Pathologic tumor size is significant in the treatment of breast carcinoma and is routinely measured on excision.
To analyze the need for measuring size of invasive mammary carcinoma on biopsy.
Nine hundred twenty-two cases of invasive carcinoma whose size was measured (greatest linear measurement) on biopsy and excision was correlated, including imaging when available (110 cases).
Patient mean age was 62 years. Most (90%; 830 of 922) carcinomas were ductal and sampled by ultrasound and graded as follows: well, 13% (113 of 922); moderately, 58% (532 of 922), and poorly differentiated, 28% (258 of 922); 19 microinvasive not graded. Tumor mean size was 7.5 mm on biopsy and 14.4 mm on excision. Biopsy modality was as follows: ultrasound, 7.8 mm (92%, 844 of 922); mammotome, 3.3 mm (7%, 65 of 922); and magnetic resonance imaging, 5.9 mm (1%, 13 of 922). Size comparison on biopsy versus excision was biopsy > excision: 8% (72 of 922), biopsy = excision: 10% (95 of 922), and biopsy < excision: 82% (755 of 922). Half (36 of 72) of the biopsy > excision tumors were less than 5 mm, 96% (726 of 755) of biopsy < excision tumors were greater than 5 mm, while those equal on both were predominantly (88%, 84 of 95) less than 10 mm, 20% (19 of 95) of which were microinvasive. Stage changed in 600 cases, staging based on excision in 581 (63%), and staging based on biopsy in 19 (2%). Radiologic-pathologic correlation (n = 110) showed perfect concordance in 11 (10%), 83 (75%) were ±1 to 2 mm, and 16 (15%) were ± more than 3 mm. Difference between the biopsy and excision ranged from a lower limit of 1.3 mm for T1a tumors to 18 mm for T2.
While most carcinomas are larger on excision, 18% (167 of 922) are larger or equal on biopsy. Factors predictive of biopsy > excision tumors include stage 1 tumors (P < .001), especially less than 5 mm, and sampled by mammotome. We recommend measuring invasive carcinoma on biopsy and excision.
After lymph node status, pathologic tumor size is the next most important prognostic factor in the staging and treatment of invasive breast carcinoma. Thus, it is incumbent upon the pathologist to accurately measure the size of the invasive carcinoma to the closest millimeter. With increased and improved screening, most invasive breast cancers are initially diagnosed and sampled on needle core biopsies. There is much variability among pathologists whether or not to measure invasive carcinoma at the time of biopsy. In certain scenarios, depending on the needle gauge and tumor size, some breast cancers may need to be measured on biopsy. As per the most recent College of American Pathologists guideline, given that most invasive breast cancers are larger on surgical excision, size is routinely determined on these specimens, and recommended but not required on biopsy, unless faced with the afore mentioned scenario above.1 Few studies have attempted to evaluate the relationship between tumor size on breast biopsy to corresponding excision specimens.2–4 The goal of our study was to evaluate this relationship, particularly in relation to the method of biopsy, including ultrasound guided, mammotome, and magnetic resonance imaging (MRI), and establish much needed guidelines and validity for this practice.
METHODS
We initiated the practice of measuring invasive carcinoma on biopsy in 2017. Using the pathology database, we retrospectively identified 922 cases from 873 patients with follow-up excision specimens (lumpectomy/mastectomy) from 2017 to 2019. In all cases, we ensured that the location of the invasive carcinoma on biopsy and excision (with biopsy site changes) matched, especially in patients with multiple ipsilateral and or contralateral biopsy. Neoadjuvant-treated breast cases were excluded because tumor size on biopsy and excision may not correlate due to treatment effects. Clinical features such as age, history, and size on imaging were gathered. Pathologic features including presence of ductal carcinoma in situ, histologic type, and grade of invasive carcinoma were reviewed. The greatest linear extent of invasive carcinoma was measured and compared microscopically for all biopsies and excisions (Figure 1, A and B), and with imaging when available. Cases with fragmented invasive carcinoma for which reliable tumor size could not be measured were excluded. In excision cases where the invasive tumor was too large to be submitted in a few paraffin blocks, or if size could not be reconstructed based on gross examination, the gross size was considered more accurate and corroborated with supplemental radiologic correlation.5,6 Tumors were staged based on the largest tumor dimension, regardless of whether present on biopsy or excision. Categoric data comparisons were performed using the X2 test and Z test. Comparison of data between/among groups was performed using t test and analysis of variance test. Correlation was established using Pearson test.
A, Greatest linear extent of invasive carcinoma on biopsy measuring 6 mm with associated ductal carcinoma in situ. B, Greatest linear extent of invasive carcinoma on excision measuring 2 mm with associated ductal carcinoma in situ. Staging is best performed on biopsy (hematoxylin-eosin, original magnification ×40).
Figure 2. A, Greatest linear extent of invasive carcinoma on biopsy measuring 6 mm. B, Greatest linear extent of invasive carcinoma on excision measuring 6 mm, equal to biopsy (hematoxylin-eosin, original magnification ×40).
Figure 3. A, Greatest linear extent of invasive carcinoma on biopsy measuring 4.5 mm. B, Greatest linear extent of invasive carcinoma on biopsy measuring 6 mm. Staging is best performed on excision (hematoxylin-eosin, original magnification ×40).
A, Greatest linear extent of invasive carcinoma on biopsy measuring 6 mm with associated ductal carcinoma in situ. B, Greatest linear extent of invasive carcinoma on excision measuring 2 mm with associated ductal carcinoma in situ. Staging is best performed on biopsy (hematoxylin-eosin, original magnification ×40).
Figure 2. A, Greatest linear extent of invasive carcinoma on biopsy measuring 6 mm. B, Greatest linear extent of invasive carcinoma on excision measuring 6 mm, equal to biopsy (hematoxylin-eosin, original magnification ×40).
Figure 3. A, Greatest linear extent of invasive carcinoma on biopsy measuring 4.5 mm. B, Greatest linear extent of invasive carcinoma on biopsy measuring 6 mm. Staging is best performed on excision (hematoxylin-eosin, original magnification ×40).
RESULTS
The age of the patients ranged from 27 to 94 years (mean = 62 years). The size of the invasive tumors ranged from a lower limit of being microinvasive (<1 mm) to 23 mm on biopsy and 100 mm on excision. By histology, most invasive carcinomas were ductal (90%, 830 of 922), the remainder being lobular. Table 1 shows the Nottingham grades, including percentage of cases with Ductal carcinoma in situ further sorted by tumor size on biopsy versus excision. Within the ductal category, special types were identified as follows: mucinous, 47; micropapillary, 71; apocrine, 17; anaplastic, 20; tubular, 13; metaplastic, 3; tubulo-lobular, 12.
Tumor size (922) was larger on biopsy than excision in 72 of 922 (8%) (Figure 1, A and B), equal in both in 95 of 922 (10%) (Figure 2, A and B), the remaining 755 of 922 (82%) cases being larger on excision than biopsy (Figure 3, and B) (Table 1). Lower stage tumors (T1mi and T1a) comprised a high proportion of the tumors that were larger than (37% for stage T1a; 36 of 97) or equal (100% for stage T1mi; 19 of 19; and 33% for stage T1a; 32 of 97) to excision, and this was statistically significant (Figure 4). In contrast, higher stage tumors (T1c, T2, and above) comprised a higher proportion of tumors that were larger on excision than core biopsy (94% for stage 1c; 345 of 366; and 100% for stage 2 and above; 192 of 192; Figure 3, A and B).
Percentage of biopsy > excision, biopsy = excision, and biopsy < excision classified by stage. T1mi tumors have statistically significant higher percentage of core biopsy = excision versus tumors of other stages (Z test, P < .001). T1a tumors have statistically significant higher percentage of core biopsy ≥ excision and lower percentage of core biopsy < excision versus tumors of other stages (X2 test, P < .001). T1c tumors have statistically significant lower percentage of core biopsy ≥ excision and higher percentage of core biopsy < excision versus tumors of other lower stages (X2 test, P < .001). T2 and above tumors have statistically significant higher percentage of core biopsy < excision versus tumors of other stages (Z test, P < .001). Abbreviation: Cbx, core biopsy.
Percentage of biopsy > excision, biopsy = excision, and biopsy < excision classified by stage. T1mi tumors have statistically significant higher percentage of core biopsy = excision versus tumors of other stages (Z test, P < .001). T1a tumors have statistically significant higher percentage of core biopsy ≥ excision and lower percentage of core biopsy < excision versus tumors of other stages (X2 test, P < .001). T1c tumors have statistically significant lower percentage of core biopsy ≥ excision and higher percentage of core biopsy < excision versus tumors of other lower stages (X2 test, P < .001). T2 and above tumors have statistically significant higher percentage of core biopsy < excision versus tumors of other stages (Z test, P < .001). Abbreviation: Cbx, core biopsy.
The tumor stage of the invasive carcinoma changed in 600 cases (65%, 600 of 922), staging based on excision in 581 (63%, 581 of 922) cases, and vice versa in 19 (2%, 19 of 922) as shown in Figure 4. Next, we retrieved the biomarker profile of the cases in our series as shown on Table 2. Most of the cases (788 of 922; 85%) were estrogen receptor positive. Of the small subset of triple negative (92 of 922; 10%) and human epidermal growth factor receptor 2 (HER2)-positive tumors (92 of 922; 10%), only 29 (16%) cases were larger or equal in size on core biopsy compared with excisions, all of which were T1a or T1b tumors. The remaining 155 (86%) cases were evenly distributed between T1c and T2 tumors. In approximately half of these cases, the core biopsy was off by a few millimeters from the final size on excision, such that while most staged as T1c tumors on core biopsy, 78 turned out to be T2 tumors on excision.
Biopsy modalities included the following techniques for specific indications with particular gauge needles: USG for a mass or density (14–18 gauge), mammotome for calcifications (9–11 gauge), and MRI for (non) mass enhancement (9–11 gauge). The mean size of the invasive carcinoma on biopsy, by modality and correlation with excision is shown in Table 3. Further analysis of the size difference between biopsy and excision by dimension (mm) and stage is displayed in Figure 5. The size difference between biopsy and excision directly correlated with final pathologic tumor size and was statistically significant (coefficient = 0.92, P < .001).
The correlation of final pathologic tumor size and the tumor size difference between biopsy and excision.
The correlation of final pathologic tumor size and the tumor size difference between biopsy and excision.
Nineteen cases were microinvasive carcinomas and could not be evaluated due to insufficient tumor volume, all of which were mammotome biopsies done for calcifications that were associated with ductal carcinoma in situ. There were an additional 19 microinvasive carcinomas diagnosed on biopsy, 7 by mammotome, and 12 by ultrasound, which were changed to T1a tumors in 10, T1b in 4, and T1c in 5.
In 110 cases, we had detailed radiologic information regarding tumor size, which ranged from 3 to 20 mm (mean = 9 mm). There was perfect radiologic pathologic concordance in 11 of 110 cases (10%), 1- to 2-mm radiologic overestimation in 31 of 110 (28%), and 1- to 2-mm underestimation in 52 of 110 cases (47%). The remaining 15% of cases were overestimated by 20 mm in 1 of 110 case (1%) and underestimated from 3 to 20 mm in 16 of 110 cases (14%). No difference between ductal and lobular carcinoma was seen.
DISCUSSION
Pathologists have an important role in measuring invasive breast cancers to the nearest millimeter given the associated diagnostic staging and therapeutic implications. This is routinely done on excision specimens due to the usually larger size on these specimens compared with biopsies. Depending on the mode of detection and feasibility, biopsies of the breast are performed via different needle gauges; larger needles (9–11 gauge) used for mammotome and MRI-guided biopsies and smaller needles (14–18 gauge) for ultrasound-guided biopsies. With improved breast imaging modalities, most invasive carcinomas are detected on screening and reported to be smaller.7 Thus, it is quite conceivable that a small invasive breast cancer may be partially or completely removed by a large needle biopsy sample. In this kind of situation, it behooves the pathologist to also measure the invasive carcinoma on biopsy to ensure accurate sizing of the invasive carcinoma for treatment purposes. While some pathologists routinely measure the largest linear extent of invasive carcinoma on biopsy, others prefer not to unless necessary at the time of excision. Given the inconsistency of this practice, we sought to find a solution to this issue.
In our series, the largest of its kind both in total number of cases and by modality, we found that the largest linear extent of invasive carcinoma size was by far captured on excision specimens (82%, 755 of 922), the mean size being 14.4 mm, almost twice that on biopsy. This reinforces the College of American Pathologists guideline to measure invasive tumor on excision specimens.1 Of note, the majority of invasive breast cancers were indeed detected and sampled via ultrasound guidance (92%, 844 of 922), owing to the more obvious targeted screened mass lesion compared with mammotome and MRI-guided biopsies, done for nonmass-like indications, such as calcifications and enhancement. This was reflected in the mean tumor size on ultrasound guided biopsy measuring 7.8 versus 5.9 mm for MRI biopsy and 3.3 mm for mammotome. In the remaining cases, invasive tumor size on biopsy was either equal to (10%, 95 of 922) or larger (8%, 72 of 922) compared with excision, impacting staging in 26% (19 of 72) of the latter, or 2% (19 of 922) of all cases. When further subclassified by imaging modality, this was true for 16% (133 of 844) of ultrasound-guided biopsies, 38% (5 of 13) of MRI, and 44% (29 of 65) of mammotome biopsies, albeit underpowered for the latter 2.
By stage, 100% of T1mi (19 of 19), 70% of stage T1a (68 of 97), 24% of stage T1b (59 of 248), and 6% of Stage T1c (21 of 366) tumors were larger or equal in size on core biopsy compared with excision. The average size of these cancers was 5.5 mm compared with 16.3 mm for tumors that were larger on excision. Half of the invasive carcinomas that were larger on biopsy than excision were less than 5 mm, whereas those that were larger on excision compared with biopsies were greater than 5 mm, and this finding was statistically significant. Invasive carcinomas that were equal in size on both biopsy and excision were predominantly (88%; 84 of 95) less than 10 mm, 20% (19 of 95) of which were microinvasive carcinomas. The latter were present in 4% of biopsies, half of which were changed to stage 1 tumors mostly biopsied by ultrasound, the other half remained microinvasive on excision with no change in stage, diagnosed exclusively by mammotome, performed for calcifications that were associated with the ductal carcinoma in situ. The difference in size between the biopsy and excision ranged from a lower limit of 1 mm for all stages to 5 mm for T1a and T1b tumors, 10 mm for T1c and T2 tumors; thus, directly correlated with stage, and was statistically significant. Most of the carcinomas were well-to-moderately differentiated and estrogen receptor positive (Tables 1 and 2).
Based on our study, it seems necessary to measure invasive carcinoma on biopsy in 18% (167 of 922; biopsy tumor size larger [72 of 922; 8%] and equal [95 of 922; 10%] than excision) of cases, or approximately 1 in 5 cases, similar to values reported by others.2–4 In these types of scenarios, invasive carcinoma on core biopsy may be the sole determinant of pathologic tumor staging and ancillary testing. Factors predictive of larger or equal tumor size on biopsy than excision include stage 1 tumors (statistically significant), particularly those that are less than 5 mm, and sampled by mammotome biopsy. Approximately half of the invasive tumors sampled by mammotome biopsy were equal or larger on biopsy compared to excision. Thus, radiologic tumor size information when available may help identify stage 1 tumors and provide a correlative guide to measure the invasive carcinoma on biopsy. Our study showed good correlation between radiologic and pathologic tumor size in 85% (94 of 110) of cases (within 0–2 mm). While tumor size and needle gauge were affected by residual tumor size on excision, presence of ductal carcinoma in situ, grade, and subtype of invasive carcinoma was not.
Since our initiative to measure invasive carcinoma on biopsy in 2017, we have continued to do so and are in favor of continuing this practice. Breast cancer diagnosis almost always begins with a biopsy, which is either performed at our institution or at an outside institution(s), the latter required to be reviewed preoperatively as a consult at our institution. Depending on the institution performing the surgery, this involves an exchange of biopsy slides between institutions. Retrospective retrieval of archival biopsy or returned consult slides may be impractical and cause delays. Lack of knowledge of tumor size on biopsy may lead to a false assumption of final tumor size on measurement, which may impact staging of invasive carcinomas, particularly Stage 1 tumors. As we approach the era of digital pathology, this issue will most likely be eliminated because the core biopsy material would be both easily retrieved and measured. But for the time being, we prefer to provide tumor size on biopsy to enable comparison with the excision and accurate staging, regardless of the origin of the slides. Those who are against this practice fear that tumor size on biopsy and excision might erroneously be added leading to overestimation of tumor size and undesired consequences of over treatment. However, the College of American Pathologists and American Joint Cancer Commission clearly reject and warn against this practice.1,5 To help mitigate against this practice, we suggest issuing an addendum at the time of excision for cases that are larger on biopsy compared with excision, to provide the dimensions of the larger size and stage accordingly.
As the future of breast treatment evolves toward nonsurgical methods, particularly in pandemics, or with the increased use of neoadjuvant treatments, including endocrine treatment, measuring invasive carcinoma on biopsy may become necessary. Most tumors in our series were estrogen receptor positive, but a small subset was triple negative (10%) and HER2 positive (10%), for which determining tumor size on core biopsy maybe important for choice of adjuvant therapy. This is particularly true for the T1c and T2 tumors, which comprised the majority of the HER2-positive and triple-negative carcinomas. While most of these were best measured on excision specimens, the core biopsy was accurate in tumor size stage assignment in approximately half of the cases and thus may serve as a preliminary accurate guide to the treating oncologist. Similar to biomarkers (estrogen receptor, progesterone receptor, HER2), which used to be performed and reported exclusively on excisions, the widespread use of preliminary biopsy to sample invasive carcinomas has become routine practice and is now expected in the final report. Providing as much information as possible on biopsy empowers treatment decisions. We conclude that it is prudent to measure invasive carcinoma on both biopsy and excision, and correlate with imaging studies, so as to best determine the largest linear extent of tumor size for optimal patient treatment.
References
Author notes
The authors have no relevant financial interest in the products or companies described in this article.
This study was presented as a poster at the United States and Canada Academy of Pathology; March 20, 2018; Vancouver, Canada.