taken into account for truly individualizing follow-up based on the risk for recurrence.

Risk-based breast cancer follow-up stratified by age.

Although personalization of cancer care is recommended, current follow-up after the curative treatment of breast cancer is consensus-based and not differentiated for base-line risk. Every patient receives annual follow-up for 5 years without taking into account the individual risk of recurrence. The aim of this study was to introduce personalized follow-up schemes by stratifying for age. Using data from the Netherlands Cancer Registry of 37 230 patients with early breast cancer between 2003 and 2006, the risk of recurrence was determined for four age groups (<50, 50-59, 60-69, >70). Follow-up was modeled with a discrete-time partially observable Markov decision process. The decision to test for recurrences was made two times per year. Recurrences could be detected by mammography as well as by self-detection. For all age groups, it was optimal to have more intensive follow-up around the peak in recurrence risk in the second year after diagnosis. For the first age group (<50) with the highest risk, a slightly more intensive follow-up with one extra visit was proposed compared to the current guideline recommendation. The other age groups were recommended less visits: four for ages 50-59, three for 60-69, and three for ≥70. With this model for risk-based follow-up, clinicians can make informed decisions and focus resources on patients with higher risk, while avoiding unnecessary and potentially harmful follow-up visits for women with very low risks. The model can easily be extended to take into account more risk factors and provide even more personalized follow-up schedules.

Comparison of Logistic Regression and Bayesian Networks for Risk Prediction of Breast Cancer Recurrence.

PURPOSE: For individualized follow-up, accurate prediction of locoregional recurrence (LRR) and second primary (SP) breast cancer risk is required. Current prediction models employ regression, but with large data sets, machine-learning techniques such as Bayesian Networks (BNs) may be better alternatives. In this study, logistic regression was compared with different BNs, built with network classifiers and constraint- and score-based algorithms. METHODS: Women diagnosed with early breast cancer between 2003 and 2006 were selected from the Netherlands Cancer Registry (NCR) ( N = 37,320). BN structures were developed using 1) Bayesian network classifiers, 2) correlation coefficients with different cutoffs, 3) constraint-based learning algorithms, and 4) score-based learning algorithms. The different models were compared with logistic regression using the area under the receiver operating characteristic curve, an external validation set obtained from the NCR from 2007 and 2008 ( N = 12,308), and subgroup analyses for a high- and low-risk group. RESULTS: The BNs with the most links showed the best performance in both LRR and SP prediction (c-statistic of 0.76 for LRR and 0.69 for SP). In the external validation, logistic regression generally outperformed the BNs in both SP and LRR (c-statistic of 0.71 for LRR and 0.64 for SP). The differences were nonetheless small. Although logistic regression performed best on most parts of the subgroup analysis, BNs outperformed regression with respect to average risk for SP prediction in low- and high-risk groups. CONCLUSIONS: Although estimates of regression coefficients depend on other independent variables, there is no assumed dependence relationship between coefficient estimators and the change in value of other variables as in the case of BNs. Nonetheless, this analysis suggests that regression is still more accurate or at least as accurate as BNs for risk estimation for both LRRs and SP tumors.

Personalisation of breast cancer follow-up: a time-dependent prognostic nomogram for the estimation of annual risk of locoregional recurrence in early breast cancer patients.

The objective of this study was to develop and validate a time-dependent logistic regression model for prediction of locoregional recurrence (LRR) of breast cancer and a web-based nomogram for clinical decision support. Women first diagnosed with early breast cancer between 2003 and 2006 in all Dutch hospitals were selected from the Netherlands Cancer Registry (n = 37,230). In the first 5 years following primary breast cancer treatment, 950 (2.6 %) patients developed a LRR as first event. Risk factors were determined using logistic regression and the risks were calculated per year, conditional on not being diagnosed with recurrence in the previous year. Discrimination and calibration were assessed. Bootstrapping was used for internal validation. Data on primary tumours diagnosed between 2007 and 2008 in 43 Dutch hospitals were used for external validation of the performance of the nomogram (n = 12,308). The final model included the variables grade, size, multifocality, and nodal involvement of the primary tumour, and whether patients were treated with radio-, chemo- or hormone therapy. The index cohort showed an area under the ROC curve of 0.84, 0.77, 0.70, 0.73 and 0.62, respectively, per subsequent year after primary treatment. Model predictions were well calibrated. Estimates in the validation cohort did not differ significantly from the index cohort. The results were incorporated in a web-based nomogram ( http://www.utwente.nl/mira/influence ). This validated nomogram can be used as an instrument to identify patients with a low or high risk of LRR who might benefit from a less or more intensive follow-up after breast cancer and to aid clinical decision making for personalised follow-up.