A Simple Model to Assess the Probability of Invasion in Ductal Carcinoma In Situ of the Breast Diagnosed by Needle Biopsy

Abstract

Objectives. The aim of the study was to develop a clinical prediction model for assessing the probability of having invasive cancer in the definitive surgical resection specimen in patients with biopsy diagnosis of ductal carcinoma in situ (DCIS) of the breast, to facilitate decision making regarding axillary surgery. Methods. In 349 women with DCIS, predictors of invasion in the definitive resection specimen were identified. A model to predict the probability of invasion was developed and subsequently simplified to divide patients into two risk categories. The model's performance was validated on another patient population. Results. Multivariate logistic regression revealed four independent predictors of invasion: (i) suspicious (micro)invasion in the biopsy specimen; (ii) visibility of the lesion on ultrasonography; (iii) size of the lesion on mammography >30 mm; (iv) clinical palpability of the lesion. The actual frequency of invasion in the high-risk patient group in the test and validation population was 52.6% and 48.3%, respectively; in the low-risk group it was 16.8% and 7.1%, respectively. Conclusion. The model proved to have good performance. In patients with a low probability of invasion, an axillary procedure can be omitted without a substantial risk of additional surgery.

A prediction model for underestimation of invasive breast cancer after a biopsy diagnosis of ductal carcinoma in situ: based on 2892 biopsies and 589 invasive cancers

“…A suspected invasive component has also only been reported in a limited number of studies. 23,24 ; all found a high risk for underestimation for biopsies with a suspected component.…”

“…Palpability of the lesion has consistently been reported as a risk factor, which this study could confirm. 15,16,18,19,[22][23][24]26,37 The BI-RADS score is an assessment categorization that should give an indication of the likelihood of cancer based on the interpretation of the radiologist. We showed that it is associated with the underestimation rate; the difference between BI-RADS score 4 and 5 was 23% in underestimation rate, which is much larger than the 7-8% found by others.…”

“…The sensitivity or the AUC was higher in several other studies. 14,17,22,24,28 Each study with a prediction model used different risk factors and therefore the models are not easily comparable. This has also been demonstrated in external validation of studies that applied published models to their cases; one study demonstrated a tendency towards lower or higher numbers of observed underestimates than expected 29 , and another previous study demonstrated validation AUCs of 0.59-0.66, whereas the studies they validated reported validities of 0.70-0.85 14 .…”

“…The most frequently reported risk factors for underestimation are DCIS grade and factors found with radiological diagnostic work-up, such as the size of the lesion, mass on mammography or ultrasonography, and the Breast Imaging Reporting and Data System (BI-RADS) score. [14][15][16][17][18][19][20][21][22][23][24][25][26][27] Furthermore, these studies reported that the risk of underestimation was www.nature.com/bjcListed are 12 studies with at least 100 cases of underestimation d References in bold are of the five studies that developed a prediction model e Reference 16 presents results of random-effect logistic regression models in a meta-analysis f The multivariable significant variables of ref. 23 were used in a prediction model, as described in ref.…”

“…An overview of the found risk factors for underestimation is given in Table 1. Based on risk factors, several studies developed prediction models with the purpose to select patients for SLN biopsy 14,17,24,[28][29][30] .…”