Diagnostic Performance and Outcomes of Breast Imaging during Pregnancy and Lactation in an Academic Clinical Practice

Author

Lu Mao

Statistical analysis

Patient demographics

Patient demographics are summarized in Table 1.

Table 1:

Patient demographics.
Characteristic N = 1,0001
Sex assigned at birth
    Female 1,000 (100%)
Age (years) at first pregnancy
    11-14 3 (0.3%)
    15-19 54 (5.4%)
    20-24 122 (12%)
    25-29 296 (30%)
    30-34 381 (38%)
    35-39 132 (13%)
    ≥40 12 (1.2%)
Race
    American Indian or Alaska Native 8 (0.8%)
    Asian 70 (7.0%)
    Black or African American 58 (5.8%)
    Native Hawaiian or Other Pacific Islander 7 (0.7%)
    White 838 (84%)
    Other 19 (1.9%)
Ehtnicity
    Not Hispanic or Latino 916 (92%)
    Hispanic or Latino 82 (8.2%)
    Unknown 2 (0.2%)
BMI during imaging 25.7 (22.8, 29.5)
1 n (%); Median (IQR)

BI-RADS 3 in pregnant vs. lactating by modality

The frequency and rate of BI-RADS 3 in pregnant and lactating women are summarized in Table 2 by imaging modality.

Table 2: Frequency and rate of BIRADS-3 in pregnant vs lactating subjects by imaging modality.
Modality Pregnant Lactating
Mammo-Screening 0/0 0/89 (0%)
Mammo-Diagnostic 0/16 (0%) 40/397 (10.1%)
Ultrasound-Screening 0/2 (0%) 0/2 (0%)
Ultrasound-Diagnostic 34/244 (13.9%) 88/605 (14.5%)
MRI-Screening 0/0 1/50 (2%)

Pathology vs final diagnosis

Cross-tabulation of biopsy (Yes vs No) against final diagnosis status is shown in Table 3.

Table 3:

Biopsy vs final diagnosis.
Final Diagnosis Total p-value1
Benign Malignant Unknown
Biopsy



<0.001
    No 792 (90%) 0 (0%) 84 (9.6%) 876 (100%)
    Yes 110 (89%) 14 (11%) 0 (0%) 124 (100%)
Total 902 (90%) 14 (1.4%) 84 (8.4%) 1,000 (100%)
1 Fisher’s exact test

In biopsied patients, 110 (89%) have benign diagnosis, comparable to 902 (90%) overall.

Malignancy rate after BI-RADS 3

Across all modalities, 123 patients got BI-RADS 3. Among them, none (0%) had a final malignant diagnosis.

Mammogaphy with tomosynthesis (DBT) and digital mammography (DM)

Table 4 below shows that the diagnostic performance (specificity) of DBT vs DM is similar.

Table 4: Diagnostic performance (specificity) of DBT vs DM.

Type Result n
DBT False positive 26
DBT True negative 157
DBT True positive 4
DM False positive 26
DM True negative 151
DM True positive 9
(a) Specificity (95% CI)
DBT DM P
0.858 (0.797, 0.903) 0.853 (0.79, 0.9) 1