Objective:

To e

xplore the calcification characteristics that may predict the invasiveness of breast lesions based on the morphology and distribution of calcification in mammography.

Methods:

A total of 267 patients with suspicious calcification detected by mammography were analyzed retrospectively. All these patients were females

with an average age of (48.1±9.7) years and were also diagnosed by biopsy. According to the pathological results

invasive carcinoma and ductal carcinoma in situ with microinvasion were defined as invasive lesions while ductal carcinoma in situ and various benign lesions were defined as non-invasive lesions. The morphology and distribution of calcification that may predict the invasiveness of the lesions were explored.

χ

2

test was used in univariate analysis while logistic regression analysis was used in multivariate analysis to establish a predictive model. Then

receiver operating characteristic (ROC) curve was drawn and area under curve (AUC) was calculated to assess the predictive value of the model. At last

Hosmer-Lemeshow test was used to assess the overall fit of the model.

Results:

Univariate analysis showed that the suspicious calcification with punctate/round calcification and amorphous calcification had predictive value for non-invasive breast lesions (

χ

2

=10.567

P

=0.001;

χ

2

=31.153

P

＜0.001)

while fine linear or fine linear branching calcification and suspicious calcification in segmental distribution had predictive value for invasive breast lesions (

χ

2

=36.275

P

＜0.001;

χ

2

=5.147

P

=0.023). Multivariate logistic regression analysis showed that amorphous calcification was independent predictor of non-invasive breast lesions (OR=0.273

95% CI 0.135-0.553

P

＜0.001)

while fine linear or fine linear branching calcification was independent predictor of invasive breast lesions (OR=5.211

95% CI 1.819-14.931

P

=0.002). The AUC of predictive model was 0.759 (95% CI 0.686-0.833) and Hosmer-Lemeshow test showed a good fit of the model (

P

＞0.05).

Conclusion:

It is feasible to predict the invasiveness of breast lesions based on the morphology and distribution of calcification in mammography and which significance is to take corresponding intervention or management measures for lesions of different natures as early as possible according to mammography screening.