| Table 1. Patient and mass characteristics | |
| Characteristic | Summary |
|---|---|
| Patients | |
| Distinct patient IDs, n | 168 |
| Age at first listed biopsy, years | 24 (20–27) |
| Multiple masses | 31 (18.5%) |
| Multiplicity not recorded | 1 (0.6%) |
| Biopsy records | |
| Biopsy records, n | 172 |
| Greatest dimension, mm | 22 (16–31) |
| Shape | |
| Oval | 132 (76.7%) |
| Irregular | 34 (19.8%) |
| Round | 4 (2.3%) |
| Multilobulated | 1 (0.6%) |
| Not recorded | 1 (0.6%) |
| Margin | |
| Circumscribed | 109 (63.4%) |
| Indistinct | 37 (21.5%) |
| Microlobulated | 11 (6.4%) |
| Lobulated | 3 (1.7%) |
| Angular | 2 (1.2%) |
| Spiculated | 1 (0.6%) |
| Obscured | 1 (0.6%) |
| Mixed descriptors | 5 (2.9%) |
| Not recorded | 3 (1.7%) |
| Echogenicity | |
| Hypoechoic | 126 (73.3%) |
| Heterogeneous | 41 (23.8%) |
| Isoechoic | 1 (0.6%) |
| Complex cystic and solid | 1 (0.6%) |
| Not recorded | 3 (1.7%) |
| Posterior features | |
| Enhancement | 30 (17.4%) |
| Shadowing | 1 (0.6%) |
| None recorded explicitly | 3 (1.7%) |
| Not recorded | 138 (80.2%) |
| Pathology | |
| Fibroadenoma, including variants | 138 (80.2%) |
| Benign phyllodes tumor | 1 (0.6%) |
| Other benign pathology | 29 (16.9%) |
| Borderline phyllodes tumor | 1 (0.6%) |
| Invasive ductal carcinoma | 3 (1.7%) |
| Values are median (IQR) or n (%), unless otherwise indicated. Patient-level denominator: 168; biopsy-record denominator: 172. | |
| Blank descriptors remain 'Not recorded.' Repeated IDs with differing ages require linkage verification. | |
| 11 records have follow-up, pain, or missing indications; palpable status requires confirmation. | |
Malignancy yield of biopsy for palpable breast masses
Women aged 30 years or younger
Statistical analysis
Analyses included biopsy records from women aged 30 years or younger. Age at the first listed biopsy and the presence of multiple masses were summarized once per patient; other characteristics were summarized per biopsy record. Continuous variables were described using medians and interquartile ranges (IQRs), and categorical variables using counts and percentages. Missing descriptors were reported explicitly.
Malignancy yield was estimated overall and for BI-RADS categories 4A, 4B, 4C, and 5, with two-sided Clopper–Pearson exact 95% confidence intervals (CIs). These intervals assume independence between biopsy records. Sensitivity analyses selected one record per patient by highest BI-RADS category and then largest diameter, included all ages, and restricted the cohort to records explicitly documenting a palpable indication. Borderline phyllodes pathology was classified as nonmalignant.
For records flagged as upgraded from BI-RADS 2 or 3 because of growth, percentage change was calculated as 100 × (current diameter − previous diameter) / previous diameter, using the tabulated dimensions in millimeters. Growth was summarized by median (IQR), mean (standard deviation [SD]), and the proportion increasing by at least 20%, with an exact 95% CI. The threshold was a descriptive reference because examination intervals were unavailable. A sensitivity calculation used the largest dimensions in the accompanying measurement text. The two phyllodes tumors were described individually without a formal statistical comparison. No regression or hypothesis testing was performed because malignancies were sparse. Analyses used R version 4.2.3.
Results
Patient and mass characteristics
The analysis included 172 biopsy records with 168 distinct patient IDs, after excluding one record at age 31. Median age was 24 years (IQR, 20–27), and median mass dimension was 22 mm (IQR, 16–31). Most masses were oval and fibroadenomas (Table 1).
Malignancy yield
Malignancy was identified in 3/172 biopsy records (1.7%; 95% CI, 0.4–5.0%). All three cancers were invasive ductal carcinomas at age 30: one BI-RADS 4C and two BI-RADS 5. No cancers were observed in BI-RADS 4A or 4B (Table 2). The 4C and 5 estimates were based on only one and two records, respectively.
| Table 2. Malignancy yield by BI-RADS category | ||||
| Category | Biopsy records | Malignancies | Yield | 95% CI |
|---|---|---|---|---|
| Overall (BI-RADS 4 and 5) | 172 | 3 | 1.7% | 0.4–5.0% |
| BI-RADS 4A | 155 | 0 | 0.0% | 0.0–2.4% |
| BI-RADS 4B | 14 | 0 | 0.0% | 0.0–23.2% |
| BI-RADS 4C | 1 | 1 | 100.0% | 2.5–100.0% |
| BI-RADS 5 | 2 | 2 | 100.0% | 15.8–100.0% |
| Yield concerns performed biopsies. Exact CIs use a working independence assumption; borderline phyllodes pathology is nonmalignant. | ||||
Yield was 3/168 (1.8%; 95% CI, 0.4–5.1%) with one record per patient ID, 3/173 (1.7%) including age 31, and 3/161 (1.9%) among records explicitly documenting a palpable indication.
Size changes among upgraded masses
The 44 growth-flagged records were all BI-RADS 4A and nonmalignant, including one benign and one borderline phyllodes tumor. Median growth was 26.8% (IQR, 14.5–46.3%), and mean growth was 36.7% (SD, 33.5%). An increase of at least 20% occurred in 27/44 records (61.4%; 95% CI, 45.5–75.6%). There were no malignant upgrades for a separate growth comparison.
The benign phyllodes tumor increased from 29 to 56 mm (93.1%), and the borderline phyllodes tumor increased from 15 to 41 mm (173.3%). Both are highlighted in Figure 1 and are included in the 44 growth records, leaving 42 other records.
| Table 3. Size changes among masses upgraded to biopsy | ||
| Measure | Summary | 95% CI |
|---|---|---|
| Growth-flagged records, n | 44 | — |
| Previous greatest dimension, mm | 21.0 (17.8–25.2) | — |
| Current greatest dimension, mm | 28.0 (22.5–34.2) | — |
| Absolute change, mm | 5.0 (3.0–10.0) | — |
| Percentage change, median (IQR) | 26.8 (14.5–46.3)% | — |
| Percentage change, mean (SD) | 36.7 (33.5)% | — |
| Percentage change, range | 0.0–173.3% | — |
| Increase ≥20% | 27/44 (61.4%) | 45.5–75.6% |
| Increase <20% | 17/44 (38.6%) | — |
| Malignancy yield | 0/44 (0.0%) | 0.0–8.0% |
| Continuous values are median (IQR), unless labeled otherwise. Upgrade status was missing for 2 other records. Examination intervals were unavailable. | ||
| Four growth records have conflicting numeric and text dimensions. Using text maxima gives 30/44 (68.2%) meeting 20%; these measurements require verification. One flagged record has 0% growth. | ||
Summary
Malignancy yield was 1.7%, with all three cancers in BI-RADS 4C or 5 at age 30. No cancers occurred among the 44 growth-flagged records, although the upper 95% confidence limit was 8.0%. Based on tabulated measurements, 61.4% met the 20% growth reference. Patient linkage, palpable eligibility, and conflicting dimensions need confirmation before finalizing the abstract. These biopsy-selected data do not establish the safety of avoiding biopsy or reducing surveillance.